There is Some Truth in That

I’ve been thinking about the relation between reference and skepticism. A conversation I had with David Chalmers and John Hawthorne a couple of weeks ago has me worried about a principle I used to like.

The principle is this: it is sufficient for a possibility P to be a skeptical one, with respect to subject S’s belief b, that (a) b is false in P, and (b) S’s experience cannot distinguish the actual world from P.

So, for instance, my belief that I have a philosophy degree is threatened by these skeptical scenarios:

(1) I’m only dreaming that I have a philosophy degree
(2) I’m the victim of an evil demon who is tricking me into thinking I have a philosophy degree
(3) I have a fake philosophy degree that looks to me just like the real thing
etc.

I have always thought this principle was true — I’m even in print defending it, and arguing that it has interesting consequences about the nature of skepticism. But now I’m not sure.

Here’s a case: Angela is exploring an abandoned castle. She finds a room, R1, and decides to investigate it. She observes that the room has green walls; she also reads a plaque on the wall with the inscription: “This room is called Grammytoo.” Taking the relevant experience at face value in the usual way, Angela says to herself, “Grammytoo has green walls.”

Presumably, we should understand Angela as expressing, and knowing, the true belief that R1 has green walls.

Now the wrinkle. Preston is also exploring the castle, and he’s found another room, R2. R2 looks just like R1; in particular, R2 has green walls, and a plaque reading “This room is called Grammytoo.” He says to himself, “Grammytoo has green walls.” By parity of reasoning, we should suppose that he expresses, and knows, that R2 has green walls. “Grammytoo” in his mouth has a different referent than it does in hers.

So far, so good. But now the puzzle. Just as Angela knows what she expresses with “Grammytoo has green walls,” so too should she know what she expresses with “I am in Grammytoo.”

I think my principle above has the implication that this belief is skeptically threatened by the scenario where Angela is in R2. After all, even though in that scenario, she speaks truly when she says “I am in Grammytoo,” the content of her actual belief — that Angela is in R1 — is false in that scenario; but her evidence doesn’t distinguish that scenario from actuality. (The rooms look just the same from the inside.)

Is this the right result? I’m not sure. Should the possible case where Angela is in R2 — where she has a different true belief — put skeptical pressure against the belief she expresses with “I am in Grammytoo”?

Intuitions, anyone?

If you’re a philosopher who is interested in playing fantasy baseball, let me know. I’m trying to start a casual league.

Sorry not to have blogging much the past few weeks. I found myself very occupied by attention to some of life’s non-blogging demands. I’m pleased to say, however, that those attentions have culminated in some very exciting news.

I applied for a four-year research postdoctoral fellowship at the Arché Centre at St. Andrews, as part of their new project on intuitions and philosophical methodology. Last weekend I flew to Scotland for an interview; yesterday, I was offered one of the two posts.

It will be a gigantic transition for me — both moving to Scotland, and moving to a tiny town, will change things dramatically. I’m very excited about the research opportunities, though; the philosophical environment at Arché appears to be extremely exciting and intense. And the philosophers there seemed very quick, exciting, and friendly.

The post will begin September 1. That means it’s now time to finish up my dissertation. (This is news — I hadn’t planned on going out on the job market until next year, but I made an exception for this one job. I’m not at all sorry to skip the full job market process, at least for a few years!)

I’m not sure how this will affect blogging in the short term; maybe I’ll blog more because I’m thinking about dissertation material, or maybe I’ll blog less because I’m too busy to be bothered with the internet. We’ll see. I’ll certainly expect to be active here once I start up at Arché.

Thanks for the comments on the last post on this topic. Here is a limited thesis: the claim that ‘intuitions are important evidence in philosophy’ is ambiguous. Only the weaker sense is true. (I suspect, but won’t argue just now, that much of the experimentalist critique relies on the stronger sense.)

The assumption that intuitions serve as important evidence in traditional epistemology appears like a very innocuous one. (This, presumably, is why it has received so little explicit defense.) But it is worth examining. The most explicit extended defense of the claim that intuitions are used as important evidence in epistemology I am aware of is chapter 1 of Joel Pust’s book, Intuitions as Evidence. (Thanks, Jonathan!) Pust argues for the apparently-obvious claim in the obvious way: by pointing to myriad apparent examples of uses of intuitions as evidence in philosophical matters. He cites thought experiments about knowledge, justification, reference, moral rightness, justice, rationality, personal identity, and explanation — in each case, he points to an ‘intuitive’ judgment that plays a central role in the argument.

