SORITES, ISSN 1135-1349

Issue #14 -- October 2002. Pp. 16-20

Copyright © by SORITES and Jussi Haukioja

Water, Phlogiston, Brains, and Vats

Jussi Haukioja


Ted A. Warfield (1998) has presented a new version of the PutnamianFoot note 2_1 argument for the conclusion that we are not brains in a vat (henceforth BIVs).Foot note 2_2 Warfield's version relies on widely held views about content externalism and self-knowledge, but he intends his version to avoid reliance on other, more questionable background assumptions which other versions have made. Anthony Brueckner (2001) claims, however, that Warfield's version fails. In this paper I will argue that, even if Brueckner's reply succeeds, it leaves room for a new version of the argument. The new version relies on assumptions no more objectionable than Warfield's, yet it is immune to Brueckner's objection.

Warfield's version of the argument goes as follows (Warfield, 1998, p. 129):

(A) I think that water is wet. [Self-knowledge]

(B) No brain in a vat in an otherwise empty world can think that water is wet. [Externalism about content]

(C) So, I am not a brain in a vat in an otherwise empty world. [A, B]

This kind of a version of the BIV argument was, as far as I know, first suggested by Thomas Tymoczko (1989).

Warfield's version avoids reliance on two questionable assumptions. Firstly, we need not assume the implausibly strong externalist thesis that «one can think about Xs only if one has had causal contact with instances of X.» (Warfield, 1998, p. 130) Here Warfield's version departs from what seems to be Putnam's own most recent formulation (Putnam, 1992). It is easy to see that this externalist thesis is implausible -- otherwise we could prove that we have had causal contact with unicorns or phlogiston, on the basis of the apparent fact that we can have thoughts about them. Since we can have thoughts about entities with which we have not had causal contact, (B) does not follow from the mere fact that there is no water in the BIV world. Warfield (1998, p. 131) presents two ways in which one might think about water without having interacted with water. First, one might theoretically construct the concept of water from the concepts of hydrogen and oxygen. Second, one might defer to others in one's language community. But Warfield claims that neither of these ways is available to the BIV -- first, we are assuming that there is no hydrogen or oxygen in the BIV world; second, there are no experts on `water' for the BIV to defer to. So (B) seems to follow without reliance on the implausible principle noted above.

Secondly, Warfield avoids taking a stand on the question of what, if anything, the BIV's words or concepts refer to. Many versions of Putnam's argument rely on the assumption that, for example, the term `water' (or the corresponding concept), as used by the BIV, refers to the computer states responsible for its `watery' experiences, or to phenomenal images.Foot note 2_3 Warfield's version only assumes that the BIV's water-thoughts are not about water. And this much, Warfield claims, follows from the externalist considerations above.

Brueckner objects that, in assuming (B), we are assuming that we are not in the same predicament with `water' as we were with `phlogiston':

          In defending [B] without reliance upon any problematic existence assumptions, I must countenance the possibility that my term `water' is on a par with my `phlogiston'. But if there is no water and `water' thus fails to refer to any existing natural kind, then nothing in content externalist theory allows me to argue that a brain in a vat is barred from thinking that water is wet -- barred from thinking exactly what I am now thinking via my sentence `Water is wet'. (Brueckner, 2001, pp. 111-112)

Hence, according to Brueckner, Warfield is covertly assuming that water exists.

One might disagree with Brueckner about whether the BIV and I would think the same thought via `Water is wet', were `water' to fail to refer. In other words, one might try to respond to Brueckner by arguing that even if both our `water' and the BIV's `water' fail to refer, it does not self-evidently follow that our water-thoughts and the BIV's water-thoughts have identical contents, and count as the same thought.

However, I will not attempt such an argument here. Rather, I will suggest a new argument which, like Warfield's, relies on neither of the two assumptions noted above, but is immune to Brueckner's objection against Warfield. My version does, however, rely on the following equivalence principle:

(EP) If there are Ps, my concept of P refers to Ps.

