Leiter Reports: A Philosophy Blog

News and views about philosophy, the academic profession, academic freedom, intellectual culture, and other topics. The world’s most popular philosophy blog, since 2003.

  1. Fool's avatar
  2. Santa Monica's avatar
  3. Charles Bakker's avatar
  4. Matty Silverstein's avatar
  5. Jason's avatar
  6. Nathan Meyvis's avatar
  7. Stefan Sciaraffa's avatar

    The McMaster Department of Philosophy has now put together the following notice commemorating Barry: Barry Allen: A Philosophical Life Barry…

Catch-all Solutions to Philosophical Problems

I’m now entering my second decade as a professor, and feel
like I have amassed enough experience to point out certain trends. I’m not sure
if the trends I want to talk about are the “normal science” of philosophy, or if
they’ve been particularly noticeable in recent years, as philosophy has gotten
more professionalized. What I’ve noticed is that a certain methodology for
addressing a philosophical problem will arise in one particular sub-literature,
where by “methodology” I mean some kind of quasi-technical mechanism for
resolving disputes in any area. Then, that strategy will start to replicate
(much like a computer virus); people will apply it to all the other fields
where it has not yet been applied. For example, at first fictionalism was
advanced to treat the problem of negative existentials; then it was applied by
Gideon Rosen to modal metaphysics. In the intervening years, fictionalists have
appeared in every discipline (fictionalism about morality, fictionalism about
abstract objects, etc.), and in the late 1990s, being on a job-search committee
meant wading through endless stacks of papers advocating fictionalism about
this or that. Contextualist solutions
to philosophical problems arose in a more haphazard way; in philosophical
logic, in application to the liar paradox and the sorites paradox, in
epistemology to the problem of skepticism. Now, a contextualist solution is
part of the standard tool-box of possibilities when faced with an apparent
philosophical problem. Now, we’ve got relativism
about truth
, the idea that the truth of a proposition is relative to an
evaluator, and no doubt serving on a junior job search committee will involve
wading through stacks of papers applying relativism to this or that apparent
dispute.

Over a year ago, Brian Weatherson declared that relativism
was going to be a central topic in philosophy in the next decade:

 http://tar.weatherson.net/archives/002685.html

Brian has turned out to be correct.
For example, at the Pacific Division of the American Philosophical Association in
Portland next year (not the MLA!),
there are two invited symposiums on Saturday on relative truth; you can spend
all day listening to philosophers dispute the topic. It’s like fictionalism five years ago.

Over the years, I’ve written papers or chapters of books criticizing
fictionalism, contextualism, and relativism. At each point, I thought there
were specific features of the methodology and its applications to which I was
reacting (and certainly, each strategy raises different issues). But I’ve
realized that I have a uniformly negative reaction to methodologies application
of which would resolve a host of apparently distinct philosophical problems. I
think there are two sources for this. First, I think the philosophical problems
are sufficiently distinct from one another that I’m immediately suspicious of
any attempt to resolve several of them by appeal to one mechanism (be it covert
fictions, context-sensitivity, or relativism). I could be convinced
(though I haven’t been so far) that some traditional philosophical problem (be
it future contingents, or skepticism, or the problem of abstracta) is due to
our failure to recognize context-sensitivity, or truth-relativity, or that
we’ve stumbled unknowingly into a fiction. But I think it is prima facie pretty unlikely that the
right account of the sorites paradox will have much to do with the right
account of skepticism. Secondly, since the methodologies are usually awfully
easy to apply, I worry about how to constrain
them. In short, each methodology seems to make philosophy too easy. I’m certainly
not saying these worries can’t be answered by advocates of catch-all methodologies. Many of the profession’s very best and most careful philosophers are attracted
to them. But I do think there is a special burden, when advocating a catch-all
methodological solution to a philosophical problem, to respond to these two
worries. 

