Defectiveness of Concepts

There’s a kind of pragmatism, call it Carnapian pragmatism, that concerns the adoption of languages. More specifically, Carnapian pragmatism, as I am using the term, combines two theses: (1) there are no a priori rationally indispensable languages, and (2) the adoption of a language ought to depend on the weight of the various benefits that speaking that language confers on one and one’s community. (1), rules out, for example, the a priori indispensability of a language involving material substances that persist through time and underly change. That is, of course, compatible with that language’s being as a matter of fact the obviously best language in which to (e.g.) conduct inquiry. (2) gives us a criterion scheme for choosing among eligible languages.

Now consider the following claim, a rough first approximation: there are some truths that one ought never to come to believe, because they involve defective concepts. Everyone, I think, will grant that there are some truths we ought not to come to believe for some significant sense of ‘ought’, perhaps because they would be too damaging or morally corrosive. But suppose the ‘ought’ is the ‘ought’ of inquiry: given the goal is furthering inquiry, one ought to φ (in general I don’t think this will be what is called the epistemic ‘ought’). The claim, then, is this:

(*) There are some unambiguous truths (where an unambiguous truth is simply true, and thus not also false) and subjects S such that (A) S is warranted in believing p, and (B) has considered whether p, but (C) ought not to believe p because p contains (or is expressed with) a defective concept.

Another way of putting things: are there concepts defective from the point of view of inquiry not just because they have no instances (or, more generally, cannot be used to express positive truths)?

I consider (*) to be stronger than Carnapian pragmatism. According to Carnapian pragmatism, there are languages we ought not to speak in view of what best furthers inquiry. Important progress in science can be made by hitting on these languages. But this thesis is just a thesis about language, and not about rational belief. Carnapian pragmatism does not forbid you from believing content composed of or expressed by bad concepts. As a view, it has a large amount of prima facie plausibility, whereas the claim of the previous paragraph looks pretty controversial from the first.

There are some concepts it would be better if we did not have, if only because they add clutter without doing much corresponding work for us. Hirsch’s incar and outcar might be good examples. Other concepts are more pernicious, because they encourage thoughts it would be prudentially better for us not to have. Slurs might be one example here. But if our only goal is inquiry, are there really concepts with which we can express truths, but where those truths simply ought not to be believed? It is just not plausible that, where we find ourselves thinking about incars, and we realize some true thing about them, that we ought not to believe that true thing. It’s just that it’s unfortunate that we ever set out thinking in terms of incars, since that’ll mostly be a waste of our intellectual resources.

A variation on Prior’s tonk might be the kind of thing we’re after. It’s characterized (incompletely, for present purposes) by its introduction rule (from A, infer A tonk B) and its elimination rule (from A tonk B, infer A). But these rules don’t give us any truth conditions for tonk, so let’s stipulate some: A tonk B is true iff A is true or B is true. Suppose we have a sentence like ‘Grass is green tonk Gibbard invented penicillin’ (call it G). is true, since the corresponding disjunction is true. But we ought not to believe G, since then we would be “licensed” to infer that Gibbard invented penicillin, a manifest absurdity. So perhaps this is a counterexample. I don’t think it is, though. Though it is true that G analytically entails that Gibbard invented penicillin, we are often not permitted to believe even what’s logically entailed by what we believe, as in preface cases, or where we have contradictory beliefs (not just anything goes in that kind of case). Similarly, though the concept is characterized by certain sorts of inference rules, it is impermissible to infer according to those rules, because we know that a huge variety of tonk-inferences do not preserve truth, and do not even pretend to.

To forestall an objection: we still have the concept tonk even if in most cases we don’t infer according to its characteristic rules, if only because we can argue about it and have good reason to say things like ‘tonk is a dumb concept’ with complete justification (compare Williamson on McGee and modus ponens). That is, I think I have at least one tonkish belief (G), though I haven’t yet inferred that Gibbard invented penicillin. So, we might be permitted (required, if the truth and our warrant are obvious enough) to believe truths expressed with this version of tonk, though we would also have to be sure not to make risky inferences.

