Does the Skeptic Win? A Defense of Moore I argue that Moore s famous response to the skeptic should be accepted even by the skeptic. My paper has three main stages. First, I will briefly outline G. E. Moore s proof for the existence of the external world in response to the skeptic. Moore s argument will serve to illustrate three of the methodological elements essential to Moore s response the Moorean Shift, Moorean Facts, and common sense which I will explore in the second section. With a clearer understanding of these methodological elements in hand, I argue that Moore s use of them ensures that his response to the skeptic is in no way dialectically inappropriate, and moreover gives him the dialectical upper hand. I. Moorean Methodology In A Proof of the External World, Moore argues as follows: (1) If there is one hand here, then there is an external world. (2) Here is a hand, and here is another (ostensibly gesture toward one s own hands). (3) Therefore, there is an external world. The success of this argument, according to Moore, depends crucially on whether we can know and not merely believe that the major premise premise (2) is true. 1 Although Moore concedes that we cannot prove that we can know the world exists, Moore insists that we can know things that one cannot prove. And so Moore takes the argument to prove the existence of an external world, not that he has knowledge of the existence of an external world. That is not to say, however, that Moore lacks knowledge that there is an external world. 1. The Moorean Shift A fuller treatment of Moore s response to the reveals a number of interesting methodological elements at work, the primary one being what has come to be called The G. E. Moore Shift, or the Moorean Shift. A Moorean Shift occurs when one counters an argument by denying its conclusion and shifting its form to a modus tollens from a modus ponens (or vice versa). Moore s example is as good as any other. Confronted with skeptical arguments of the form 1 I.e., whether the argument constitutes a proof of its conclusion depends on its meeting this condition. There is no need here to consider whether the argument so-outlined meets the other conditions for a proof that Moore considers (i.e., that the argument s main premise has to be different from the conclusion, and that the premise has to follow logically from the conclusion), because it does. But I suppose this, too, could always be disputed. 1
(4) If S doesn t know that S is not dreaming, S does not know there is an external world. (5) S does not know that S is not dreaming. (6) Therefore, S does not know there is an external world. (4, 5 MP) Moore responds by pointing out that belief in the existence of an external world (e.g., Moore s premise here is a hand, and here is another ) is more obviously true, or more reasonable to believe, than the conclusion of the skeptic s argument. 2 Thus, Moore argues instead that (4) If S doesn t know that S is not dreaming, S does not know there is an external world. (5*) S knows there is an external world. (6*) Therefore, S knows that S is not dreaming. (4, 5* MT) And now with the two arguments in hand (pun intended), Moore thinks that the only way of deciding between my opponents argument and mine, as to which is better, is by deciding which premise is known to be true. 3 For reasons I mention below, many think Moore s response to the skeptic is problematic; they say it is in some way dialectically inappropriate or fallacious. But even supposing it is, that is not to say the Moorean shift is no good as a general argumentative strategy. Perhaps a deeper appreciation of the Moorean Shift can be gained by illustrating how it can function in a philosophical context other than in response to the skeptic: Zeno s paradoxes. Where Zeno would argue 4 via modus ponens that (7) If it is impossible to cross an infinitely long distance, then motion is impossible. (8) It is impossible to cross an infinitely long distance. (9) Therefore, motion is possible. (7, 8 MP) In response to Zeno, one might shift the argument into a modus tollens as follows: (7) If it is impossible to cross an infinitely long distance, then motion is impossible. (8*) Motion is possible. (9*) Therefore, it is not impossible to cross an infinitely long distance. (7, 8* MT) 2 G. E. Moore, Philosophical Papers (George Allen & Unwin, 1956), p. 226. 3 ibid., 121-122. 4 My formulation of Zeno s argument is influenced by R. M. Sainsbury, Paradoxes (New York: Cambridge University Press, 3 rd ed. 2009), p.11. In what modal context (logical, metaphysical, physical, etc.) the argument is intended to function is an open question here. 2
and insist that because we have more reason to believe (8*) than (8), Zeno s argument fails (perhaps this is the origin of the phrase one s modus ponens is another s modus tollens?). Thus, a Moorean shift provides a satisfying solution, at least to the philosopher with Moorean sympathies, to philosophical arguments for extraordinarily radical or revisionary conclusions, such as the ones reached by the skeptic and by Zeno. 5 2. Moorean Facts The Moorean Shift highlights other important epistemological notions at work in Moore s thought. One such notion is that of a Moorean Fact. Indeed, The cogency of the Moorean Shift crucially depends on the notion of a Moorean Fact. In the Moorean Shift, Moore shifts the epistemological burden on the skeptic by asserting that (5*) is more reasonable to believe than (5), and so concludes with (6*). Moore s confidence that (5*) is more reasonable to believe than (5) is precisely what is supposed to make the shift cogent. He writes: [W]e may safely challenge any philosopher to bring forward any argument in favor of either of the proposition that we do not know [some common sense belief to be true], of the proposition that it is not true, which does not rest upon some premise which is, beyond comparison, less certain than is the proposition which it is designed to attack. 