Saturday, February 25, 2006
Stanley and 'catch-all solutions'
How is this for a pointless and annoying failure of transitivity? To get a state issued ID card, the only ID I need is my passport. To buy alcohol at H.E.B., the only ID I need is a state issued ID card. Yet my passport isn't accepted as sufficient ID to buy alcohol at H.E.B. Sigh.
I really liked Stanley's Knowledge and Practical Interests a lot, but I've been trying to figure out what he's up to in the final chapter on philosophical paradox. Stanley is generally worried by standard moves to tackle difficult philosophical issues; his examples are going fictionalist, contextualist, relativist, and attributing the seeming unacceptability of certain statements to pragmatic rather than semantic grounds. It's not that he thinks that it's always inappropriate to propose such a solution to a given problem, it's just any such proposal should appeal to special features of the target objects or concepts that explain why they in particular should be ameniable to such a treatment; any such proposal that appeals just to the (purported) availability of, say, a contextualist or fictionalist resolution is to be regarding with suspicion. Why such suspicion? Stanley seems to have an intuition (which I share to some extent) that philosophical solutions should be discriminating; we shouldn't expect solutions to seemingly disparate and deep philosophical problems - the Sorites and scepticism, for example - to be given by a standard piece of philosophical machinery that can be applied with equal success and plausibility (or lack thereof) to almost any other problem. (For Stanley's own statement of his dissatisfaction with 'catch-all' solutions, plus some discussion from myself and others, visit here.)
Now, in the final chapter of his book Stanley is keen to show that he hasn't appealed to such a catch-all method in accounting for certain intuitions we have regarding the dependence of correctness of knowledge ascriptions on features of either the attributer or the subject's circumstances. (I'm leaving that deliberately vague for now). Stanley's contention is that an account of knowledge on which whether a subject knows a given proposition can vary (even keeping their evidence for that proposition fixed) depending on their practical interests explains why we might be strongly and systematically inclined to attribute knowledge of a given proposition to a subject if the consequences of them being wrong (given their practical interests) are minor, and yet give a different verdict regarding a subject with the same evidence for whom being wrong would be disastrous. Contextualism about knowledge also seeks to respect these intuitions, but Stanley thinks that it incurs costs which his position neatly avoids.
The danger that Stanley has given a catch-all solution arises because Stanley's position stands to contextualism about knowledge just as Delia Graff's interest relative account of vague expressions stands to contextualist solutions to the Sorites. According to Graff (Shifting Sands, 2000), contra the contextualist, vague predicates like 'heap' and 'bald' express the same property throughout their respective suitably constructed Sorites series. Yet changes in the practical interests of a subject making judgements with such predicates determine shifts in their extensions (shifts not attributable to changes in the relevant comparison class, etc). A similar invariantist position stands opposed to contextualist treatments of the Liar paradox (though Stanley restricts his attention to the Sorites). The worry, then, is that interest-relative invariantism (IRI) is not one jot more discriminating about which philosophical problems are amenable to the kind of treatment it offers than contextualism.
Stanley takes one of the lessons of his cool little Analysis paper from 2003 criticising both contextualist and Graff's solutions to the Sorites to be that IRI isn't a catch-all solution. The argument is just that his criticisms of Graff's view show that the Sorites isn't in fact amenable to an IRI treatment. But to finally get to my worry, I don't see how problems with Graff's account in any way show that IRI isn't a catch-all solution. Stanley takes contextualism to be one of the paradigms of such a solution, but he doesn't think there's a good contextualist solution to the Sorites either in the offing either. So I'm just not clear on how he's supposed to have shown that IRI is appropriately more discriminating than contextualism.
Update: I was working from memory yesterday, and so when I got home I checked Stanley’s book on a hunch that I was probably being unfair to him. And that’s right; Stanley does spend a little time (177-8) pointing out specific features of vague expressions that make an IRI account inapplicable which are not present in the case of knowledge attributions. But I take it this doesn’t invalidate what I said above. We want specific features of knowledge attributions that explain why they are amenable to an IRI treatment; what Stanley offers in chapter 8 are specific features of statements featuring vague predicates that make such a treatment unavailable there. In the absence of reasons for holding that these features are also present in most other candidates for an IRI treatment, it’s still hard to see on what grounds we should conclude that IRI might not be a catch-all solution after all.
I really liked Stanley's Knowledge and Practical Interests a lot, but I've been trying to figure out what he's up to in the final chapter on philosophical paradox. Stanley is generally worried by standard moves to tackle difficult philosophical issues; his examples are going fictionalist, contextualist, relativist, and attributing the seeming unacceptability of certain statements to pragmatic rather than semantic grounds. It's not that he thinks that it's always inappropriate to propose such a solution to a given problem, it's just any such proposal should appeal to special features of the target objects or concepts that explain why they in particular should be ameniable to such a treatment; any such proposal that appeals just to the (purported) availability of, say, a contextualist or fictionalist resolution is to be regarding with suspicion. Why such suspicion? Stanley seems to have an intuition (which I share to some extent) that philosophical solutions should be discriminating; we shouldn't expect solutions to seemingly disparate and deep philosophical problems - the Sorites and scepticism, for example - to be given by a standard piece of philosophical machinery that can be applied with equal success and plausibility (or lack thereof) to almost any other problem. (For Stanley's own statement of his dissatisfaction with 'catch-all' solutions, plus some discussion from myself and others, visit here.)
