WHO WOULDA THOUGHT?

It's scary when Michael Moore writes one of the most sensible things I've read about American politics in weeks.

WHAT ARE ARGUMENTS FROM ERROR ARGUMENTS FOR?

I've been thinking about arguments from error for quite some time now as they figure prominently in the arguments for justificatory internalism. Right now I'm putting the finishing touches on a paper of mine ('The Process and the Property of Justification') and have been thinking about a problem for any view on which S's being personally justified in believing something entails that S's belief being justified. The problem is structurally similar to the problem that arises for reliabilism in the guise of the new evil demon problem.

Let us suppose that:

(A1) Personal justification suffices for doxastic justification.
(A2) Doxastic justification suffices for propositional justification.

It is my view that the intuition that is generated by the NED thought experiment and the judgment concerning the beliefs of the deceived are personally justified--we have the intuition on the basis of the fact that we think that the dupe's epistemic performance is impeccable and conclude that the belief is justified. If I'm right, this mistake is similar to a mistake J.J. Thomson thinks underlies the arguments for thinking that killing is worse than letting die--we wrongly move from the observation that in the killing cases the agent is bad to the mistaken conclusion that there is some set of considerations that show the action to be impermissible. Anyway, set this aside for now.

Suppose that we accept (A1) and (A2) while also maintaining a view of propositional justification with the following feature: for a belief that p to be justified, there must be a justification for the belief and this depends on there belief's meeting some condition C that is (i) non-luminous (ii) such that S could find herself in a position such that it would be irrational for her to refrain from judging that her belief that p meets C when it doesn't.

In a case where all these details would be met, we would judge, as we judge of the dupe in the demonic scenario, that she'd be justified in believing p even though her belief that her belief meets C is false. Thus, we'd either have to revise the account of justification or deny (A1) or (A2). Note that it doesn't matter whether C is a belief's being produced by a reliable process or whether it is something more internal. Unless the condition is so internal that it can't meet (i) and (ii), the difficulty will arise.

This suggests that the problem that is supposed to be a problem for the reliabilist is in fact a problem that arises for any extant theory of justification as there is no such theory that specifies a necessary condition on justification C that is luminous and such that if one's belief fails to meet this condition it is impossible to find oneself in a position where it would be irrational given one's commitments to refrain from judging (wrongly) that one's belief meets C.

UPDATE
Why Kerry is the one to fight the war on terror.

OH LAURA ...

Here's a gem. Time interviews the first lady:

TIME
And did you have a take on this gay-marriage question?

BUSH
Well, I think it's a debate. People want to be able to debate the issue, and that's exactly what the call for a constitutional amendment does. It opens the debate up. The people of the United States didn't really want the Massachusetts Supreme Court or the San Francisco mayor to make the choice for them. And we're seeing a debate on it. And I think that's good.

TIME
Did you have a take on the amendment yourself?

BUSH
I also think there should be a debate on the issue. People want to be able to talk about it—and come to terms with it, if that's what people decide.

TIME
Right, but are you of an open mind about the amendment?

BUSH
Sure

I guess I had it all wrong. I had thought amending the constitution is a way of ending discussion but Bush is really the pro-discussion candidate.

DEFEAT AND JUSTIFICATION

In Cohen (1984), he points out that a certain kind of justification internalist cannot make sense of both undermining and overriding defeat. The problem is that we define the notion of undermining defeat in terms of evidence which, when acquired, can be combined with one's present evidence in such a way that the evidence is retained but the justification for a certain belief is lost. For example, if you taste some OJ and it tastes funny, you can't justifiably believe that there is something wrong with the OJ when you remember that you've just brushed your teeth. The justificatory internalists who claim that there being something or other that makes your belief probably correct isn't a necessary condition on justification can't explain why justification would be lost when one acquires information to the effect that given one's present position there is nothing that makes it objectively likely that one's beliefs are correct.

I haven't seen much in response to this problem apart from the response Cohen mentions which is that it is just a brute fact about justification that this is so. Does anyone know of an article in which a justificatory internalist takes up Cohen's challenge?

