Brains in a Vat

“Brains in a Vat”: Whose Brains? Which Vat?

Skeptical Thoughts on Reference and Knowledge

‍

There is in Western thought and literature a vertiginous topos, perhaps most terrifyingly illustrated in Satan’s despairing soliloquy of Book IV of Paradise Lost:

Me miserable! which way shall I fly

Infinite wrath, infinite despair?

Which way I fly is Hell; myself am Hell;

And in the lowest deep a lower deep

Still threat’ning to devour me opens wide,

To which the Hell I suffer seems a Heav’n.  (ll. 73-78)

No matter how far Satan descends into the hell of despair, there is an incommensurably deeper hell--one so deep as to make the present hell of suffering “seem [...] a Heav’n.”  Versions of this mise en abyme are not always so unpleasant, or so paralyzingly infinite; but they are often as thought-provoking.  Aquinas, for example, speaks in the Summa Theologica of a great hierarchy of angels, such that lower angels can receive enlightenment from higher angels, but cannot impart enlightenment to them (Part I, Question CVII, Second Article).¹  And Jonathan Swift’s Gulliver, in a passage revealing of the recent 17th-century microscopic and telescopic extensions of our world, declares:

Undoubtedly Philosophers are in the Right when they tell us, that nothing is great or little otherwise than by Comparison:  it might have pleased Fortune to let the Lilliputians find some Nation, where the People were as diminutive with respect to them, as they were to me.  And who knows but that even this prodigious Race of Mortals [the Brobdingnagians] might be equally overmatched in some distant Part of the World, whereof we have yet no Discovery?    (Book 2, Chapter 1)

And in still another vein, the history of atomic and particle physics seems in the 20th century to be the story of ever-new “worlds” or entities being discovered further down the “chain of being,” revealing parts in what were previously thought to be indivisible wholes.²  I’ll wish to return to this distinctive topos in more purely philosophical fashion shortly.

1. In a recent reprise of his notorious essay “Brains in a vat” (see Preface), Hilary Putnam set out what he sees as the essential task for someone who thinks that the argument he mounts—essentially an argument about the nature of, and constraints upon, reference—is not persuasive.  If one doubts its cogency, or the claims that he makes, Putnam has suggested, then challenge the premises—or show how conclusions other than his must follow.  Much has been written in answer to the first form of challenge. But it is the latter that is my interest here, especially insofar as it is reminiscent of the Greek skeptical enterprise embodied in Pyrrhonism.  And in the course of my effort to accept Putnam’s challenge—to show that more follows from his argument than he wishes, or can control—I’ll eventually have recourse to the characteristic vocabulary of skepticism, especially ataraxia (αταραξια) and isosthenia (ισοσθενεια)—terms which suggest, respectively, the attaining of peace of mind, equanimity, and the state of mental suspense, equilibrium, or equipoise.³

Such a skeptical tack would seem to provide a useful focus for the energies of those who have been made distinctly uneasy by the conclusions of the “brains in a vat” argument, since as Putnam acknowledges, those conclusions often produce the impression of trickery: for all the logical acuity of the piece, many readers have a sense of something not said, some qualification not acknowledged, some skeptical possibility not confronted.  There seems to be an enduring suspicion that the argument from self-refutation, which is adduced with such clinching force, must have been arrived at illegitimately—hence the intense focus on the premises of Putnam’s argument (even on the apparently incontrovertible premise that “we can ask whether we are brains in a vat”).  

My strategy will not be to attack Putnam’s premises.  Though I, too, have shared the sense of “trickery,” the problem seems not to be in the nature of the premises, or the inferences from them. My unease comes from the sense that something else does indeed need to be said beyond “if we are brains in a vat, then the sentence ‘We are brains in a vat’ says something false (if it says anything).  In short, if we are brains in a vat, then ‘We are brains in a vat’ is false.  So it is (necessarily) false” (Putnam 1981, p. 15).  