Now, in the final chapter of his book Stanley is keen to show that he hasn't appealed to such a catch-all method in accounting for certain intuitions we have regarding the dependence of correctness of knowledge ascriptions on features of either the attributer or the subject's circumstances. (I'm leaving that deliberately vague for now). Stanley's contention is that an account of knowledge on which whether a subject knows a given proposition can vary (even keeping their evidence for that proposition fixed) depending on their practical interests explains why we might be strongly and systematically inclined to attribute knowledge of a given proposition to a subject if the consequences of them being wrong (given their practical interests) are minor, and yet give a different verdict regarding a subject with the same evidence for whom being wrong would be disastrous. Contextualism about knowledge also seeks to respect these intuitions, but Stanley thinks that it incurs costs which his position neatly avoids.
The danger that Stanley has given a catch-all solution arises because Stanley's position stands to contextualism about knowledge just as Delia Graff's interest relative account of vague expressions stands to contextualist solutions to the Sorites. According to Graff (Shifting Sands, 2000), contra the contextualist, vague predicates like 'heap' and 'bald' express the same property throughout their respective suitably constructed Sorites series. Yet changes in the practical interests of a subject making judgements with such predicates determine shifts in their extensions (shifts not attributable to changes in the relevant comparison class, etc). A similar invariantist position stands opposed to contextualist treatments of the Liar paradox (though Stanley restricts his attention to the Sorites). The worry, then, is that interest-relative invariantism (IRI) is not one jot more discriminating about which philosophical problems are amenable to the kind of treatment it offers than contextualism.
Stanley takes one of the lessons of his cool little Analysis paper from 2003 criticising both contextualist and Graff's solutions to the Sorites to be that IRI isn't a catch-all solution. The argument is just that his criticisms of Graff's view show that the Sorites isn't in fact amenable to an IRI treatment. But to finally get to my worry, I don't see how problems with Graff's account in any way show that IRI isn't a catch-all solution. Stanley takes contextualism to be one of the paradigms of such a solution, but he doesn't think there's a good contextualist solution to the Sorites either in the offing either. So I'm just not clear on how he's supposed to have shown that IRI is appropriately more discriminating than contextualism.
Update: I was working from memory yesterday, and so when I got home I checked Stanley’s book on a hunch that I was probably being unfair to him. And that’s right; Stanley does spend a little time (177-8) pointing out specific features of vague expressions that make an IRI account inapplicable which are not present in the case of knowledge attributions. But I take it this doesn’t invalidate what I said above. We want specific features of knowledge attributions that explain why they are amenable to an IRI treatment; what Stanley offers in chapter 8 are specific features of statements featuring vague predicates that make such a treatment unavailable there. In the absence of reasons for holding that these features are also present in most other candidates for an IRI treatment, it’s still hard to see on what grounds we should conclude that IRI might not be a catch-all solution after all.
Labels: Epistemology, Jason Stanley, Vagueness
# posted by Aidan @ 5:19 PM 
Comments:
<< Home
Aidan,
Here is the explanation. IRI about knowledge isn't a 'method', like contextualism or relativism. IRI about knowledge is simply the claim that the nature of knowledge is more complicated than epistemologists have realized. IRI about knowledge is therefore simply a claim about what it is to know something. So I don't view it as a method, in the way that appealing to context-sensitivity or relative truth to evade contradiction supplies one with a general method.
Maybe one way to realize this is to think about how utterly hopeless it would be to try to appeal to the interest-relativity of truth to evade the Liar Paradox (I have some remarks about this in the book). In contrast, contextualism and relative truth can be mechanically applied anywhere there is the appearance of disagreement.
Here is the explanation. IRI about knowledge isn't a 'method', like contextualism or relativism. IRI about knowledge is simply the claim that the nature of knowledge is more complicated than epistemologists have realized. IRI about knowledge is therefore simply a claim about what it is to know something. So I don't view it as a method, in the way that appealing to context-sensitivity or relative truth to evade contradiction supplies one with a general method.
Maybe one way to realize this is to think about how utterly hopeless it would be to try to appeal to the interest-relativity of truth to evade the Liar Paradox (I have some remarks about this in the book). In contrast, contextualism and relative truth can be mechanically applied anywhere there is the appearance of disagreement.
# posted by 
: 4/02/2006 10:57 PM
: 4/02/2006 10:57 PM
Thanks, I think the point is well taken with respect to the Liar. I'm not sympathetic to Delia's IRI treatment of vagueness, but I guess the point I was trying to make in the post is that I couldn't quite get in focus the relevant contrast you wanted to make with contextualism, given that you were concerned to argue against both. But I do remember scattered comments on the Liar in the book, and perhaps the hopelessness of an IRI treatment there does serve to suggest there isn't a handle we can crank here. I'll give it some more thought.
I'm really looking forward to meeting you next week - we should have a lot to talk about.
Post a Comment
I'm really looking forward to meeting you next week - we should have a lot to talk about.
<< Home