                                                                                                     -Jason

Leave a Reply to Kenny Easwaran Cancel reply

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

13 responses to “Catch-all Solutions to Philosophical Problems”

  1. Unless there is some reason simply to ignore methodological questions, there ought to be some way (method?) of assessing the adequacy of methodological theories. (If there were no such way, that might constitute a reason to ignore such questions.) One way that suggests itself is to ask what consequences would follow if a proposed methodological theory were true. If there were no such consequences, that in itself would be a reason to reject the theory. (Having a wide range of application is not normally thought of as a vice for a theory–in fact "fruitfulness" is widely recognized to be one mark of a successful theory.) If there are such consequences, then the question arises whether they are plausible. Can't the application of a methodological theory to a more specific, "applied", issue be seen as a contribution to a standard way of assessing such a theory? That gloss seems more charitable than that of dismissing such efforts as mindless trend-following.

    However haphazardly a methodological theory may have arisen, if it suggests what seem, independently, to be plausible if novel responses to seemingly disparate philosophical problems, that should count in its favor even if it does not fully dispel the presumption that prima facie distinct philosophical problems are not to be conflated merely for methodological convenience. As for the worry that application of a methodology makes philosophy "too easy," it might be enough to say in response that a successful methodology has to yield plausible answers to multiple questions–and, if that too proves easy, then more power to that methodology.

  2. Sigh. Yeah, that's what my friends keep saying.

    So, right, one would hope that there are methods of assessing the adequacy of methodological proposals. That's why I've written a number of papers, for each such methodology, disputing it for particular cases. But I think you're wrong about one point. If it's too easy to apply that methodology to particular cases, that is often a serious worry for it. For example, as Keith DeRose points out in a number of places against "Warranted Assertibility Maneuvers" (basically, appealing to "pragmatics" whenever one faces recalictrant data for one's theory), one could use the strategy too easily to defend any semantic proposal whatsoever. For example, one could say that "bachelor" semantically expresses the property of being a man, and to defend one's proposal, argue that it's just pragmatically odd to apply it to married men. The fact that it's so easy to use Warranted Assertibility Manuevers to explain away any worry with a view is a serious problem with the strategy. That's not to say that it's *never* legitimate to appeal to pragmatics to explain away apparent problems with a theory. There are a number of cases (many discussed by Grice) in which it is the right thing to say. But it is to say that there had better be clear rules about when it's o.k. and when it's not o.k..

    I think there are such rules that are emerging with contextualism — if you're claiming that a construction is context dependent, it had better bear certain resemblences to uncontroversial context-sensitive constructions. Furthermore, like the appeal to pragmatics (Warranted Assertibility Manuevers), contextualism has the advantage that there are certain clearly correct applications of it to disputes. For example, if John utters "I am tired" and Bill utters "I am not tired", the account of why they are not genuinely disagreeing is obviously the contextualist account (I'm also attracted to contextualist accounts of indicative conditionals and epistemic modals). In contrast, there is not a single uncontroversial case in which fictionalism and relativism are clearly the right account of an apparent disagreement (Like Peter van Inwagen, I'm not even a fictionalist about fiction!). So there is an extra burden of proof on those who advocate these strategies.

  3. No need to be conservative! We can countenance context-relativities of many heretofore unsuspected kinds. Just make sure that there is something definite to say about how the situation in which a conversation takes places determines the values of the context variables in any particular case. Also, it should not be too hard for interlocutors to be sensitive to the pertinent features of the situation. Also, whatever your semantic values and whatever your context variables, for each semantic value V it should be possible to give a recursive "definition" of "S has V in context C".

    Speaking of methodologies run wild, how about possible worlds semantics, huh?

  4. Like Prof. Edmundson, I'm more inclined to see unified answers to difficult philosophical problems and paradoxes as plus point. Obviously we don't want to uncritically apply methods that have worked against one problem to disparate problems, and perhaps that has been happening to a certain extent with fictionalism and contextualism.