To say something more about these remarkable sentences, however, I don’t think we need to make either of the two errors Putnam specifies as likely to occasion our resistance to his conclusion: (1) “taking physical possibility too seriously” (I will posit only a peculiar extension and replication of the same set of physical possibilities that undergirds Putnam’s argument); or (2) “unconsciously operating with a magical theory of reference” (for the sake of my argument, I will take Putnam’s account of reference to be exactly right, at least for English, and as an adequate philosophical basis for denying the possibility of our world being referred to in “vat English”).  

2. Putnam’s important thought-experiment may be most consequentially qualified if we consider the possibility that there exists a complex nesting of “English to ‘vat English’” relations.  By this I mean that the conclusion of his argument is “true” only in the situation (which is, most notably, ours) of “English vis-à-vis ‘vat English’”—and that his argument can say nothing, other than perhaps by means of speculative analogy, about the truth of English vis-à-vis a possible “supra English” (which would be to English as English is to “vat English”).⁴  

Why the metaphor of “nesting”⁵?  We know that it is possible to imagine that those controlling Putnam’s vat may manipulate the brains in it in such a way that those brains imagine they are conducting their own “brains in a vat” experiment—indeed, so that the vatted brains may imagine creating brains that, in turn, would imagine they are creating brains in a vat, and so on ad infinitum.  I would claim that given the possibility of these infinitely extending nested, or interlocking, relationships—on the model of English-to-“vat-English”—we have no logical reason to suppose that the nesting of possibilities can’t extend in both “directions”: down to a nesting of “vat English,” “vat/vat English,” “vat/vat/vat English,” etc., but also up to a nesting of English, “supra English,” “supra/supra English,” ad infinitum.⁶

What might follow from such nesting?  We can’t say any more, as “true,” than Putnam says we can in his original chapter.  And the argument from the logic of self-refutation endures.  But we now face the question: is “truth” anything but “truth-as-a-causal-constraints-theory-of-reference-permits-at-our-point-in-the “...‘vat-English’ to English to ‘supra English’... series”?  What would it mean to think of such “truth” in any other sense?  How would we know we had reached a vantage from which some other sense of the word could be made?

The upshot is that we can’t know⁷—in any perspectival sense—where we are in the series; moreover, we can’t know fully what it means not to know where we are in the series.  Most important, we can’t know we are not brains in a vat except insofar as we are (inevitably) speaking about a particular point in the series (which we might label the “English::‘vat English’ point”).  We don’t know that we aren’t brains in a vat in any ultimate sense—only vis-à-vis the first stage of the brains-in-a-vat experiment that our language allows us to formulate, and which Putnam articulates so fully.  This might account for why many readers experience the sense of “trickery,” but also find that they can’t avoid the conclusion that Putnam’s argument is sound.  I would suggest that there is a radically different meaning to that “trickiness,” and a context for this “soundness.”

Even so, Putnam might still wish to claim that there exists some version of “causal constraints upon reference” at every point, “up” or “down,” in the series.  But it’s not clear what non-tautological meaning such a claim might have for “supra English,” or that we can usefully say anything about what it might mean, or that we can specify what might make it meaningful.  (One feels at this moment a peculiar proximity to the conclusion of Wittgenstein’s Tractatus).  Putnam might also wish to argue that the argument from self-refutation obtains at every point in the series; but again it’s not clear what non-tautological meaning such a claim might have.  If it is impossible for us to imagine a language “above,” or more “powerful,” than ours in which the argument from self-refutation does not obtain, this is because the very idea of any such language and the argument from self-refutation are logically co-extensive.  It makes no sense to say that “supra English,” or some version of it, is any way distinguished—or is somehow specifically referred to—by virtue of our being able to say that argument from self-refutation obtains: there is no other logical possibility imaginable.