6 So we might say that, in the most general terms, a Moorean Fact is a claim about which of two beliefs (that serve as major premises in arguments leading to contradictory conclusions) is more reasonable to believe. The Moorean Fact in play in the above arguments therefore is: (10) It is more reasonable to believe (5*) than (5). Obviously there is a bit of epistemic relativity inherent in the concept of a Moorean Fact, so defined. There may be and predictably will usually be disagreement about which premise is more reasonable to believe, depending on which argument one endorses. Simply pulling a Moorean Shift and labeling the new key premise more reasonable to believe than the original argument s key premise doesn t exactly advance the discussion. For example, suppose you re a die-hard scientific realist who sees the Minkowskian model of spacetime geometry as not only the best interpretation of the Special Theory of Relativity, but is the sober truth about the structure of reality itself. You are caught in the throws of a passionate argument with your 5 Though, arguably, the skeptic s argument is perhaps the most radical; for if it is true, then not even the conclusions of other, radically unintuitive arguments such as Zeno s argument can be known. I will return to this important difference below (section III). 6 G. E. Moore, Philosophical Papers, pp. 227-228. 3
colleague, a Newton-loving proponent of Neo-Lorentzian Relativity. Your colleague presents you with the following argument: (11) If the Minkowskian interpretation of STR is correct, then there are no relations of absolute simultaneity. (12) There are relations of absolute simultaneity. (13) Therefore, the Minkowskian interpretation of STR is not correct (11, 12 MT) Naturally enough, you will see things differently. Accordingly, you will pull a Moorean Shift on your colleague and insist that (12*) The Minkowskian interpretation of STR is correct. and therefore conclude with your good friend modus ponens that (13*) Therefore, there are no relations of absolute simultaneity. (11, 12* MP) Now, it will do no good to simply assert that the following is a Moorean Fact (14) It is more reasonable to believe (12*) than (12). for which premise (12) or (12*) is more reasonable to believe is precisely what s at issue. In other words, nothing would prevent your colleague from either denying (14), or asserting what he takes to be a Moorean Fact of his own: (15) It is more reasonable to believe (12) than (12*). In such a case, Moore s methodology doesn t seem too helpful. And if the cogency of the Moorean Shift depends on Moorean Facts, one might wonder how effective the Moorean Shift is, too. All that is needed to dispel these worries, however, is a little more clarity on what counts as a Moorean Fact. 3. Common Sense The problem above is that the general characterization of Moorean Facts is too general. Not just anything can count as a Moorean Fact. This is where Moore s commitment to common sense comes in. In particular, when a major premise of one of the competing arguments is an item of common sense, one is in a position to state a Moorean Fact. That s why the Moorean Shift seems to be a cogent move in response to, say, Zeno s paradoxes, but not in response to your Neo- Lorentzian colleague: that there is motion is an item of common sense, whereas that the 4
Minkowskian interpretation of STR is correct is not. But this doesn t entirely solve the problem. What exactly is common sense, anyway? And why do some beliefs count as common sense and others not? This is a hard question, and one to which most philosophers even sympathetic to Moore are not prepared to give a definitive answer. 7 But rather than think of common sense beliefs as meeting a rigid set of necessary and sufficient conditions, maybe it s better to think of a continuum of beliefs, on the one end being commonsensical beliefs and on the other end, radical, revisionary beliefs. Examples of beliefs falling on the commonsense side would be beliefs like Here is a hand, and here is another or There is an external world, or motion is possible. In A Defense of Common Sense, Moore lists a host of other beliefs he considers to be items of common sense, among which are: There exists at present a living human body, which is my [Moore s] body. This body [i.e., Moore s body] was born at a certain time in the past, and has existed continuously ever since, though not without undergoing changes;......at every moment since it was born, there have also existed many other things, having shape and size in three dimensions,... from which it has been at various distances...; also there have... existed some other things of this kind with which it was in contact......there have, at every moment since its birth, been large numbers of other living human bodies... the earth... existed also for many years before my body was born... 8 Examples of beliefs on the other end of the continuum might include beliefs like motion is impossible, or all that exists are ideas, or there are no chairs. Of course, some beliefs may be closer to the far ends of the continuum than others. But the point is that the greater the disparity between any two beliefs on the continuum, the stronger the Moorean Fact will be, and hence the more cogent the Moorean Shift. To illustrate: Common Sense Neutral Radically Revisionary p q r s Where p, q, r, and s are beliefs, the strongest Moorean Facts will be ones like It is more reasonable to believe p than s, which might be represented by (10) above, whereas weak or ineffectual Moorean Facts (which are really pseudo-moorean Facts) will be ones like It is more 7 The most detailed analysis of what does and does not count as common sense that I am aware of is William Lycan, Moore Against the New Skeptics, Philosophical Studies 103 (2001), pp. 35-53. 8 Moore, Philosophical Papers, pp. 33-35. 5