    But in general that is something it is going to be hard to establish. It is going to be very very difficult to judge when to take the first worry seriously and when it will stifle work on a problem. How do we decide when a proposed connection between two problems is bogus or an insight? These questions look just as difficult as any other philosophical questions. There's no more consensus on the inter-relations between, to pick up the example, the relationship between the sorites paradox and scepticism as there is on what the solutions should look like. So I'm inclined to think that there is no additional burden on a theorist who proposes, say, a contextualist solution to both to explain why we shouldn't be suspicious. If the proposed solutions are bunk, then they are bunk; if they're plausible, it is hard to see on what grounds common methodology should defeat that plausiblity, because there don't seem to be any readily available independent criteria for deciding when two questions are suitably interconnected to permit such common solutions.

    And as for making philosophy too easy, it seems partly up to the opponents of fictionalist, contextualist and relativist proposals to make sure that isn't so. They seem to have been doing a good job so far.

    (I've just seen Stanley's response when hitting the preview button, but I won't attempt to revise what I've written.)

  5. Chris,

    The problem with what you're suggesting is that one needs some constraint on the postulation of contextual variables, or else one will be able to get any result one wants. In short, there needs to be some empirical constraint on the postulation of contextual variables (as I've argued in my papers in philosophy of language), or else one ends up with a theory that allows any sentence to express any proposition. For example, Hanna Kim, a graduate student currently at Michigan, argues against Joseph Stern's clever arguments that there is a 'metaphor' parameter in the syntax (an unpronounced operator) that there is no syntactic evidence for it, and one can't just postulate contextual variables (or operators) on semantic grounds.

    It's true that explicit quantification over possible worlds yields a system that seems to have greater expressive power than ordinary English. But arguments by Max Cresswell (*Entities and Indices*) and Graeme Forbes (in his brilliant *Languages of Possibility*) suggest that this is in fact false, and that indexed actually operators (which Cresswell argues are a good interpretation of "actual" in English) can give one the full power of quantification over worlds. In any case, this kind of worry with possible world semantics is a very different kind of concern than the ones I've raised for fictionalism, contextualism, and relativism, isn't it? Possible worlds semantics doesn't provide us with a catch-all method of cranking out a solution to any apparent disagreement, as the other frameworks do.

  6. It seems clear that "catch-all solutions to philosophical problems" are generally problematic – philosophy just generally isn't that easy. However, if someone does come up with a good new argument, it often does have some relevance elsewhere. Grice may have come up with his thoughts on pragmatics to argue that various connectives are in fact truth-functional though the appear not to be – but whether or not that was the original application, there are many other situations where his method is applicable. Of course, it's not universally applicable.

    To some extent, it seems to me that the exportation of methods from one area to the rest of philosophy (virtue epistemology is the first example that comes to mind beyond the ones you mention) could be seen as a sign that philosophy is at least semi-scientific. It's quite common in the sciences (especially mathematics, which which I'm most familiar) to take a method first used on one problem and apply it to a similar-looking problem in a somewhat distant field. This is often how important breakthroughs are made. But I suppose when some method seems to be applicable absolutely everywhere, it's a sign that the method is most likely being misapplied.

  7. Jason: Thanks for this response. Notice that you're not falling back on conservativism in this answer, and my main point was just that there is no special reason to expect all context-relativities to resemble the context-relativity of demonstratives. But I didn't really mean that there were no other constraints beyond those I listed. Syntactic arguments are good when you can get them, but I definitely cannot agree that "one can't . . . postulate contextual variables . . . on semantic grounds". I left out the word "just" in quoting you, because I agree that one cannot "just" do this. There has to be some good reason to do so. An additional good reason to do it is the need to capture the logic of natural language. In defining logical validity we will require that the conclusion be evaluated relative to the same context as the premises. (How exactly to do that is a question.) Seeking a definition of logical validity of this kind that draws a credible distinction between the valid arguments containing conditionals and the invalid ones can give us a reason to treat the semantic value of a conditional as essentially relative to features of the context.

    As for possible worlds semantics, the excesses I am thinking of are the attempts to explicate all kinds of things that have nothing especially to do with possibility and necessity in terms of possible worlds. The list includes: causal relations, fictional discourse, propositional attitudes, epistemic possibility, and even conditionals. Even the idea that we should explain modal operators such as "possibly" and "necessarily" did not come from a serious examination of the logic of natural language. (If it had, I don't think people would ever have assumed that possible worlds are complete.) Maybe there's an important difference between this case and the others you suggest, but if so, I don't see it yet. On the other hand, I don't offer this as an objection to any particular application of the possible worlds idea.