If I’m right that both the conclusion and “rightness of argument” in Putnam’s chapter have a context, then two possibilities follow: (1) that other speculative conclusions might interestingly be drawn from this line of thinking; (2) that no conclusions might be drawn, except that we can become aware of this weird way of thinking about what we can know about the possibility of our being brains in a vat.

If the first of these is the case, we might be inclined to speculate about what, if anything, could be said about “supra English” (beyond the logical condition of the argument from self-refutation necessarily obtaining).  Whatever speculative possibilities we entertain, however, we must emphasize again that we cannot refer to “supra English,” only to the logical possibility created by our reversing (on the basis of logic alone) “direction” in thinking about the “...English to ‘vat English,’ ‘vat English’ to ‘vat/vat English’...” series.⁸  This claim is crucial, for Putnam would argue that if we can refer to “supra English,” then we have already begun, in some sense, to speak of the “referents” of “supra English,” thereby creating a powerful disanalogy between the relation of “vat English” to English and English to “supra English.”  To be sure, there is a disanalogy, an asymmetry, at issue here, but it needs to be perceived in a different sense: the disanalogy is precisely in our inability (contra the case with “vat English” from the perspective of English) to refer in any meaningful sense to “supra English.”  Thus perhaps what we should say first is that any speculation about “supra English” (or some higher version of it) would be asymmetrical with (potentially referential) knowledge about “vat English” (or some lower version of it).  We can look “down” from the perspective of “original manipulators”; but we can’t look “up” in anything like the same way if we have been the “manipulated.”  It seems, in other words, utterly incoherent to speak referentially of “supra English” (conclusion [2] above).  If we presume to speak of “supra English,” we do so only in terms of logical possibility.  

So governed, we may speculate that, at the very least, “reference” in “supra English” must also be determined, in some sense, by some version of what we call “causal constraints”; but of course we could not know anything of the nature of such “causal constraints,” or the manner of their constraining, or even whether or not they are a function of manipulation (electronic or otherwise).  “Supra English” may, or may not, be a language in which the “brains in a vat” question can be philosophically (as opposed to referentially) posed⁹; again, we simply cannot know, or speak about how we could know.  We might at least imagine that there is a last step “up” the “supra English” series; indeed, we seem obliged to say (though without being able to do so referentially) that, finally, there must be something outside the vat.  But we might also imagine, as we (speculatively) “step outside the vat,” that manipulations aren’t actually made by conscious beings at all, but are (as Putnam has suggested) part of an unconscious process.  We could even speculate that at some “upward” point in the “supra English” series causal constraints upon reference turn out to be an amalgam of “referential” and “manipulated” knowledge.  In other words, constraints might be analogous to those that would be produced if a real (“English-speaking”) brain, with knowledge derived from our (“English-speaking”) real world—our familiarity with actual objects and experiences from our world (e.g., “cabbages,” “sex,” “chemistry labs,” “philosophy books,” “autumn”)—were put in a vat and then hooked up and manipulated.  (Such manipulation—including possibly inducing in the newly vatted brains amnesia about the possibly unpleasant vatting process—would, of course, still have to come from outside the vat, the “world” in which the now vatted brains gained their referential knowledge.)

These are mere speculative possibilities, and in their very peculiarity suggest the absence of referential constraint. That such possibilities induce a strong sense of cognitive vertigo seems another argument against the appropriateness of “knowledge,” in particular referential “knowledge,” as the term for defining what we might say of “supra English.”  Finally, all we can do is imagine analogies.  And however interesting imagined analogies may be, they are revealing only of mere possibility—there is nothing referential about them.  