(EP), it seems to me, can be known to be true a priori by any thinker. The first person formulation is crucial: I can know that, if there are Ps, my concept of P refers to them. But, of course, the externalism supported by Twin Earth thought experiments entails that someone else's qualitatively identical concept may in fact not refer to Ps, if her environment is different.

Warfield (1998, p. 134) considers a principle closely related to (EP), as a possible premise in anti-BIV arguments, but concludes that such an argument will be «difficult to provide» (ibid.). However, here is such an argument:

  1. I have the concept vatFoot note 2_4. [Self-Knowledge]

  2. Either there are vats or there are no vats. [Law of Excluded Middle]

  3. If there are vats, my vat refers to vats. [EP]

  4. If there are vats, the BIV's vat does not refer to vats. [Externalism]

  5. If there are vats, I am not a BIV. [1, 3, 4]

  6. If there are no vats, there are no BIVs. [Definition of `BIV']

  7. If there are no BIVs, I am not a BIV. [Universal Instantiation]Foot note 2_5

  8. If there are no vats, I am not a BIV. [6, 7]

  9. I am not a BIV. [2, 5, 8]

(4) seems to follow from the same line of reasoning as Warfield gave for (B). The BIVs, we are assuming, have not had causal interaction with vats. Were they to be able to refer to vats, the reference would have to arise in some other, indirect way. Clearly deference to other members is out of the question. Could the BIVs vat come to refer to vats if they constructed it from other concepts? For example, could they not refer to the vat in which they are with the definition «the container in which I am»? Maybe they could, maybe they could not. But note that even if they could, and their vat were associated with this definition, that would mean that their vat would not be qualitatively identical to our vat -- for our vat is not associated with such a definition. Our vat refers to the kind of thing we have had causal interactions with (assuming, as we do in the antecedent of premiss 4, that vats exist).Foot note 2_6

In this argument, it is crucial that we are discussing vats and not, for instance, water -- there is no suitable analogue of (6) for water. For a version which employs water to work, we would need to show that the BIV's water denotes something -- most plausibly, the states of the computer to which the BIV is connected. But this would be to diverge from Warfield's original project of not relying on such assumptions. By stating the argument in terms of vat instead of water, we avoid Brueckner's objection. If our vat fails to refer, that can only be because there are no vats, and hence no BIVs; if it does refer, the externalist considerations above show that the BIV's vat cannot refer to vats. Either way, we can know that we are not brains in a vat (of the Putnamian kind).

Finally, it might be objected, have we not just described a counter-example to the equivalence principle (EP)? Isn't the BIV scenario one in which there are vats, but vat does not refer to vats? This objection would miss the importance of the first-person formulation of (EP). The principle holds of our vat in the BIV scenario, because the concept refers to all and only the vats we are assuming to exist. On the other hand, the BIVs would also be correct in believing (EP) -- that is, they would be correct in thinking `If there are vats, my concept of vats refers to vats', because none of the occurrences of `vat' in that sentence would refer to the vat in which they are.

My argument does not establish that our vat in fact does refer, or that any of our concepts refer. But, if sound, it does prove that our thoughts about BIVs of the kind discussed here are not about us. Of course, one might also take this to be an argument against the compatibility of self-knowledge and content externalism and claim that, if the conjunction of these theses enables us to deduce that we are not brains in a vat, the conjunction must be false.Foot note 2_7 However, here I echo Warfield's (1998, pp. 137-140) sentiments. The argument given here is not a general refutation of sceptical scenarios, but a highly restricted one. I grant that it may be surprising that such an argument can be given. But if one feels that the premisses have considerable independent plausibility (as many do), the possibility of an argument of the kind given here should not lead one to abandon them.Foot note 2_8


Jussi Haukioja
Department of Philosophy. University of Turku
FI-20014 Turku