  8. OK, I'll bite. I admit to not understanding what it would mean for possible world semantics to be a methodology in the relevant sense. I would have thought that this was a framework — a toolbox if you like — for expressing various theories about the semantics of natural language. It turns out to be a pretty powerful set of tools, allowing a more-or-less unified approach to all kinds of constructions including the stuff Chris lists. (EVEN conditionals? Lots of presupposition failure going on there for me….) Of course, if the data suggested that there were differences between these kinds of constructions that treating them in this more-or-less unified way essentially distorts or masks, then that's a reason to prefer a different tool. Whether that's the case is an empirical question. (No surprise here but I do have a view about how the data line up.)

    The case of contextualism (and the other -isms Jason mentions) looks a bit different. I take it what Jason finds objectionable is that there is a standard "Contextualist Move" to be made for just about every philosophical problem we might encounter. I have at least a dim view of what that amounts to, but it is hard to imagine what a standard "Possible Worlds Move" would amount to — unless, of course, the "Move" were so gerrymandered that it amounts to saying that we can propose a fragment that treats this or that construction as intensional. But that's not much of a move over and above saying that the language of possible worlds is capable of expressing a particular theory about the semantics of this or that construction. But that's not terribly objectionable: it doesn't provide a one-size-fits-all solution to any problem; but it may provide a common framework for expressing a solution to a problem, should you have a solution in your back pocket.

  9. I'm sympathetic to Jason here. A methodology that promises to solve all problems ends up potentially solving none, for it doesn't tell us why the problematic cases are different to non-problematic cases. But I disagree with this…

    In contrast, there is not a single uncontroversial case in which fictionalism and relativism are clearly the right account of an apparent disagreement (Like Peter van Inwagen, I'm not even a fictionalist about fiction!).

    PvI is a fictionalist about fictional characters, not I think about fiction itself. Fictionalism is clearly the right way to resolve the apparent problem that the following three sentences all sound true (or at least highly plausible), but nevertheless are contradictory.

    1. Gladstone met Sherlock Holmes.
    2. Sherlock Holmes was the greatest detective of the 19th century.
    3. Gladstone never met the greatest detective of the 19th century.

    If you want to put this in dispute form, imagine A strongly endorsing 1 and 2, while B endorses 3. The simple resolution is that 1 and 2 are true in the fiction, while 3 is true simpliciter.

    As for relativism, the example won't be entirely uncontroversial, but I think most people believe something similar about world relativity. A is in a world where pigs fly only on weekdays, B is in a world where pigs fly only on weekends. They utter the following claims:

    A: Pigs fly on weekdays.
    B: Pigs don't fly on weekdays.

    Again, contradiction it seems, but both are true. Now there is a contextualist move here (I believe somewhat popular at Rutgers) to say that A is really uttering the proposition, Pigs fly on weekdays in w1 where w1 is her world. But the most common view among philosophers, by far, is that they utter the same proposition, and this proposition is true relative to one world, and false relative to another. Obviously this is not a very close analogy to the uses of relativism that MacFarlane, and Egan, and I etc want to use it for, but then neither is 'knows' very much like 'here'.

    If the upshot of the relativism debates is that people decide that even world-relativism like this is implausible, and we have to offer a contextualist resolution of A and B's 'dispute', that would be serious progress by my lights. It would mean I was monumentally wrong about a lot of things, but at least I'd be wrong in a useful way. (Brian as the useful idiot of philosophy…)

  10. Let me concede something to Thony and then take two bits back. If what you're calling contextualism is just the "contextualist move", then you're right: Possible worlds semantics is not like that. I guess the contextualist move is something like this: Someone provides an analysis of something. Someone provides a counterexample. Then, instead of revising the analysis, the first person restricts the analysis to certain contexts that exclude the context in the counterexample. Is that what you had in mind?