Here’s a worry about that response to the tonk variation. Tonk will still differ from incar, because in worlds where incars are extremely important, truths about them become very important. That is, there are some possible beings for whom it would be useful to think with incar. But with tonk, no matter which possible world and which possible beings, or beats it in view of inquiry. Or and tonk make the same contribution to truth conditions, but with or, you need not worry about making the inferences that come with possession of the concept; with tonk you need to be on your guard. So or will always be at least as good as tonk. While I agree, I don’t see how this gets us (*). It would be better to reason with or than with tonk, no doubt about that. But ought we really to disregard the tonkish truth, just because or does better? That just doesn’t follow. The worry, then, is just an instance of Carnapian pragmatism, and does not establish (*).

(*) is an exciting claim, while Carnapian pragmatism seems old hat to me. It would be a way of combining conservatism with revisionism: “granted, the people here are saying true things, but what they’re doing is nevertheless defective”. So, it would be very interesting to find the truths it claims are out there, though I haven’t yet found any. For my own part, I’m inclined to think something like this: from the point of view of inquiry, anyone is permitted to believe any unambiguous truth for which they have undefeated warrant. If, for example, theological discourse is defective from the point of view of inquiry, it is defective because its positive claims are all of them false, since they would involve a concept with no instances. I don’t know if I have more of an argument for that claim than the ridiculousness of conceding to someone that some claim is true, and they have great reason to believe that claim, but even so they oughtn’t to believe it. We all just need to be careful about what we do with even the truths we believe. That might sometimes be difficult, practically speaking, but it seems always at least possible.

(Thanks to Umer, Boris, Zoe, Paul and Nick for the interesting discussion much of which I’ve translated here.)

About these ads

14 Responses to Defectiveness of Concepts

  1. Aaron Boyden says:

    Is G actually true? The proposed truth conditions for tonk say it is, but then the rules for tonk produce contradictions. I thus have at least some inclination to think that once you’ve started using the proposed truth conditions for tonk, you’ve stopped talking about truth, and started talking about some other concept, truth*, where truth* and falsehood* are not incompatible and truth* is not very much like truth.

  2. druckerd says:

    Dialectically, it’d be nice if it weren’t, since I wouldn’t need any more of a response to the example. And that was my first instinct when I heard this example. But I think we can say G is true, by analogy with the following case. Say that y is some slur. It seems like from “x is a y”, the concept itself puts me in a position to infer “x is bad”, even though (let’s say) badness is no part of what the first sentence attributes to x in what it actually says. (Presupposition triggers will work this way, too, even though they’re less deviant.) Let’s also (begrudgingly) say it’s _true_ that “x is a y”, as I feel with at least some slurs we will do. Even so, I’d reject thinking in terms of the concept, because the concept “says” we can make what I see as bad inferences. That doesn’t mean the sentence (“x is a y”) is no longer true. So long as we can make sense of conceptually licensed inferences that are nevertheless bad inferences to make, which I feel like we should anyway, then we can make sense of a concept with inoffensive truth conditions but wacky stipulations about inferences.

  3. Umer Shaikh says:

    Okay, I need to think about this more, but here are a couple off-the-cuff responses:

    1) “I consider (*) to be stronger than Carnapian pragmatism. According to Carnapian pragmatism, there are languages we ought not to speak in view of what best furthers inquiry. Important progress in science can be made by hitting on these languages. But this thesis is just a thesis about language, and not about rational belief. Carnapian pragmatism does not forbid you from believing content composed of or expressed by bad concepts. As a view, it has a large amount of prima facie plausibility, whereas the claim of the previous paragraph looks pretty controversial from the first.”

    Grant that it is psychologically taxing to speak in terms of the incar/outcar language, or some combination of risky and taxing to speak in terms of the tonk language. If this is right for speaking, it seems it’d be a factor when thinking to oneself, too. I don’t see any motivation for adding to my belief box a tonk truth for every or belief, and it seems just as bad for belief as for spoken language: it seems to be courting those risks in my internal mental life.

    (I do have to depend on an idea of belief here where it makes sense to believe, simultaneously, the content of or statements and not to believe that of tonk statements. I am fine with this, and I think it is much harder to get (*) going if you don’t accept such an idea, or maybe even if you accept such an idea only grudgingly to avoid explosion.)

    A clarification. I _do_ have some beliefs about tonk; I’m expressing them right now. I probably have beliefs using tonk as a connective, too. I do seem to have the concept tonk. But (*-tonkversion) says at least sometimes we shouldn’t believe a statement with tonk in it, not we should _never_ believe a statement with tonk in it, or we should purge tonk from our mental lives entirely. Go ahead and have beliefs containing tonk in them, and go ahead and express them in the philosophy classroom. It is their use in your everyday speaking and mental life that would be bizarre. Perhaps this was obscured by talking about whole languages to start, but it seems fine as far as (*) is concerned, which looks only for some true warranted beliefs you shouldn’t have because of their defective concept.