GEACH HAS ANOTHER PROBLEM

I'm not interested in the problem that bears his name, I'm interested in a lesser known problem that should be less unknown. I don't know whether this is standard practice (it seems to be from the few sources I've checked), but he is railing against the use of propositional functions such as 'It ought to be the case that...' that yield propositions when propositions are given as arguments. Geach comments:

There is no manifest inconsistency in saying: 'I do not allow that John ought to beat up Tom; Tom is no doubt a rascal, but it is not John's business to beat him up, John has no right to. But Tom ought to get beaten up by John; he thoroughly deserves it, he asked for it, serve him right' Nor do I see how anyone could show this moral attitude to be latently inconsistent.

He offers this as evidence for thinking that 'ought' operates on predicables to form predicables and 'it is a muddle' to ask of a state of affairs, equally describable as 'John beating up Tom' and 'Tom being beat up by John', whether it ought to be. The argument, which I find impressive, is that in SDL ('standard deontic logic') where we represent 'ought' as an operator 'It ought to be the case that', the attitude above (which may well be false for substantive reasons) is a manifest contradiction but this should not turn out to be a contradiction.

Has Geach's observation any significance when it comes to characterizing the relations between personal, doxastic, and propositional justification? I leave that as an open question for now because I don't have the article I need at hand. Any thoughts? Is Geach's moral sentiment hopelessly inconsistent? Are there better examples of the phenomenon he has in mind?

ACCESS, PREROGATIVE, AND REASONS

An agent's reasons are the things the agent took to weigh in favor of some decision or belief, let us suppose. It is tempting to say that the agent's reasons belonging to the agent entails and that this is best accounted for by internalism. I disagree. The accessibility of reasons can be a matter of:

ACCESS: An agent knows, or can know upon reflection, whether he has reasons and what they are.
PREROGATIVE: An agent knows, or can know, what his reasons are.

These conflict and that they conflict is important. When an agent knows p (where 'p' is a proposition about the external world) it is tempting to say that p can be the reason A X's and that in such a state, A satisfies both access and prerogative. But, when we turn to the 'bad' case, access and prerogative come apart. Why? Because if you ask the agent why he X'd, he'd say 'p'. He wouldn't justify his behavior by reference to this unless he believed p. However, according to the internalist, he's wrong about what his reason is--it his belief that p, not p itself.

Internalists tend to say that if we give up access, we end up harshly evaluating agents, saying they weren't justified or rational in doing what they did just because they had a false belief. Ah, but internalists who say this deny prerogative and thereby are paternalistic. I prefer tough love to paternalism and say that an agent who is justifed in X'ing or believing p satisfied prerogative without having to satisfy access as access means that a justified agent knows that he has a justification, or can know this upon reflection.

This comes out in the case of error were a third party tries justifying the agent while maintaining that there is no justification for what the agent did or believed. The defender of access would, I think, have to say that since the agent who is justified could know on reflection that this is so and come to know there is a justification for believing what she does. This underwrites what I've labeled the 'passive aggressive' fallacy in honor of a number of right wing bigots who reason from 'I'm justified in believing that SSM is wrong' to 'There is a justification for believing this'. On this version of access internalism, I can virtuously move from a justified belief that I'm justified in believing to the claim that it is the case that there is a justification for believing what I do. I think we should distinguish two JJ principles, one that is true, one that is not, one that I accept, one that the access internalist accepts:

(JJ1) If S is PJ in believing her belief that p is DJ --> S's belief that p is DJ.

(JJ2) If S is PJ in believing her belief that p is DJ --> S is PJ in believing p.

That (JJ1) could be false when (JJ2) isn't strikes me as important. An inference that conforms to (JJ1) is an instance of the passive aggressive fallacy. If I am justified in believing (mistakenly) that my belief that p satisfies the conditions on justification, it seems the antecedent of (JJ1) and (JJ2) is true but the consequent isn't in (JJ1) but is in (JJ2). SO assuming that DJ requires that there is a justification for believing p, I think the way to deal with this is to deny ACCESS, go with PREROGATIVE, and let the agent's decide what their reasons are and whether they'd have them if their beliefs turned out to be false.