reasonable to believe q than r, which might be represented by either (14) or (15) above. 9 The obvious downside to thinking of common sense this way is that there is a lot of subjectivity about where one will locate a particular belief on the continuum. 10 Nevertheless, a somewhat more clear understanding of a Moorean Fact though admittedly still pretty vague is this: a fact about which of two beliefs (that serve as major premises in arguments leading to contradictory conclusions), is closer to common sense than the other. III. Does the Skeptic Win? Plausibly, the Moorean Shift is most persuasive when you can assert a strong Moorean Fact; i.e., when the premise of your interlocutor s argument is radically revisionary and the premise of your shifted argument is a common sense belief. Thus, everything about Moore s methodology here is fundamentally epistemological in nature common sense beliefs are essential to Moorean Facts, Moorean Facts are essential to the Moorean Shift. This point helps highlight an important difference between the application of the Moorean Shift to the skeptic s argument and to arguments like Zeno s. The difference is that the skeptic s arguments are about the possibility of all knowledge, including any premise Moore thinks is known with certainty to be true. So one might say that the skeptic s argument is metaepistemological in nature. For this reason, many have cast doubt on Moore s response to the skeptic. They say that even if the premise in Moore s shifted argument is true, Moore s response begs the question against the skeptic, or is in some other way dialectically unacceptable. But this can t be said of a Moorean response to Zeno s argument, or some other such argument for a metaphysical, not meta-epistemological, conclusion. So we could say with some plausibility that Moore s strategy is more effective, at least prima facie, against the metaphysical revisionist like Zeno than the epistemological skeptic like Russell. But maybe this is only prima facie the case. Recall that the main premises in the skeptic s and Moore s arguments are, respectively (5) S does not know that S is not dreaming. (5*) S knows there is an external world. 9 For the sake of illustration we might say so. But actually I think the belief that there are relations of absolute simultaneity is solidly on the common sense side and the belief that the Minkowskian interpretation of STR is correct is solidly on the radically revisionary side. In fact, the Minkowskian geometry of spacetime might even be a paradigm radical belief. But to be fair, what s most important is not really a belief s status on the continuum by itself, but the belief s status on the continuum in relation to another. 10 This is admittedly an extremely simplistic way of illustrating the matter. But just because the illustration is simple doesn t mean its would necessarily fall apart during more complicated applications. 6
with the Moorean fact being (10) It is more reasonable to believe (5*) than (5). from which Moore concludes, contra the skeptic, that S knows that S is not dreaming. Those persuaded that Moore s response is dialectically impotent or inappropriate might ask what prevents the skeptic from countering (10) by asserting (16): (16) It is more reasonable to believe (5) than (5*). And the skeptic could assert (16). But the problem is that it, unlike (10), is plainly false; Moore at least has what appears to be swaths of evidence favoring (5*), namely, all appearances of an external world, the least of which being his two hands. Of course Moore can grant the skeptic that such appearances could be deceptive, which is all the skeptical scenarios warrant. And even the skeptic should agree that we at least seem to perceive an external world. But there is no such comparable evidence in favor of (5). It is true that one s evidence would be the same if (5) were true, but it s not like we re having experiences that lead us to believe (5) is true (if the apparent evidence didn t favor (5*), then there would be no point to the skeptic asking how do you know you re not being deceived? ). So the skeptic shouldn t have any problems admitting that (16) is false, unless he is prepared to deny that Moore even seems to be perceiving an external world. But maybe the skeptic can regroup and counter (10) in a different way. Even if the skeptic agrees that (16) is false, i.e., it is true that (16) It is not more reasonable to believe (5) than (5*). the skeptic will be quick to point out that from (16) it does not follow that (10) is true. Indeed, the skeptic can return the complement and say that it is precisely the skeptical scenarios that lead him to think (10) It is not more reasonable to believe (5*) than (5). By countering Moore s (10) with (10), the skeptic avoids having to defend a positive assertion like (16), so in that sense (10) is more modest. However, I don t think this move is any more promising. After all, (10) is not making a claim about what is reasonable to believe at all; it is denying a claim about what is reasonable to believe. And the skeptic denies that of (10) as well. 7
But now consider: if it s not more reasonable to believe (5*) than (5), as (10) says, and it s also not more reasonable to believe (5) than (5*), as (16) says, then the skeptic has just as much reason to believe (5*) as he does (5). But if our two alternatives are knowledge and no knowledge, one wonders why anyone would choose the latter over the former. Surely knowledge has intrinsic epistemic value, such that, all else being equal, it would be better to have than lack. 11 Even if the skeptic prefers no knowledge, he must at least acknowledge that Moore s proof works for him. For if two alternatives A and B are equally reasonable, and George chooses A over B for no other reason than sheer preference, then Bertrand, who for the same reason choses B over A, could hardly criticize George s choice (and vice versa). 11 For defense, see Jonathan Kvanvig, The Value of Knowledge and the Pursuit of Understanding (New York: Cambridge University Press, 2003). 8