Let’s take a look at Pust’s first example intended to show that intuitions are used as crucial evidence:

Here is a case (derived from Lehrer …) from that massive literature:

Nogot’s Ford. Suppose your friend Nogot comes over to your house to show you the new Ford automobile he has just purchased. [standard Gettier story] … Do you know that a friend of yours owns a Ford?

Most philosophers take the fact that they have the intuition that S does not know that p in this case to show that S does not know that p. (p. 5, my emphasis)

Pust offers no defense of this sociological claim, apparently considering it obvious. It is certainly true that most philosophers take S not to know that p in this case; and it is also certainly true that most philosophers have the intuition that S does not know that p in this case. But I am inclined to doubt Pust’s claim that most philosophers take the fact about their own mental states to show that the fact about S is true. (Suppose they were asked to defend the judgment about S. The appropriate response would be to cite, for instance, the fact that S’s belief that p was derived from a falsehood. It is probably true that, upon sufficient questioning, they might exclaim, “I just have an intuition!” But this is just a way of ending the train of inquiry, not a serious attempt to explicate the evidence.)Very shortly after the passage I quoted above is another bit from Pust that can provide insight into the mainstream diagnosis of intuitions as evidence:

The analysis of justified belief proceeds in exactly the same fashion. A theory is proposed …, and it is tested by its ability to account for intuitive judgements regarding the justifiedness or unjustifiedness of particular actual and hypothetical beliefs. That this is so is recognized by many philosophers who reflect on their practice. For example, the epistemologist John Pollock claims that in epistemological analysis:

[O]ur basic data concerns what inferences we would or would not be permitted to make under various circumstances, real or imaginary. This data concerns individual cases and our task as epistemologists is to construct a general theory that accommodates it. (Pollock 1986, 182)

(Pust p. 5, emphasis in original)

This passage strikes me as remarkable. Pust cites Pollock as an epistemologist who reflects on his own practice, and recognizes the crucial role that Pust suggests intuitions play in it; yet the quotation Pust selects does not include the word ‘intuition’ or any of its cognates. Pollock is explicit: the basis data is the acceptability of inferences. Somehow, Pust takes away the moral that Pollock recognizes that the basic data are intuitions. What has happened?

The answer, I think, is that there is a crucial ambiguity in the suggestion that intuitions are evidence, and that Pust has allowed himself to slide between uses. Take a prototypical case in which intuitions are allegedly important evidence in traditional epistemology. I consider a thought experiment and I have the intuition that Jones doesn’t know that Smith owns a Ford. What do we mean when we say that this intuition is evidence for a philosophical claim — say, the claim that knowledge isn’t identical to justified true belief? We should distinguish between a strong and a weak reading of “the intuition that p is important evidence for q.”

Evidence is propositional. According to a strong reading, the intuition that p is evidence for q just in case the proposition that I have the intuition that p is important evidence for q. Pust clearly has in mind — at least sometimes — the strong reading. He is sometimes explicit, as when he says, as quoted above, that ‘most philosophers take the fact that they have the intuition that [q] to show that [q]’.

According to a weaker reading, the intuition that p is evidence for q means something like the intuited proposition — p — is evidence for q. It is on the weaker reading that the Pollock quote given by Pust plausibly lends credibility to the claim that Pollock treats intuitions are evidence in his epistemology. That such-and-such is permissible — an intuited proposition — is the starting point for Pollock’s theorizing.

(The weak reading is not an abuse of English. When Holmes investigates a crime scene, there is a perfectly acceptable sense in which he is relying crucially on his beliefs. But this is not to say that facts about his belief states play evidential roles in his theorizing. He reasons with propositions about the crime scene, not his own psychology. In a parallel way, we can express a truth by saying that “all we have to go on is our knowledge”; but “our knowledge” here refers to something like the set of known propositions — not any collection of our own mental states of knowing. It is on this model that we may understand the weak reading of “the intuition that p is important evidence for q”.)

Only this weak reading, I think, is plausibly true. There is no obvious place for psychologistic facts to play evidential roles in traditional epistemology. Consider a standard Gettier argument:

(1) Such-and-such is a possible story.
(2) Such-and-such involves a case of justified true belief that is not knowledge.
(3) Therefore, it is possible for there to be justified true belief without knowledge.

If I came to know (3) in this way — and like many epistemologists, I did — then, if someone asks me to cite my evidence for (3), then I will cite (1) and (2). I won’t mention any facts of the form, I have the intuition that…. Indeed, it’s hard to see how facts of that sort could play an important role in an argument anything like this one.

I was going to go on to anticipate some objections, but this is already long for a blog post; I’ll turn it over to the internet for objections, and maybe plan to write a follow-up post soon.

Last night I was perusing my copy of The Complete Works of Lewis Carroll, and came across this gem. It begins with an old, cute clever trick — but it ends with one I think engages with some deeper issues.

Which is better, a clock that is right only once a year, or a clock that is right twice every day? ‘The latter,’ you reply, ‘unquestionably.’ Very good, now attend.

I have two clocks: one doesn’t go at all, and the other loses a minute a day: which would you prefer? ‘The losing one,’ you answer, ‘without a doubt.’ Now observe: the one which loses a minute a day has to lose twelve hours, or seven hundred and twenty minutes before it is right again, consequently it is only right once in two years, whereas the other is evidently right as often as the time it points to comes round, which happens twice a day.

So you’ve contradicted yourself once.

Ah, but,’ you say, ‘what’s the use of its being right twice a day, if I can’t tell when the time comes?’ Why, suppose the clock points to eight o’clock, don’t you see that the clock is right at eight o’clock?

Consequently, when eight o’clock comes round your clock is right.
‘Yes, I see that,’ you reply.

Very good, then you’ve contradicted yourself twice: now get out of the difficulty as best you can, and don’t contradict yourself again if you can help it.

You might go on to ask, ‘How am I to know when eight o’clock does come? My clock will not tell me.’ Be patient: you know that when eight o’clock comes your clock is right, very good; then your rule is this: keep your eye fixed on your clock, and the very moment it is right it will be eight o’clock. ‘But—,’ you say.

There, that’ll do; the more you argue the farther you get from the point, so it will be as well to stop.

The more externalist one is about knowledge and justification, the closer this parody hits to home.

It is a commonplace that philosophers often use intuitions as evidence.

Here’s a really simple question: why do people think that’s true?

It’s an important premise in experimentalist critiques of philosophical methodology, but I haven’t found any kind of extended — or even brief but explicit — defense of it. Is there an obvious reason to think it’s true, or is there some discussion of it I can look to?

Eventually I’ll explain why I’m asking what may appear to be a really naive question.

Why is there typically a gap of many months between submission of a journal article and a decision? As I understand the process, the author submits her paper, then the editor invites a couple of reviewers, who agree to review the paper, then read the paper, then prepare a report and a recommendation for the editor. And then the editor makes a decision based primarily on the reports. Maybe it’s just my perspective mostly on the outside of this process, but I don’t see why that should typically take months.

I gather that the major hold-up comes between reviewers being assigned and reviewers submitting reports. But the actual act of reading and reviewing a paper doesn’t, in most cases, take so long that reviewers should need weeks and weeks to do it. I’ve reviewed papers a few times — it might take me an hour or two to read the paper, then half an hour to an hour to write up a report. (I’ll think about the paper in the background over a day or so in between reading and writing.) Here’s one hypothesis: I’m not spending anywhere near enough time reviewing. Somebody let me know if that’s the case. Here’s an alternative hypothesis: it takes weeks and weeks and weeks to get reviewers’ reports because reviewers have put the papers into their drawers and aren’t thinking about them. Reviewers take weeks and weeks to review because they’re procrastinating for weeks and weeks.

Here’s a system that appears to make sense to me. It’s probably hopelessly naive, but I’m curious as to why — or if it’s not, I’m curious why it isn’t the standard.

The author submits a paper. The editor notes the subject matter and writes to a couple of potential referees. “I have a journal submission in your area; would you be willing to referee it? Do you have time to review this paper this week?” If the reviewer says yes, then the editor has his report in a week, or perhaps a bit more if the editor have to start bugging them about it a week later. Probably sometimes, the reviewer will look at his schedule and realize that the best time to write the review in the next week is right now, and the report’s in in a day or two. If he says no, the editor turns to the next candidate.

One obvious consequence of this system is that lots more potential referees would decline, because they’re too busy this week. But those are the ones who would have taken forever to get the reports in anyway. If there are enough potential referees to invite, this doesn’t seem like a drawback. (If the editor can’t find somebody else qualified, he may extend another invitation with a different deadline. Perhaps better: he could have invited a “I’m too busy now, but could do it for you later if you don’t have somebody else first” response as an option in the first place.)

If the standard time to provide a report is one week, then, even allowing the time it takes to obtain referees and consolidate reports into decisions, I would think that the standard time from submission to decision could be somewhere in the ballpark of two to three weeks.

I’m dreaming, I’m sure. Tell me why this idea is unrealistic?

Steve Stich thinks you’d have to be pretty weird to value knowledge intrinsically. He writes:

Other languages and other cultures certainly could and probably do invoke concepts of cognitive evaluation that are significantly different from our own, just as they invoke different conceptions of etiquette. For many people — certainly for me — the fact that a cognitive process is sanctioned by the venerable standards embedded in our language of epistemic evaluation, or that it is sanctioned by the equally venerable standards embedded in some quite different language, is no more reason to value it than the fact that it is sanctioend by the standards of a religious tradition or an ancient text, unless, of course, it can be shown that those standards correlate with something more generally valued or obviously valuable. Unless one is inclined toward chauvinism or xenophobia in matters epistemic, it is hard to see why one would much care that a cognitive process one was thinking of invoking (or renouncing) accords with the set of evaluative notions that prevail in the society into which one happened to be born. (The Frangmentation of Reason, p. 94)

Grant the empirical claims for the purpose of argument. I agree that it’d be pretty weird to value having a cognitive process that “accords with the set of evaluative notions that prevail in the society into which one happened to be born.” But I insist that it’s not at all weird to value having knowledge. The property of having knowledge isn’t the same property as the property of having a cognitive process that accords with the set of evaluative notions that prevail in the society into which one happened to be born. These properties have different modal extensions.

(1) I’ve just received word that my paper, “Dreaming and Imagination,” has been accepted for publication at Mind & Language. That’s my defensive of the imagination model of dreaming — something I’ve been thinking about for practically my entire graduate career. Thanks very much to the people who have helped me develop these ideas via this and other blogs.

(2) I’ve applied for the Arche postdoctoral fellowships in intuitions and methodology. They’re four-year research positions, starting this September. I hadn’t been planning to go on the job market until next year, but this opportunity struck me as exciting enough to make a provisional change of plans worthwhile. Maybe I’ll be finishing my Ph.D. and moving to Scotland this summer!

There is some limited data that suggests that people in different cultural groups have systematically different epistemic intuitions. Many Americans of East Asian descent, for instance, appear likely to react to Gettier cases by affirming that the subject ‘really knows’ the true, justifiably believed proposition. Stephen Stich and other ‘experimental philosophers’ think that this sort of systematic disagreement with standard Western philosophical intuitions is devastating for traditional philosophical practices.

One possible defense against the challenge is to suggest that these subjects mean something different by ‘knowledge’ than do we mainstream epistemologists. In a fascinating exchange in Steve Stich’s seminar this morning, Ernie Sosa tried to motivate that line against challenges from Steve and Jonathan Weinberg. Steve and Jonathan were pushing a burden of proof move against Ernie: “sure, it’s possible that there’s a kind of ambiguity at work, but for this to be an adequate defense, it must be shown that there is such an ambiguity — so you need to tell me more about the individuation conditions for concept types.”

This move struck me as surprising. Surely it’s possible for two different groups of people to use the word ‘knowledge’ with two distinct but largely overlapping extensions. How would we recognize such a situation? Presumably, by finding actual or hypothetical cases in which, even after careful reflection and consideration, and mutual acquaintance with the more basic features of the situation, they disagreed whether the term ‘knowledge’ applied.

What better evidence could one ever hope to find that my word ‘knowledge’ doesn’t mean the same thing as yours?