JUSTIFICATION, JUSTIFIEDNESS, AND ASSERTION ... AND A SURVEY

Yesterday I distributed a survey in my Phil 106 (this is the 'sex and violence' course of the department as I've heard it described as opposed to the 101 which I've heard described as 'the class where they try to trick you into thinking that God doesn't exist'). Some interesting results. When asked whether the world would be a better place if people stopped eating meat 100% of the students said 'No'. When it came to the war in Iraq, when asked whether we should have gone 70% of them said 'Yes' (a slightly smaller percentage said that the costs of this war will outweigh the benefits). Nearly 60% of them said that same sex couples should be allowed to marry which is surprising in a state where 76% of the people voted in favor of the defense of marriage act.

I've just spent a good hour or so arguing with Gibbons about justification, in particular, whether claims such as (1) or (2) could be true and what follows:

(1) John is justified in X-ing but there is no justification for X-ing.
(2) John is rational in X-ing but there isn't really any reason for John to X.

Examples might help. I have in my hand a bottle filled with petrol marked 'Gin' and know that John wants a little gin, would prefer not being poisoned with petrol than ingesting some gin, and believes that the bottle contains unadulterated gin. I think this is a case in which (2) is true where 'X-ing' is 'drinking what is in the bottle'. I think this is a case in which (1) is true where 'X-ing' is believing that he should drink what is in the bottle.

My initial argument for (1) was that if John asked me either 'Do I have a reason to drink what is in the bottle?' or 'Is there any reason to drink what is in the bottle?', I'd be lying if I said 'Yes'. However, if some third party asked whether John was rational in drinking if he did, I could say 'Yes' as well. In fact, I could quickly qualify by asserting (2).

A complication must be noted. If John asks 'Is there any reason for me not to drink?' or 'Do I have a reason not to drink?', it looks like I should say 'Yes'. That is questionable evidence since the saying might alter John's reasons. If I say 'No', however, it seems to me that I'm lying.

Another more interesting complication must be noted. Could it be that if John addressed someone else who believed what he did the right answers to his questions would differ? My view is that they wouldn't, in which case this means that the fact that John as the reasons he does is determined by the facts about his situation and his preferences, not facts about his beliefs and which beliefs of his he's justified in having. One hypothesis we discussed is whether the truth of my assertion about John's reasons might depend on my evidence or beliefs rather than his. The semantic intuition I have is that if some third party who believes all that John does says that there isn't a reason to refrain from drinking, he's just wrong--the bottle has some petrol in it.

IN DEFENSE OF DOUBLE BILLING

I've just sent a letter to Randy Cohen, (aka 'The Ethicist') at the New York Times regarding his discussion of Double Billing.

Dear Mr. Cohen,

While generally I find your verdicts agreeable, I must take issue with your response to the graduate student from New York who wanted to know whether it would be permissible for her to bill her University for two hours were she to simultaneously hold an office hour and grade. There is a perfectly good sense in which the University would receive two hours worth of work from her were she to grade papers while waiting dutifully in her office for students who happened never to arrive so I fail to see why she shouldn’t be paid accordingly. Were a student to arrive, contrary to fact, she would cease grading and attend to the student. This is not something she’d be obligated to do were she spending that time working for the University as a grader. If the University failed to take this into consideration, they’d be paying her as a mere grader, something that she is not. As was made explicit in the letter, she was not being paid to meet with students; rather, she was being paid to be available to meet with students. If she satisfies two obligations, both of which consume an hour, the fact that there was one hour during which she met both obligations shouldn’t exempt the University from paying her for doing so even if the University is paying her an hourly wage.

Sincerely,

Me

A PROUDFUL DAY FOR ME AND MY PEOPLE

As someone who was adopted into a family with some Lithuanian heritage I couldn't feel any more proudful than I do now having heard that the boys from Lithuania defeated the 'dream team'.

MORE ON BOGHOSSIAN AND INCOMPATIBILISM

Boghossian argues for incompatibilism appealing to 'uncontestable' assumptions about non-intrinsic properties (relational or extrinsic properties) in 'Externalism and Inference', arguing:

(1) You cannot know that an object has a non-intrinsic property by inspection.
(2) According to the content externalist, content properties are non-intrinsic.
(C) You cannot know the content properties of your thoughts by inspection if the content externalist is right.

