Boghossian presents the following argument (suitably modified to avoid some misleading aspects) as a starting point toward a reductio of externalism about content:

            (A) If I am thinking that water is wet, then water exists in my environment.
            (B) I am thinking that water is wet.
            (C) Water exists in my environment.

    This is, of course, a valid argument.  (A) is supposed to follow from the externalist doctrine.  (B), as it stands, simply describes an episode of thinking that I might engage in.  As such, it may well be true on some occasion.  So we can take the argument as sound, too.  So far, no threat to externalism.  However, Boghossian thinks externalism is threatened by recognizing that both premises of this argument are available to me apriori.  He thinks the truth of externalism is something establishable (if at all) without empirical investigation (after all, we only use thought experiments!).  Now (B) represents a potential item of so-called self-knowledge.  And Bog thinks the distinctive way we know our own present states of mind (which makes no use of inference, evidence, or observation) makes knowledge of (B)’s truth also apriori.  We should note that Bog’s use of “apriori” here is very promiscuous.  It is possible (and sensible) to question the apriority of both A and B.<1>   But, for now, let us just try to reconstruct his reasoning.  At least in some sense, it looks like (A) and (B) can be known by reflection.  And Bog thinks that this implies that the conclusion (C) should also become available to me through reflection (once I engage in the reasoning).  But, clearly, if water does exist, that is not something anyone can know through reflection alone.
    In class, I (DB) suggested that Bog’s reasoning does not take proper account of the Davidson-Burge-Heil (“DBH”, for short) line regarding how content-externalism can be reconciled with self knowledge.  (The line is summarized in a previous handout.)  Zena questioned my claim, pointing out that Bog can be understood to be offering a reductio which accepts that we can know our minds in a privileged way (=apriori) even if externalism is true.  The question is reasonable, since Bog appeals in his argument to both the truth of externalism and the supposition that we know apriori the contents of our present thoughts.  Pursuing this question, Bill offered the following, explicit argument on Bog’s behalf, asking where exactly the DBH line cut in:

            (1) If externalism is true, then I can know (A) apriori.

(This is because, according to Bog, if externalism is true, it is knowable apriori, and its truth entails (A).)

            (2) If I have privileged self-knowledge, then I can know (B) apriori.
            (3) If I can know (A) apriori and I can know (B) apriori, I can know (C) apriori.

(This is by one additional step of Modus Ponens.)

But,
            (4) No one can know (C) apriori.

So,
            (5) Either externalism is false, or I do not have privileged self-knowledge.

Bill suggested that the externalist could only reject (1) or (2), which, granting Bog’s loose sense of “apriori” may not be so attractive.  (In any case, it may suffice for Bog’s argument to point out that, however we know that water exists, it must be in a very different way from the way we know the truth of externalism and the content of our thoughts.  Yet his argument supposedly shows that we can know about water’s existence simply by reasoning to the truth of a general and broad philosophical doctrine and by telling what the content of one of our present thoughts is.  And that seems patently wrong.)  Bill puts the ball in my court: to say exactly where the DBH line impacts the explicit, seemingly valid argument.

    Let me clarify, first, that I do not reject the initial (A)-(B)-(C) argument, which is valid, and (when I am thinking that water is wet) also sound.  But that argument has nothing to do with knowledge, and does not threaten externalism in any way.  Now let me articulate what I take to be the DBH line.  It says that the (metaphysical) conditions placed by externalism on someone having WATER-thoughts (i.e., thoughts involving the concept WATER, as opposed to TWATER) need not be known, let alone known apriori, by individuals who actually satisfy those conditions.  Further, and crucially, knowing the content of your thoughts does not require knowing that you yourself are satisfying those metaphysical conditions.  So: if you accept the DBH line, I can know (“apriori”) that I am thinking that water is wet, without thereby knowing (let alone knowing apriori) that I am in a water environment.  This is so, even though I cannot have a water-thought without being in a water environment.
    The DBH line can be summarized through the following claims:
    (i) If I am thinking that water is wet, then water exists.  (I.e., externalism is true – it lays down, as a condition on my having a WATER-thought that there be water in my environment).
    (ii) If I am thinking that water is wet, then I (can) know “apriori” that I am thinking that water is wet.  (I.e., privileged self-knowledge is possible.)
    (iii) I can know (“apriori”) that I am thinking that water is wet, without knowing (let alone knowing apriori) that water exists in my environment.  (This is because my knowledge of my thought’s content does not require knowing that I am satisfying the relevant conditions on having a thought with that content.

    Now, (i) is the same as (A) above, and (ii) spells out the privileged self-knowledge claim.  (It’s not quite the same as Bill’s (2), because it puts in the antecedent “If I am thinking…” explicitly, which I think is needed for perspicuity.  But I doubt anyone would object to that.)  But look at (iii).  It flatly contradicts the conclusion of Bog’s argument that I should know (C) apriori (if externalism is true and privileged self-knowledge is possible).  So what is going on?
    What I think this brings out is that the crucial step in Bog’s reasoning is provided by the claim that I know (A) apriori (roughly, Bill’s (1)).  So the question is: if the truth of externalism plus the possibility of self-knowledge doesn’t entail the apriori knowability of the contents of my environment, why should accepting that I can know the truth of externalism “apriori” change the picture?

    Presumably, it is my recognition or realization that (A) is true – that if I am thinking that water is wet, then there should be water in my environment - plus my concurrent recognition that right now I am thinking that water is wet that puts me in a position to deduce that there should be water in my environment.
    But what I find problematic is this.  If I am indeed to accept the DBH line, then at the same time as I affirm (so presumably believe) the truth of (A) I also know that my knowledge of the content of my thoughts in no way secures my knowledge of the existence in my environment of whatever actually gives my thought that content (i.e., of the thought’s ‘causal determiners’).  Yet Bog would have us believe that my knowledge of externalism plus my self-knowledge should put me in a position to have apriori knowledge of the nature of my environment.  So I submit that Bog’s argument must be inconsistent with genuine acceptance of the DBH line.
    If I have to pin down the point where DBH impacts the argument, I guess I’d place at Bill’s step (3).  Accepting the DBH line means accepting that my (“apriori”) knowledge of externalism plus my “apriori” knowledge of my thought’s content does not after all imply that I can have apriori knowledge of the existence of what fixes my thought’s content.   Even as I know what I am thinking – say, that water is wet, and even as I realize that my thinking that thought requires the existence of water around, I am not thereby in a position to ‘deduce’ that water exists in my environs.  I still need to check the world to ascertain the latter claim.  Indeed, if externalism is true, it may be part of the requirement on my having the concept WATER that I recognize that need!
 

1.  The apriority of (A) is questioned convincingly in Falvey and Owens (Phil Rev 1994).