    Similarly, it’s true that we don’t always infer from our beliefs, and it’s probably true that it’s not always epistemically permissible to infer from one’s beliefs. But the tonk-objector doesn’t need to rely on the premise you used, that having a tonk-belief in the box licenses one to infer from it. All I need is that the sorts of mechanisms you use to prevent explosion in the self-contradictory belief system case, whatever they are, are at least somewhat psychologically costly, or indeed, have any sort of drawback at all. (see below)

    Notice too that the restriction in the tonk case is also much sharper than the preface paradox case: _any_ elimination of tonk either replicates what you already believe by other means, replicates what you could already infer from elimination on the corresponding or statement, or takes you to an unjustified belief. (I think I got all the cases?) Again, this seems important: the tonk concept doesn’t add anything, even when handled perfectly; the only thing that would suffer on purging the concept entirely would be your philosophical work. But clearing out all the contradictions in your belief box generally would be a destructive thing to do, even if you’re not a philosopher.

    2) “Or and tonk make the same contribution to truth conditions [...] or will always be at least as good as tonk. While I agree, I don’t see how this gets us (*). It would be better to reason with or than with tonk, no doubt about that. But ought we really to disregard the tonkish truth, just because or does better? That just doesn’t follow. The worry, then, is just an instance of Carnapian pragmatism, and does not establish (*).”

    Here’s how to get “don’t use the tonk language” from or’s being always at least as good as tonk. Let OR be the language that has the or connective and TONK be the language that has both the or and tonk connective. For simplicity let’s ignore the fact that we do want to keep TONK around for some features, like philosophical discussion. A more detailed argument either systematically lets OR use tonk, but only in certain contexts or something, or we can just switch our example supporting (*) to someone who hates philosophy. Obviously these are not intended to be just spoken languages, but the languages in which you “think to yourself.”

    i) TONK requires the users of the language to not infer (in a non-hypothetical context) from tonk sentences (if the user’s beliefs are not to explode).
    ii) The restriction in (i) is difficult to manage psychologically.
    iii) If TONK requires a difficult to manage restriction in order to avoid explosion, that is a drawback to using TONK.
    iv) TONK does not have an advantage (e.g., an expressive advantage) over OR.
    v) We should not use a language that does not add any advantages to its competitors and requires a difficult to manage restriction to avoid explosion.
    ____
    vi) We should not use TONK.

    A word about (ii). You probably have little trouble managing the problem now. But we just need one case, so imagine the case where a tonk belief matches every or belief, where tonk is pervading the mental economy. It might not be a _lot_ of trouble. But it just has to be some, in order to be an unbalanced drawback.

    Another way to go is not to talk about difficult restrictions but useless redundant beliefs, or better, both drawbacks. I won’t repeat the argument with this consideration added in; I trust it is easy to see how to do it, and I am lazy.

    I certainly have the intuition that the tonkish truth is no real addition to the orish truth, and disregarding it is neither a practical nor epistemic cost. Notice that if you disagree, you should say something about what is special about tonk. Surely it is not always worth not disregarding true beliefs in more and more n-ary connectives, equivalent to combinations of the old connectives…. We need something in between “all truth-equivalent beliefs are epistemically just as good as one another” and “any syntactic difference in expression / any difference in concepts used in expression makes a new potential belief, worth believing in addition to the old one.” I don’t know how to thread the needle, but there does seem to be some space there (yay for mixed metaphors).

    One last thing….

    3) “That is, there are some possible beings for whom it would be useful to think with incar. But with tonk, no matter which possible world and which possible beings, or beats it [or does just the same] in view of inquiry. [...] While I agree, I don’t see how this gets us (*).”

    The point of the incar-beings was not to get you (*). The point was to show one way in which badness of concepts might differ from concept to concept, to begin a typology of bad concepts. While the badness of incar seems to have as much to do with our natures (or the natures of cars and garages, I suppose, as in your example) as with the content of the concept, tonk plausibly seems less like this. It’s badness comes almost entirely from its own content, combined only with the fact that our cognitive resources are not infinite or incredibly cheap. Tonk seems much closer to being an a prioi defective concept, or necessarily bad, or something of this sort; it relies on less contingent facts about human nature. Even for beings with unlimited cognitive resources, for whom there is no cost to using tonk, it does not seem to me there is any particular positive reason to use it, either.