In a previous post, I argued that (1) is mistaken. I focused on our knowing right from left to show that (1) is false. John Gibbons pointed out that knowing that something is nearby seemed to be another counterexample to (1). What about (2)?

I had granted (2) initially, the thought being that the Twin thought experiments are ones in which intrinsic duplicates differ in that one, not the other, is such as to believe that water is wet. If we employ this test, then (2) is easily enough established:

NI Test: If x and y are intrinsic duplicates such that x is F but y is not F, then F is a non-intrinsic property of x.

But then I naively thought passing this test meant that F is non-intrinsic so that for all z such that z is F, F is a non-intrinsic property of z. I less naively am assuming that saying that F is a non-intrinsic property means that there is nothing that has F as an intrinsic property. I just realized today, thinking about a conversation with John from the fbc, that something is terribly amiss.

I had said quite recently that the conditions under which one could grasp a concept such as water might be highly disjunctive, but there is no disjunct such that we could say that someone could know that this disjuct was satisfied without knowing something substantive about the external world. But Burge in 'Other Bodies' points out that someone could, in principle, introduce the term 'water' in a waterless world that picked out only collections of H2O and would thereby grasp our concept of water. Such a super-scientist could have the concept of water even if there was nothing else in the world. But then by the following test of intrinsicness:

I test: If x is the lone object and x is such that x is F, then F is an intrinsic property of x.

No property can be intrinsic and non-intrinsic. Now, I doubt anyone would employ both these tests within one framework, but the problem is that these tests and the ancillary assumptions I've mentioned are reasonable. Maybe Burge was wrong in saying that someone could come up with the concept of water simply through intellectual effort and imagination, but that doesn't strike me as the right think to say here.

I suspect that the assumption that I think will get us in trouble is this one:

If there is an x such that x is F and F is a non-intrinsic property of x, then for all y such that y is F, F is a non-intrinsic property of y.

Denying this would allow us to deny (2) as that is a claim about content properties as such that is based on observations about pairs of subjects that are consistent with the possibility of a solitary superscientiest having thoughts with these content properties. However, it doesn't undermine (2) when reformulated and applied to ordinary schmoes. As I'm thoroughly convinced that (1) is false and that Boghossian's discussion trades on the distinction between properties of thoughts and properties of persons, I'm not too terribly worried about it. But it is interesting and I wonder what more could be said for or against (2).

DO ANY NEWBIES HAVE THEIR EARS ON?

I don't know whether any of you new philosophy grad students have your ears on, but if you are in Lincoln and are having any trouble finding your way around town or school, let me know. Chances are that most of us won't be around Oldfather next week, but if any of you want to catch a drink or anything, send me an email at cmlittlejohn@yahoo.com.

RELIABILISM AND THE APRIORI (CONT.)

In the post prior to the previous, I explained why I thought that there was a problem with the Brueckner/Brown inspired response to those content externalists who think that we can acquire inferential apriori(sh) knowledge of the external world. The standard argument for the possibility of KWBKM, recall, goes as follows.

(1) Audrey knows apriori she's thinking that water is wet (Content externalism & Privileged Access).
(2) Audrey knows apriori if (1), there is water (Content Externalism).
(C) Audrey could come to know apriori that there is water.

Anyway, the Brown/Brueckner inspired response which is (shamelessly) almost never discussed in the literature is that (2) is false because the content externalist is committed only (at most) to the fact that it is impossible for someone to grasp the concept of water in a waterless world, not that this is apriori. McKinsey and Boghossian object to this, but their objections don't work (trust me).

Here's a different response. If what we are trying to show is that the content externalist isn't committed to the apriori knowability of (C) we run into a difficulty. If we grant (1), we can't deny that (C) can be something you can be justified apriori in believing unless we can deny that the mere fact that drawing (C)-type conclusions from (1)-type premises is reliable would be enough to ensure that the justification Audrey has for believing she thinks that water is wet is transmitted to her belief that there is water. It seems to me that the natural view for a reliabilist to take on inferential apriori justification is simply that p is apriori justified if p is non-inferentially justified apriori or inferentially justified on the basis of something non-inferentially apriori justified where there is a reliable inferential principle linking the two thoughts. Brown won't want her argument against (C) to depend on rejecting exteranlist accounts of justification.