We can put this point inversely: for us to know that we are not brains in a vat—for the original form and import of Putnam’s argument to hold—we would have to know that a purely logical possibility (as opposed to one in some way dependent on causally constrained reference) for a relation between English and “supra English” does not and cannot exist, at least a relation on the model of “vat English” to English.  But how could we know this?  In other words, how is a (causally constrained) reference to “supra English” involved or implicated in this logical derivation of the possibility of “supra English”?  It would seem that on Putnam’s terms, if we are to know that we are not brains in a vat, then a “logical possibility” and a “causally constrained reference” must be identified with one another.  But at this point, it is not at all clear what wheels the phrase “causal constraints” (theory of reference) is turning.  Indeed, in not distinguishing between a logically derived possibility and a causally constrained reference, Putnam seems implicitly committed to what he himself calls a “magical” theory of reference and representation.¹⁰

As I suggested at the beginning of this essay, the Greek skeptics had several quasi-therapeutic terms for the consequences of their skepticism, including isosthenia—a kind of intellectual equilibrium, equanimity, equipoise (useful in eliminating intellectual anxiety, and producing ataraxia (αταραξια).  Well aware of the argument from self-refutation, Greek skepticism at its best could not be drawn into a commitment to the proposition “we can have no knowledge about the truth of a statement” (if such a proposition is ‘true,’ it is false).  Rather, in pointing out difficulties in the claims of others to knowledge, they implicitly suggested the futility of such claims, indeed precisely those of the sort, “we know we are not brains in a vat.”  Thus my argument against Putnam constitutes a modern version of Greek skeptical isosthenia (ισοσθεηειa).  One can’t, of course, argue someone into such a state.  One can only offer examples, point out the implications of assumptions, elucidate contradictions, much as Greek skeptics typically said, in effect: “you might want to consider this—it seems not to confirm what you say,” or “this doesn’t seem to square with what you claim,” or “given your premises, couldn’t one also argue that...?”  Insofar as philosophers are able to work themselves into some anxiety or confusion (Greek tarache [ταραχηε]¹¹ over whether or not they are brains in a vat, my comments on Putnam’s argument might put them exactly at the point of a very odd equipoise:

...looking “down,” they might believe that for them (and for others who seem to be in their world) brains exist in a vat only if someone (or something) in their “world”—a world the knowing of which they can understand only by means of a causal constraints theory of reference—put the brains there; and they would know that one significant question that can be entertained philosophically in English (whether or not their brains had been put in a vat) has an answer: no!

...looking “up,” at least thinking about what it might mean to look “up” in the sense of entertaining the logical possibility that I’ve outlined here for a series of the “‘vat English’ to English to ‘supra English’” sort, they can’t be sure, can’t “philosophize” (in a Putnam-like sense), can’t refer, can’t do more than entertain interesting imaginative analogies.

The philosophical clarity of “looking down” is balanced by the philosophical impossibility of “looking up”: isosthenia, of a sort.

3. If this commentary is effective in the skeptic’s sense, it suggests that the implications of Putnam’s larger argument about the nature of representation and reference need significant contextualization, and that his apparently bedrock conclusion deriving from that argument has been shaken.  If we no longer have a modern equivalent of the certainty of the Cartesian “cogito”—a certainty that has for over three centuries affected Western epistemological aspiration—then the nature of present philosophical enterprises may look rather different, with certain questions becoming newly exigent.¹² What does the logical possibility of an infinite nesting of “vat English” to English (and English to “supra English”) suggest for our thoughts about, and our speaking of, our being?  what ultimate knowledge of our “vat status,” our being in the world, can we have?  Are traditional philosophical questions about ontology and epistemology indeed reinvigorated by my skeptical challenge?  And what does all this suggest about the viability of skepticism on Pyrrho’s/Sextus’s model in modern philosophy?  Even if skepticism surrenders its original “therapeutic” ambitions, does it continue to have a consequential, and indeed defining, role in the history of philosophy?   Must, for example, the pragmatists’ wish for an end to “Philosophy” continue to be unrealized?  And finally, does the formalism of my discussion here suggest that mathematics can be of greater service or suggestiveness in philosophical inquiry?  Perhaps the relation of the Löwenheim-Skolem Paradox to the “brains in a vat” question, articulated brilliantly by the late and much missed Thomas Tymoczko, is deeper than we have heretofore realized.

Smith College

Northampton, MA

‍