    Here I’ll repeat the caveat from before: such angelic creatures might happen to get a mighty burst of energy from inferring to a tonk statement, energy only useful for logic. But this epistemic usefulness has nothing to do with tonk’s power to describe the world, with its content qua content.

    –Umer

  4. Umer Shaikh says:

    “I certainly have the intuition that the tonkish truth is no real addition to the orish truth [...] Surely it is not always worth not disregarding true beliefs in more and more n-ary connectives, equivalent to combinations of the old connectives…. We need something in between “all truth-equivalent beliefs …” [and] “any difference in concepts used in expression makes a new potential belief, worth believing in addition to the old one….”

    Sorry, this was unclear on my part. Assume p is true. The point was, I don’t think

    1) p
    2) p and p and p and p and p and p [repeat as many times as you wish]

    Are both (in general) epistemically worth believing, but only (say) limited cognitive resources prevent us from believing the second as well as the first. When we find the criterion for individuating beliefs in the way relevant here, what true content is worth having, I suspect 1 and 2 get lumped together, and hope I have not thereby committed myself to lumping together all truth-equivalent content.

    –Umer

  5. druckerd says:

    I feel _no_ impulse to reason tonkishly, even though I believe G. It seems inappropriate to even say that G is quarantined. I don’t expend any effort to not reason tonkishly. The question doesn’t even arise. I actually find it far harder, far more mentally taxing, to suspend judgment on G once I’ve considered it, once I’ve been presented with it. (Forgetting is seriously hard work a lot of the time.) So it seems for people like me (and I imagine most people confronted with G will be like me), there’s reason _to_ accept G, and little to, realistically, no reason not to.

    • Aaron Boyden says:

      I still am not sure we should be in such a rush to say people believe G. I can’t help but suspect that people really believe only that “Grass is green or Gibbard invented penicillin,” and that they are mentally substituting “or” for “tonk,” perhaps out of the sort of automatic charity we generally apply to try to make things people say make sense, when they claim to be believing the original G. Especially looking at druckerd’s comment; I’m inclined to think that if you have no impulse to reason tonkishly, you aren’t interpreting anything as being a true tonk claim.

  6. Umer Shaikh says:

    I tend to think that people draw inferences quite liberally, often without thinking about it explicitly very much. (I’m not just thinking of the inferences you draw in the calm of your philosophical consideration, sitting in front of the fire.) I guess I just have different suspicions about folks’ reasoning practices; I’m not sure there’s much more to be done here from the armchair, though.

    “So it seems for people like me (and I imagine most people confronted with G will be like me)….”

    Let me reiterate that (*) is incredibly weak in at least one way: all it takes is some concept, some truth, and some subject.

  7. druckerd says:

    There’s something interesting going on here. I wanted to put (*) in terms of an ‘ought’ having to do with inquiry. I used that for a shorthand for one of the two kinds of ‘ought’ that gets called epistemic (so the ‘ought’ in ‘He ought to think she’s in her office’ and not in ‘She ought to be in her office’ (in view of one’s evidence)). So, for a second let’s not say that the relevant ‘ought’ is the one that furthers inquiry, but rather the ‘ought’ in the sense of that first sentence.

    In that case, I think your argument, Umer, is appealing to something like an epistemic consequentialist principle like this: what grounds S’s duty to φ, from the point of view of inquiry, is that φing will maximize the amount of truths that S gets. But that’s a pretty problematic principle, precisely because it encourages us sometimes to reject obvious truths because of all the extra truths we’ll get. (I’m thinking of Berker’s paper here.) Your argument would also apply to other mentally exhausting concepts where the number of new truths we get because of new expressive power doesn’t really outweigh all the truths we’ll miss out on because of the sheer difficulty of the concepts. (Perhaps some concepts of math are like this. I don’t know, I’m trying to think of some good examples right now.) So I’m wary of the kind of argument you’re pressing.