Here's how I'll try to deal with this problem. We have to be careful in specifying the type of the belief generating process. What I hope to show is that no process could generate knowledge that there is water.

To begin, as we are trying to show that no process could generate knowledge of (C), we have to split cases between those who already know (C) and those who don't. I'm tempted to split this between those who already believe that there is water and those who don't (this will take a lengthy explanation and I just don't feel like dealing with it now). If we are dealing with people who are trying to settle the question 'Is there water?' by means of this process, then (1) can't be read as saying that Audrey knows she believes that water is wet, it has to be read as she suspects, wonders, wants, etc. Then the argument is from a premise about her mind where the relevant mental state isn't one which if it constituted knowledge would ensure that 'there is water' is true. Now the response to this reliabilist response is that under these specific conditions, the inference won't be a reliable one. For under these conditions, the inference would be from 'I have an attitude with concept C' to 'There is something that answers to C'. This is the generall strategy I hope to develop this afternoon. Fingers crossed.

THE PRESIDENT HAS A MESSAGE FOR YOU...

(here).

RELIABILISM AND THE APRIORI

A few content externalists (Brewer, Sawyer, Warfield) have maintained that one could come to know the conclusion of the following argument apriori if content externalism is true:

(1) I'm thinking that water is wet.
(2) If I'm thinking that water is wet, then there is water.
(C) There is water.

I think this is a mistake and for a while explained this much in the way that Brueckner (1992) and Brown (2004) have. Simply put, it is a mistake to think that (2) is knowable apriori. The argument for thinking that (2) is apriori is that we know content externalism on the basis of thought experiments including ones that pertain to our concept of water, but this rationale fails terribly.

What I just realized last night is that this strategy for undermining this troubling thesis is problematic:

KWBKM : If CE, it is possible for someone to come to know the world apriori knowing what their mind is like.

The standard criticism is that since CE isn't knowable apriori, (2) isn't knowable apriori, and thus (C), if known would be known via inference from (1) and (2), thus it wouldn't be known apriori.

The standard criticism, as well as the standard defense, of KWBKM is problematic.

Let's say that we are reliabilists about knowledge including apriori knowledge. On such a view, if p is known inferentially, p can be known apriori provided that the basis for q is apriori. Why can't the defender of KWBKM simply say that the basis for (C) is (1) all by itself? An inference from a (1)-type premise to a (C)-type conclusion is reliable provided that the scope of content externalism is reasonably broad (i.e., it applies to most of our kind concepts). I'm worried that a defender of KWBKM might say that (2) needn't be known apriori. Consider the parallel case of logical knowledge. Should the reliabilist say that if S comes to know q apriori from apriori knowledge of p and of p --> q S knows (i) modus ponens is reliable; (ii) modus ponens is an exceptionless rule. I don't think so. It would seem that the reliabilist would want to say that (i) simply has to be true, not known to be true, and (ii) could turn out to be false consistent with its being the case that q is known apriori.

I still think KWBKM is false, but that explanation is a little harder than I initially thought.

EVERYBODY'S GOT SOMETHING TO HIDE EXCEPT FOR ME AND JON SUTTON

A couple of things I've read have completely changed the way I think about epistemic justification and what I plan on defending in my dissertation. First, there is Jonathan Sutton's excellent forthcoming paper 'Stick to What You Know'. I came across it thanks to a comment of his on Certain Doubts and I've been convinced of something that I've been taking some flak for around these parts:

(J --> T) If S's belief that p is justified, p is true.

As everyone but Sutton and I deny this, everyone else is wrong about justification. He argues for this in arguing for his view that J is K (where the 'is' is the is of necessary biconditionality). I happen to disagree with this, but this disagreement isn't what concerns me here. He says:

Sutton: S's belief that p is justified iff S knows p.
Me: S's belief that p is justified iff S's belief that p is true.

I've argued that the motivation for Sutton's view is lacking because justification requires nothing more than truth. By my lights, his view would be like a view on which the aim of right action is bringing about the best outcome for the right reasons. It contains too much. Set that aside, let's just focus on J --> T. [If you want to know why this is unmotivated, it is because I think three things are true. (i) The aim of belief is T not K (Sutton disagrees and I'm working on the case for this); (ii) So long as you fulfill this aim there is a justification for you to believe; (iii) Doxastic justification is reduced to propositional justification (i.e., S's belief that p is justified iff There is a justification for believing p). The important disagreement here is not between me and Sutton, it is us vs. the world. Read his paper if you want a conversion experience. If you want another conversion experience that will move you towards my mere truth view, read Sorenson's 'Unknowable Obligations' and then read some bad stuff on expected outcome and rule consequentialism (It isn't bad simply because it is a rule or an expected outcome view--I'm saying read the bad stuff because if you are looking for a conversion experience, you should avoid the thought provoking stuff).

The stick that I've been getting beat with is this one. Suppose that all your evidence points against p but p is true. What should you believe? I say p. They say ~p. What are you justified in believing? I say ~p. They do too. On my view personal and doxastic justification come apart. On their view, they don't. Now, my argument against their view (which of course I find convincing) is this.

Doxastic justification implies propositional justification. That is 'S's belief that p is justified' entails 'There is a justification to believe p'. They agree and say that personal justification implies doxastic justification. Here's where we disagree. Here's why I win.

Let's suppose that you can't directly satisfy the directive believe what is true. You should satisfy it so you should take the means to satisfying it. Let's say you do this by making sure your beliefs cohere. By believing p you can satisfy the one goal (the coherence goal) while failing to fulfill the fundamental goal (the truth goal). Question: Should you be like this? Question: Is there a (n all things considered) justification for being like this? Question: Ought you be like this?

The answer to all three questions is the same, but if you accept (a) the inference from personal to doxastic to propositional justification and (b) that there are some cases in which S is justified in believing p when ~p because of coherence, you have to answer 'yes'. So now I have a stick to beat you with. There is some sense in which things aren't as they should be if you believe the false.

I'm interested in determining whether the following claim is true and if not, what sort of exceptions there are to it. The claim: There is no X-type justification for being such that you fail to fulfill an X-type obligation. A second and related claim: If by Y-ing you'd fulfill all the relevant X-type obligations, there is an X-type justification for Y-ing.

LOST IN TRANSLATION

I've been trying to pin down the details of Brewer's argument for:

(A) A person's capacity to make determinate reference to certain objects and kinds in belief depends on having demonstrative based knowledge about them.

His argument assumes the form of a reductio on two assumptions:

(1) S believes p because her embedding makes it the case that he believes p rather than some alternative content q. (Assume)
(2) By hypothesis, the facts in virtue of which S believes p rather than q are not reason giving relations. (Assume)

(3) As these relations aren’t reason-giving relations, they can’t give him reason to believe p rather than believe q.
(4) Any difference between believing p and believing q is nothing to him.
(5) Thus, S doesn’t understand p and q as alternatives.
(6) Thus, believing p just is believing q because there is nothing more and there is nothing less to believing than the subject’s taking things to be a certain way.

As (6) contradicts the assumption that believing p and believing q are distinct, we have to reject either (1) or (2). As (1) is a commitment of the content externalist, the content externalist is committed to the denial of (2) which is (pretend) to accept (A).

Now I suppose (3) is secure but what about (4)? There is something odd in this. By hypothesis the subject lacks the concepts that would enable her to believe q and this suggests that (5) isn't really established by anything about the relation between content determining and reason giving relations. Perhaps what Brewer needs is:

(5') S couldn't understand p and q as alternatives.

Why couldn't he? Well it seems that Brewer thinks this is because we as theorists know that if q had been true rather than p, S would have had the same reasons to believe had q been the case rather than p.

But this is odd. Surely we recognize the possibility of two scenarios such that were either to obtain, we'd have no reason to think one rather than the other was actual and this means that we can understand how they'd be alternatives. Moreover, it seems possible for there to be two alternatives like this where we know that one of them is true. Borrowing an example I've heard attributed to Neta (I don't know what he used the example for), I know that my dogs don't speak German. I also know that if my dogs spoke German in secret, I'd have the same reasons to believe my dogs didn't speak German as I actually do. But that doesn't change the fact that I know that neither Simon nor Ollie speaks German.

Does anyone who knows their Brewer know how to better formulate his argument for (A)?

ME VS. BOGHOSSIAN

I've added another paper to my ever growing collection of online papers. It is still far from finished, but I might as well put it online. Comments would be very much appreciated. Here it is: Boghossian's Other Argument for Incompatibilism.

BOGHOSSIAN'S OTHER ARGUMENT FOR INCOMPATIBILISM

I’ve been working on the chapter where I discuss the significance of content externalism to the debate between the justificatory internalists and externalists when I came across a rather interesting argument of Boghossian’s ('Externalism and Inference' in Ludlow and Martin (eds.)) that hasn’t received any discussion in the literature as far as I can tell. He writes:

The content properties of thoughts are individuated in terms of their relational properties. [From which it follows that] we could not know what we think merely by looking inwards. What we would need to see, if we are to know by mere looking, is not there to be seen. And … there appears to be serious objections to the suggestion that we may know our thoughts on the basis of nothing.

He provides us with the following argument:
(1) You cannot tell by mere inspection of an object that it has a given relational or extrinsic property.
(2) The content properties of a thought are relational or extrinsic.
(C) You cannot tell my mere inspection what the content properties of a thought are.

He then provides an additional argument for (1) in this passage:
This principle is backed up by appeal to the following two claims, both of which strike me as uncontestable. That you cannot know that an object has a given relational property merely by knowing about its intrinsic properties (A1). And that the mere inspection of an object gives you at most knowledge of its intrinsic properties (A2).

The first problem can be brought out by recognizing that for the argument to go through, the relational properties mentioned in (2) must be relational properties of the 'object' that is mentioned in (1). The content externalist is committed to saying that some mental properties are relational properties in virtue of the twinning arguments they use to motivate their views, but those arguments show that they are properties of a person, not properties of the thoughts themselves.

Now, there are views on which this sort of criticism makes sense. Davidson's combination of token identity theory and content externalism invites us to imagine that there are mental events with an intrinsic physical basis. He seems to think that we should think of the relation between a neural event and that event's being a thought with a particular content as we think of the relation between a condition of your skin and the fact that that condition is a sunburn. With this model in mind, it is tempting to suggest that just as we can't know whether it was the sun rather than a powerful lamp that is responsible for the burn, we can't know by inspecting the intrinsic physical core of the mental event whether it is a thought that water is wet or a thought that twater is.

But suppose the identity theory is true. I'm still sceptical.

There are examples that support (A1) and (A2). But that doesn’t show that they are true. Let me explain. To do that, let me take you back to 1768 when Kant was wrote ‘Concerning the Ultimate Foundation of the Differentiation of Regions in Space’. Kant thought he had a ‘simple’ thought experiment that showed that space couldn’t be a purely relational matter. Imagine a universe that consists only of one hand. Either it is a right or it is a left hand, there are no ‘neuter’ hands that aren’t terribly mangled and the hand I’m imagining isn’t like that. As right and left hands have no internal difference that distinguishes them, it must be by reference to something external to the hand in virtue of which it is right rather than left (or left rather than right). As the universe contains nothing but the hand, it seems the universe is responsible for this.

So according to Kant at least, being a right hand is a relational property of your right hand. No amount of non-chiral knowledge could reveal whether the hand was a right or a left but you know your right from your left hand. So I think (1) is false.

Ah but wait, you say, didn’t you convincingly show that Kant’s argument for substantivalism was inconclusive? Well I thought so, I’m surprised that you did, but thanks. Anyway, I argued that Kant had no justification for rejecting a claim open to the relationalist which is that being a right hand is an intrinsic property of such hands, not a property grounded in the relations between the parts of the hands. Graham Nerlich argues (‘Hands, Knees, and Absolute Space’, J Phil (1973)) that three-dimensional relations between parts can distinguish right from left hands. However, what is a relational matter is whether a human hand is an enantiomorph. That is, can we take one hand and make it occupy the same spatial region currently occupied by its incongruent counterpart simply by moving it about? This depends on the nature of the space in which the hand is contained.