    The first bit I take back is this: I'm not sure that many cases where people have seriously wanted to introduce context variables fit this model. Jason mentioned the liar paradox, the sorites paradox and skepticism. It's tempting to try to fit contextualism as a response to skepticism into this model, but I'm not sure it works even there. (People don't think that contextualism answers the Gettier counterexamples, do they?)

    The second bit I take back is this: Jason didn't try to define what he meant by a methodology and I don't see anything in what he said that would exclude the inclusion of the possible worlds idea. (The possible worlds idea is not just an idea for semantics. Counterfactual theories of causation do not usually pose as theories of meaning, do they?)

    I don't think possible worlds semantics is so powerful, but I would never try to make that point on the grounds that the possible worlds idea has been applied very widely. On the contrary, that is prima facie evidence of success. No, to criticize, one has to consider the merits of each specific application. The reason I brought it up was that Jason had touted conservativism as a virtue and this was another way of chiding him. People are right to view contextualism with some skepticism, but if they withhold that skeptical attitude from the possible worlds idea then—in my opinion—that can only be due to conservatism.

  11. O.k., was in New Brunswick mock-interviewing and teaching all day, and haven't gone online. Now I've returned to find everything said.

    I think I agree with every single thing in Thony and Brian's comments. What Thony says I find objectionable is exactly what I find objectionable, and what Thony finds hard to imagine about a "standard 'possible worlds move'" I find hard to imagine as well. They don't seem analogous cases.

    When I wrote the comment about "no uncontroversial cases of relativism", I paused over it for a few moments, thinking that what a relativist *should* say is that relativity to possible worlds is a form of truth-relativity (and of course this is what John MacFarlane says in conversation). Indeed, I assigned as a paper assignment to my graduate seminar last semester the task of addressing whether Frege could incorporate modality into his over-all system, with the answer I was looking for being 'not easily, because that would involve relativizing truth to worlds'.

    But my own view is that propositional truth is at best *only* relative to worlds, that worlds are the only features of circumstances of evaluation (though I find myself tempted by the view, also alluded to by Brian, that maybe worlds aren't even features of circumstances of evaluation, but part of contents). Anyway, point taken.

    And Brian is right (though I think there is a typo in his message) that PVI is not a fictionalist about fictional characters, but is a fictionalist about properties ascribed to fictional characters within the fiction. He also cleverly put this in a form in which one could see the fictionalist strategy for avoiding apparently paradoxical arguments.

    So I was engaging in hyperbole when I said "there is not a single uncontroversial example of truth-relativity and fictionalism". In each case, there is a single relatively uncontroversial example.

  12. Will somebody please tell me what the "contextualist move" is supposed to be? I tried to define it in my previous post. If that attempt was correct, then I think you cannot maintain that the current attempts (i.e., past 20 years) to deal with semantic paradoxes, sorites paradoxes and skepticism by introducing context variables are examples of the "contextualist move". If that attempt was not correct but there is still some pattern of over-reaching on exhibit in contextualism, then I don't know why I shouldn't find it in the possible worlds idea as well.

    Please bear in mind that I first spoke up here in opposition to conservativism. I don't think one can effectively criticize any theory of anything on the grounds that it borrows an idea from somewhere else, whether that be the contextualism idea or the possible worlds idea. But one should not go around deploring the fever of new ideas without looking back to see whether any old and settled ideas are guilty of the same charge—and then in the very next post bemoan the resistance of the old guard to new ideas.

  13. I should think that it is controversial that what is expressed by an utterance of a sentence is only true relative to a world. To claim this is to ignore mood. Does an utterance of (2) express the same thing as an utterance of (1)?

    (1) The trains run on time in Germany.
    (2) The trains would run on time in Germany.

    It seems rather that the content of (1) is a claim about what holds across the candidates for the actual world among the current participants in the conversation, while (2) is a similar claim about some other (contextually salient) set of worlds. One reason to think that the contents of (1) and (2) are different is that entailment holds in neither direction between them.

Designed with WordPress