    • Umer Shaikh says:

      I’m not familiar with Berker’s paper. Is this the believing something in order to get funding, in order to get further true beliefs example? I’m not sure I have a problem with the supposed absurdum. In any case I need not be committed to accepting any consequences of your consequentialist principle. Compare: we all reason ethically in terms of consequences, sometimes; this doesn’t mean that other considerations can’t also step in, in certain circumstances. That is to say, I need not invoke such an unrestricted principle as you propose.

    • Umer Shaikh says:

      (Er, now that I look back on it, I did use a pretty sweeping consequentialist principle in my first comment, (v), although it was also weak in that (I claimed) giving up on TONK wasn’t any sacrifice. I bet you’re quite right that that was ill-advised.)

  8. Zoe Johnson King says:

    [[Background: via email, Boris introduced the following rival to Daniel’s (*):
    “(**) There are some concepts F( ) and subjects S such that [i] F’s extension is non-empty, but [ii] F( ) is epistemically defective (defective from the point of view of inquiry).”

    Then Umer gave this great example:
    “The extension of is me, so you can certainly use it to state truths. But it has the curious elimination rule (if it can be called that!), understandable, it seems to me, if the corresponding property of is understandable, that you may infer any statement from a statement containing . The concept has an extension but I would think is epistemically defective if any concept is. What exactly that means you ought to do with is another question.”

    Now speaking in my own voice:

    Hahaha, nice example Umer! So, if I’ve understood the dialectic correctly so far, it looks like the problem with both and – the thing that makes them inherently defective concepts, if Umer’s right that it sometimes makes sense to think that way – is the fact that, given their introduction and elimination rules, you can infer anything from statements containing either of them. One reason why this is a problem might be that the concepts aren’t truth-preserving; if we know we’ll be permitted to infer anything, true or false, then perhaps we ought to exercise caution in believing and/or asserting even true statements using these concepts, in case we get carried away somehow and end up inferring some falsehoods! (I thought that something like this was behind the idea that these concepts are not helpful epistemically/for inquiry. Whether or not I’ve got that right, I return to the idea at the end of the post.) Another reason why it’s a problem might be that (i) given our limited cognitive capacities, it counts against a concept if we are permitted to infer anything from statements containing it bc that’s a massive headache for us, and (ii) nothing good is gained by using these concepts to counterbalance that strike against them, since we already have the perfectly decent concepts and who can do all the good things and can do but that have none of the costs.

    If this is right, then (*) and (**) are both cool and interesting claims, but they seem pretty far away from the claims Price is making in the article that originally spawned our discussion (http://prce.hu/w/preprints/pragmatism.pdf). Price certainly doesn’t seem to think that the problem with normative and theological concepts is either that their introduction and elimination rules mean that their use risks carelessly inferring some falsehoods (I’m not even sure I could state introduction and elimination rules for these kinds of concepts!), nor does he think that there are other concepts that secure all the benefits of normative and theological concepts but that are truth-preserving. So it looks like Price’s grounds for thinking that normative and theological concepts are defective, and as such ought not to be used, must be different grounds to these.

    I thought the slurs example might be more illuminating as an analogy for what Price is trying to do. We could put it in a way that’s pretty similar to the problems with and : if “X is a mick” presupposes “X is bad” (whether or not it’s a part of the concept – I 100% retract the wrong thing I said about that the other day!), then perhaps even true statements using the concept ought not to be believed or asserted, because we’ll be liable to infer that the mick-person is bad by whatever mechanism we usually infer presuppositions from statements, and yet it may well be false that s/he is bad. This is pretty similar to the problem with connectives that allow you to infer anything, whether true or false, but it works without reference to introduction and elimination rules – we just need to identify specific false things that using a concept can be expected to lead us to infer (in this case, because the false thing is presupposed by true sentences using the concept), and that’s enough for us to reject it as dodgy. This seems like it could conceivably be the kind of thing Price is trying to do for normative and theological concepts; the false presuppositions of true statements using these concepts might be e.g. that it makes sense to ask whether the concepts are representational, or that it makes sense to ask whether objects exist that fall under the concept.

    Actually, I guess we could even develop an analogue for the second kind of problem with and for slurs. If slurs have a descriptive content that gives them truth-conditions, then we could say that this content will be expressed in perfectly decent concepts that can do all the good things a slur can do but without exposing ourselves to the risk of inferring falsehoods: for , for example, we cld just say and that would do the trick! I’m not sure whether we can do the same thing for normative and theological concepts (as understood by Price). It could work pretty well for thick concepts, with respect to which the argument would be that we should just use concepts expressing the descriptive part and forget the dodgy evaluative part. It could also work for other normative concepts if you’re some kind of naturalist reductionist – but Price certainly isn’t that.