What is problematic with Nerlich’s attempt to use this in an argument for substantivalism is that now it seems that it is an intrinsic property of regions of space whether it determines a particular hand to be right or left. Isn’t this just as bad as a relational view that takes the handedness of a particular hand to be a brute fact? Well suppose that Kant and Nerlich don’t win this debate. The objection to Boghossian doesn’t require that we use the example of incongruent counterparts to disconfirm (1) directly because if I’m right and space is relational, then we know that one of the assumptions he uses to argue for (1) must be false. Why? Because even if the ground of the difference between my right and left hand is a matter of the intrinsic properties of my hands, my knowledge of the difference is surely grounded (or could be grounded) simply in terms of my knowledge of the relations between me and, say, my thumb, forefinger, and palm. This shows that (A2) is false.

It looks like Boghossian's argument for the incompatibility of self-knowledge and content externalism is plausible when we limit our stock of examples but an expanded set of examples shows that there really is no defending the assumptions of the argument.

A PUZZLE ABOUT JUSTIFYING

According to Robert Audi ('Justification, Truth, and Reliability', PPR 1988), there is a conceptual truth that links the process of justifying to the state of being justified:

(PPI) A belief is justified iff it has one or more other non-normative properties such that citing them can … both show that it is justified and … constitute justifying it (1988: 6).

This is the 'process-property integration' thesis and it is the stick Robert likes to beat externalists with. For example, I can justify your believing p knowing full well that p is false and that p was arrived at by an unreliable process by showing that the mistake was a reasonable one given your perspective, powers, and so on. Thus, by appeal to PPI it appears we have a powerful argument against reliabilism.

At least, we have a seemingly powerful argument against the necessity of reliability for being justified. It is not an argument against the sufficiency. After all, the property of being correct or being produced by a reliable process seems to me to be a pretty good way of justifying believing. Now maybe I'm missing something here (I know Robert would say that I am) but he lives in Indiana now so I won't have to worry about my missing something until he reads drafts of my dissertation. That won't be today.

I promised you a puzzle and here it is. Three claims strike me as correct and to solve the puzzle you have to explain how they could be or explain why appearances are misleading:

(1) To justify your believing p, I needn't endorse believing p. [e.g., I can justify your believing p knowing ~p and if I know ~p I can't say you should believe p. In fact, I think I can't say there is a justification for believing ~p but set that aside].
(2) To justify my believing p, I must not remain neutral on whether one should believe p but must endorse believing p.
(3) You and I can offer the same justification for believing p.

If I'm trying to justify my believing p, I can't say that this justification has been a success while recognizing or having been forced to recognize that p is false. That's why I accept (2). But then how can this be reconciled with (1)? If it can't, how could (3) be true?

The key is to appreciate that we have to distinguish between two processes of justifying distinguishing between those justifications that have truth as a requirement and those that don't. I'm tempted to say that retrospective justifications don't have a truth requirement but prospective ones do, but this wrongly suggests that the important difference is temporal. I'm also tempted to say that the difference runs along first- and third-personal lines, but this is also a mistake. If there is akratic belief, I can know I believe p without endorsing it so the kind of endorsement that comes with a p-presupposing justification isn't simply a matter of first-person perspective. Moreover, there is no reason a third-person party can't give a p-presupposing justification of my belief when the third-person justifier is herself committed to the truth of p. So my solution is simply to say that there are two kinds of processes of justification and given PPI, two distinct properties or states of being justified. How are they related?

To me it seems that the difference amounts to a difference between whether the agent or the belief is justified. To justify the belief, one aims at showing that it is true. Such a justification is defective if the belief is false and thus the property of justification isn't one that a false belief can have. To justify the agent, one aims at showing that the agent was reasonable in coming to believe what she did given her position, her perspective, and her powers. Such justifications aren't defective if the belief is false and so this property of being justified doesn't have truth as a requirement.

Unfortunately, this tidy solution faces a difficulty which is that Kvanvig has argued that there is a logical connection between personal and doxastic justification I'd have to deny if I took this line. My next post will be a somewhat detailed discussion of his propositionalist view of justification.

SURELY TERROR WARNINGS AREN'T POLITICALLY MOTIVATED...

If they're making jokes about the political motivations behind terror warnings on Comedy Central, then I don't think the suggestion that this is what drives them is all that crazy (how's that for a test?). Anyway, check out this. Apparently Bush's wife was doing a meet and greet in one of the buildings that terrorists were supposed to be targeting. What a brave woman.