    I have a bit of unease still about prospect of ruling out concepts as defective on the grounds that believing/asserting true statements that use the concepts involves riskily exposing ourselves to the possibility of inferring falsehoods. Mainly, the unease is because I don’t think I’ve/we’ve spelled out the problem very well; earlier I just spoke of getting “carried away” by and , and I guess the picture was one of some unreflective, inference-hungry agents who carelessly infer stuff from stuff whilst caring not a fig for the truth or falsity of the new statements they come to accept, because they just looove inference. Or something along those lines. If that’s the picture, I’m not sure how persuasive it is as a prudential argument against using certain concepts – surely we’re not that stupid? I feel less uneasy about the idea that true statements with false presuppositions are “risky”, because it doesn’t seem so gung-ho and irresponsibly daft to infer presuppositions from statements that have them – rather, it seems like a totally normal and legitimate part of conversation and belief-revision. But I worry that an argument of that form might prove too much, and end up casting out as defective some concepts that are in perfectly good working order, just because /some/ of their uses in true statements can have false presuppositions. I don’t have any good counterexamples yet though – I’ll have to keep thinking about it.

    — ZJK

  9. bbabic says:

    So just a clarifying comment now that (**) is in play. The idea was to rephrase the following suggestion by Daniel in the original post:

    “Another way of putting things: are there concepts defective from the point of view of inquiry not just because they have no instances (or, more generally, cannot be used to express positive truths)?”

    If I have re-described the above correctly, it is a different issue from (*). The latter is about whether or not you should believe the truth you get to via the contested concept. The former is about whether you ought to use the concept. You might not object to S’s belief that ‘2 ⊕ 3 = 5′ is true, while at the same objecting to S’s use of quaddition.

    I, like Umer, would likely agree with (**). I’m less sure what I have to say about (*). In any case, it seems to me that the issue as clarified by Zoe then becomes stating the conditions under which the second condition of (**) obtains; namely, the conditions under which a concept with true instances is epistemically defective. If an argument of the form Price gives (as I understand it from you all) is to have some weight, then you would expect to have some such conditions, or at least reasonably persuasive standards. There are easy cases — slurs, the quaddition function, Umer’s example, and so forth. But even in these cases, it is not obvious what connects them all — slurs may entail the risk of falsehoods, Umer’s example allows us to infer too much (a more general problem of the problem with slurs), and the quaddition function has problems related to mathematical insight or fruitfulness (an altogether different problem, if you agree that it is a defective concept).

    So what, if anything, justifies tossing a concept into the defective trash bin? Also, would any attempted statement of such conditions rule out too many perfectly good concepts (as some of Zoe’s concerns suggest)? If so, is it really obvious that, say, Umer’s example is epistemically defective? That is, should we be in the business of legislating which concepts are legitimate? If not, is there any persuasive weight left in Price’s argument?

    (I’ll have to look at the article before I can contribute more meaningfully.)

    – B

  10. Tristan Haze says:

    Great post! I’ve often wondered about something like (*).

    There’s something bothering me about this reply to the idea that propositions involving the incar concept might serve as examples:

    ‘It is just not plausible that, where we find ourselves thinking about incars, and we realize some true thing about them, that we ought not to believe that true thing. It’s just that it’s unfortunate that we ever set out thinking in terms of incars, since that’ll mostly be a waste of our intellectual resources.’

    Given a true proposition P involving the incar concept, one question is whether one ought – the ought of inquiry – believe P. Another is whether, given that one is thinking in terms of incars and that’s OK, one ought to believe P.

    It could be that one oughtn’t – ought of inquiry – be thinking about incars at all (because it’s a waste of resources, for example), and so even if we find ourselves thinking in incar terms, that doesn’t mean we should believe *anything* involving them. Perhaps we ought (of inquiry) to get out of the line of inquiry altogether.

    (Still, I doubt that the incar case is an example showing (*), together with its annexed (C), since ‘incar’ doesn’t intuitively seem *defective*. So this is mainly about the reply, and about teasing apart ‘true but ought (of inquiry) not be believed’ from ‘true but involves a defective concept’.)

Follow

Get every new post delivered to your Inbox.

%d bloggers like this: