Scepticism Comes Alive

SCEPTICISM COMES ALIVE This page intentionally left blank Scepticism Comes Alive Bryan Frances CLARENDON PRESS O...

Author: Bryan Frances

30 downloads 1018 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

SCEPTICISM COMES ALIVE

This page intentionally left blank

Scepticism Comes Alive Bryan Frances

CLARENDON PRESS OXF OR D

3

Great Clarendon Street, Oxford ox2 6dp Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York ß Bryan Frances 2005 The moral rights of the author have been asserted Database right Oxford University Press (maker) First published 2005 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose the same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Data available Typeset by SPI Publisher Services, Pondicherry, India Printed in Great Britain on acid-free paper by Biddles Ltd., King’s Lynn, Norfolk ISBN 0–19–928213–7 978–0–19–928213–5 1 3 5 7 9 10 8 6 4 2

for Margaret

This page intentionally left blank

Preface I once overheard a telling conversation between two of my colleagues. One asked the other about a new book on a topic of some importance to both of them. He asked whether they would have to do anything different because of the book. The second colleague said not, so the first colleague said he would not read the book. The conversation encapsulates an excellent test of the worth of a philosophical work: an idea is important if as a result of it experts will have to change what they do when they work on the idea’s topic. It’s not good enough to get it right, or to choose an important topic, or to choose a topic of contemporary interest, or to come up with something original; it isn’t even enough to do all of these things. I know now that much of the work I have done, especially as a graduate student, failed that test—even work that found its way into leading philosophy journals. However, due in part to some dumb luck, I have stumbled across what I believe to be a worthy idea, one that epistemologists cannot ignore. It is, in effect, a new kind of argument for a new kind of radical scepticism. It must be admitted that the notion of scepticism elicits strange behaviour in philosophers, especially epistemologists. Many philosophers, even contemporary ones who should know better, sometimes assert that no one is really a sceptic. Philosophers are pretty much professionally forbidden from being radical sceptics, even though we aren’t forbidden from believing any of many other comparably outlandish claims (e.g., see Chapter 5 and Section B of Chapter 11 of this essay). Scepticism is commonly, and justifiably, thought to be an epistemic pit with particularly nasty properties. If you are in that pit, so is everyone else; there is no practical way of coming out of the pit and into the light of knowledge; and you most likely have a serious cognitive or epistemic deficit, some kind of failure of your epistemic systems. But, as I will presently show, while these claims hold for traditional forms of scepticism, none of them apply to the sceptical theories I explore here. The sceptical snares that these new theories lay forth apply to just some people (e.g., many philosophers and many scientists), are caused by some pretty random sociological events, are fallen into only in adulthood, can be escaped on one’s own or with the help of others, and indicate no epistemic deficit of any significance. In fact, falling into the

viii

Preface

sceptical snare is actually an improvement in epistemic standing compared to most who have the knowledge denied to those in the snare. That’s right: the person who knows is epistemically inferior to the person who doesn’t know. If I’m not mistaken, we have an entirely new kind of scepticism. In this essay I have restricted myself to investigating the new sceptical argument. I have discussed as little as possible traditional arguments for traditional forms of scepticism. I have not presented my best guesses as to how particular anti-sceptics will react to my new sceptical arguments. I do this not out of any disrespect, but in order to make the essay relatively short, unified, and not cumbersome. I have ignored most traditional epistemological disputes and theories. I have resisted, with effort, the temptation to blather on about my own pet peeves and ideas. I have, with the exception of this preface and a few other places, attempted to avoid the word ‘I’. I have done all this in the hopes that a short book restricted to presenting the case for the new scepticism is less distracting than a long one that drifts from the main task. It’s also easier to write. This essay would have been an article of reasonable length if I had not been so anti-sceptical by both inclination and training. More than half of the essay is devoted to filling out the details of the sceptical solution to the sceptical puzzles generated by my arguments in the first part of the essay (with the intention of finding problems with that sceptical solution), and investigating anti-sceptical responses to those puzzles. The new sceptical arguments conclude that I don’t know some very ordinary facts that one would think are very easily known. For instance, I don’t know that my shirt is red, that I currently have a throbbing pain in my knee, or even that I believe that my shirt is red or that my knee is throbbing. These are difficult results to swallow, certainly quite contrary to common sense, and in this essay I have done my best to avoid swallowing them. As far as I have determined, I have failed. My current best guess is that many philosophers and most members of certain classes of scientists lack these and analogous ordinary pieces of knowledge, if not today then perhaps in our future and definitely in some very realistic possible worlds. Much of the interest in the new kind of scepticism lies in its scope and strength. Regarding strength, these results remain even if we adopt all the clever anti-sceptical fixes thought up in recent years: for instance, reliability, proper functioning, relevant alternatives, contextualism, and the rejection of epistemic closure.

Preface

ix

That is, even if one or more of those anti-sceptical strategies succeeds in defeating the traditional arguments for radical scepticism, the new sceptical arguments go through. Regarding scope, even if my sceptical arguments are sound, you are free to know millions of exotic truths such as the fact that black holes exist! You can know about the existence of black holes, but not about the colour of your shirt or even about what you believe regarding the colour of your shirt. What kind of scepticism is this? Enough teasing! Here is the brutally short and crude version of my long and exceedingly sophisticated argument. In order to know P, one must be able to rule out some :P possibilities. For instance, in order to know that the tree is a fir, one has to rule out the possibility that it’s a spruce or a hemlock. At least, one has to rule out those possibilities provided they are real, scientifically respectable, ‘live’ hypotheses; one is aware that they have such respect; and one is perfectly aware that those hypotheses conflict with one’s belief that the tree is a fir. For instance, you came across the tree while taking a stroll through a forest with a tree expert. You said the tree was a fir, but she said that it’s quite hard to tell from this vantage-point because spruces and hemlocks look the same and there are lots of them around here. Those :P possibilities are ‘relevant alternatives’, as it is often said. Perhaps the brain-in-a-vat possibility doesn’t need to be ruled out; but the spruce and hemlock possibilities do need to be ruled out. Assuming I can’t rule out the spruce possibility, I don’t know that the tree is a fir—even if the tree is a fir. At least, I don’t know it’s a fir tree once I’m aware of the live status of the spruce and hemlock possibilities. But now here’s the kicker: there are several real, scientifically respectable, ‘live’ hypotheses that can be used in the very same argument template as in the previous paragraph. It’s a real, live scientifically expert-endorsed possibility—at least in some possible worlds next door to ours—that no ordinary objects are coloured, that no one believes anything, that no one has any character traits, and that all pains are in brains only. Using the line of reasoning from the tree story, we can show that those of us aware of the live status of these four error theories don’t know that fire-engines are red, don’t know that we sometimes have pains in our lower backs, don’t know that John Rawls was kind, and don’t even know that we believe any of those truths. However, people unfamiliar with philosophy and cognitive science do know all those things, as they are well outside the

x

Preface

domain of philosophy and cognitive science. And all of this holds on the mere assumption that those error theories are false but live. Finally, there is an intriguing way to transform the live sceptic’s argument into one for universal scepticism, the thesis that we know absolutely nothing. ‘Yes, but those cases are different from the tree case, and those differences will block the route to the new scepticism.’ Maybe so, but it turns out to be awfully hard to back up this intuition. I have been unable to find any way to justify it—and I’m anti-sceptical by nature. Furthermore, even if the intuition is right, we can derive some very interesting epistemological results from the subsequent denial of the new scepticism. So, no matter what our reaction to the live sceptic’s argument, we walk away with interesting results. In that respect, we all win. Thanks are due to the University of Leeds for granting me a semester of research leave to work on a boring project that I abandoned in favour of writing this book. Thanks also to a Nouˆs referee (a presentation of the book’s main line of argument appears in my ‘When a Skeptical Hypothesis is Live’, which is forthcoming in Nouˆs), Paul Bloomfield, David Chalmers, Richard Fumerton, Sanford Goldberg, John Greco, Ken Himma, Sarah McGrath, Joseph Melia, Peter Millican, Mark Nelson, Duncan Pritchard, and Tom Stoneham for written or oral comments, and to the faculties at the University of Liverpool and the University of Connecticut for helpful discussion. Special thanks are due to my generous friend Andrew McGonigal, with whom I discussed the issues raised in this book on many very fruitful occasions, and my wife Margaret Frances, for offering substantive comments on the entire essay. Without their good sense and perceptive comments the book would have had many more mistakes and fewer insights. BRSF

Contents 1. Introduction: When a Sceptical Hypothesis is Live

1

2. When Experts Make Mistakes

8

3. How Live Hypotheses Sabotage Knowledge

18

4. The First Live Sceptical Hypothesis A The First Hypothesis and Argument B Initial Defence of the First Sceptical Argument

30 30 41

5. Additional Sceptical Hypotheses

64

6. The Sceptical Solution A Outrage: Sceptical Hypotheses B Outrage: Sceptical Conclusions C Outrage: Sceptical Hypotheses Sabotaging Knowledge

73 74 76 97

7. Universal Scepticism

106

8. The Consistency of Scepticism and Knowledge

114

9. Epistemic Threats, Context, and our Anti-Sceptical Strategies A Epistemic Threats, Context, and Live Scepticism B Our Three Anti-Sceptical Strategies C Scepticism: Live versus Classic

121 122 128 132

10. Disarming the Live Sceptical Threat A When Hypotheses are Set Aside B Knowledge Attributions in Ordinary Contexts

137 137 144

11. Do Live Sceptical Hypotheses Pose Real Threats? A Metaphysical Distance from Actuality B Philosophy versus Common Sense C The Population of the Sceptical Snare D The Ambiguity of ‘Scepticism’ E Testability and Indeterminacy

153 154 156 178 180 182

xii

Contents

12. Defeating the Live Sceptic Head-On A Safety and Sensitivity B Extreme Externalism C The Tenacity of Knowledge

185 186 192 195

13. Concluding Reflections

199

References Index

204 207

1

Introduction: When a Sceptical Hypothesis is Live Philosophy is inseparable from scepticism, which follows it like a shadow that it chases away by refuting it, only to ﬁnd it once again under its feet. Emmanuel Levinas, Otherwise than Being or Beyond Essence

When Jo was a teenager, she learned that a huge meteorite wiped out the dinosaurs. She learned this theory in the usual way, hearing it from her parents, teachers, and books. Now pretend that at the time she was told the meteor story as a child, say at the age of 8, the scientiﬁc community was sharply divided on the issue of what caused the demise of the dinosaurs. Although most scientists accepted the meteor hypothesis, many others subscribed to the idea that their death was caused by some enormous solar ﬂare. A signiﬁcant number of other scientists thought that it wasn’t a solar ﬂare or a meteor, but a particularly nasty series of supervolcanoes. These latter two classes of dissenters had decent evidence: evidence concerning the sun and supervolcanoes that the meteor advocates took seriously. Both the solar ﬂare theorists and the supervolcano theorists were highly respected professors, highly respected by the meteor theorists and at the top of their profession. Whole book series, conferences, and Ph.D. dissertations were devoted to these competing hypotheses. Suppose, further, that upon going to her university, Jo found out about the rival and highly respected hypotheses. She didn’t understand all the reasons why they were so well respected and endorsed, but she was well aware that they were well respected and frequently endorsed by the experts, even the best among them. Even so, she kept her meteor belief. As it turned out, the meteor hypothesis was the right one. But although some experts may have known that fact, surely Jo was

2

Introduction

in no position to know it after she found out about the eminent status of the rival hypotheses. She could do little or nothing to defeat the rival hypotheses. Even after studying the issues as an undergraduate, she couldn’t know that a meteor wiped out the dinosaurs, for from the moment she ﬁrst learned of the dinosaurs until she graduated with a Bachelor’s degree in palaeontology she was perfectly aware that there were two ‘live’ contrary hypotheses that she was in no position to rule out—and they had not been ruled out for her, either, for instance by her teachers. And if she was aware of all this, and the competing hypotheses weren’t ruled out, then she couldn’t know that the meteor hypothesis was correct. Perhaps in order to know the meteor truth, she wouldn’t have needed to rule out the possibility that she is a brain in a vat. And perhaps someone mostly outside the scientiﬁc community, and thus not exposed to the ultimately misleading supervolcano and solar ﬂare evidence, could know that the meteor hypothesis is true. Still, in order for Jo to know the meteor fact, she does have to be able to rule out, to some degree anyway, the solar ﬂare and supervolcano possibilities, for, unlike the crazy philosophical hypotheses, these are ‘real, live possibilities’, and she is perfectly aware of their existence, live status, and inconsistency with her belief. Perhaps she doesn’t have to completely demolish the supervolcano and solar ﬂare hypotheses in order to know the meteor story, but she certainly has to knock them down a few epistemic notches. The heart of the dinosaur argument is simple: because the supervolcano and solar ﬂare hypotheses are real, live possibilities inconsistent with the meteor hypothesis, Jo is aware of all that, she is nothing even approaching an expert or genius on these matters, and those hypotheses are not ruled out, she doesn’t know that the meteor story is true. There aren’t many ideas in philosophy that can’t be coherently and cleverly challenged, but it is a brave philosopher who thinks that the above story is misguided. Surely, one would think, in those circumstances Jo’s true and partially justiﬁed belief in the meteor hypothesis didn’t amount to knowledge. The traditional sceptic uses an argument of roughly the same form: a. Here’s a sceptical hypothesis (e.g., dreaming, evil demon, hallucination, brain in a vat, wildly defective reasoning skills, universe created ﬁve minutes ago).

Introduction

3

b. You must be able to rule out that sceptical hypothesis in order to know much of anything. c. However, you can’t rule it out. d. Thus, you don’t know much of anything. This ‘ruling-out’ type of argument is not, of course, the only way to use sceptical hypotheses to generate powerful sceptical arguments, and one can generate such arguments without sceptical hypotheses. But such an argument is not without an intuitive basis. The typical reaction to this familiar type of argument for scepticism is to ﬁrst zero in on the conclusion, pronounce it crazy, and then go back to the rest of the argument with the expectation of ﬁnding faults that cannot be repaired. This impulse seems retrospectively justiﬁed, because when we act on it, we immediately become highly suspicious of the sceptical hypotheses employed in the arguments. The suspicions of both philosophers and ordinary folk typically stem from three gut reactions to the hypotheses. i. I can’t do anything to prove or refute the classic sceptical hypotheses head on, so to speak, no matter how hard I try. For instance, there’s nothing I can do, such as pinching myself or performing clever experiments, to defeat or establish the brainin-a-vat (BIV) hypothesis. So aren’t these hypotheses idling in some important way? ii. The sceptical hypotheses just don’t seem relevant, for some reason, to our ordinary claims to know. Sure, I guess they’re logically inconsistent with what we claim to know, but so what? Why does that fact about mere logical possibility make the sceptical hypotheses threats to my knowledge that my socks are blue? iii. No one has even the slightest positive reason to think that the sceptical hypotheses are actually true. In fact, for some of them we lack reason to think that they’re even possibly true. So how could they constitute a hurdle for my knowledge, given that there’s nothing positive to be said on their behalf? None of these thoughts constitute objections, as they are just hints at objections, but some or all of them often justiﬁably incline us to reject the whole lot of hypothesis-based sceptical arguments, no matter how diverse its membership. They justify, to a signiﬁcant extent, our conveying the following challenge to the sceptic: until you can use sceptical hypotheses that get around these three problems, the sceptical

4

Introduction

arguments utilizing these weird hypotheses are never going to be good enough to convince us of their highly counter-intuitive conclusions. In truth, the analogy between the dinosaur and traditional sceptical arguments is quite weak, due to the fact that (i)–(iii) apply to the traditional sceptical story but not to the dinosaur story. That’s a large part of why we don’t buy the sceptic’s reasoning. In fact, one might think that the lesson of (i)–(iii) is that the traditional sceptical hypotheses never really posed any epistemic threat at all to our knowledge, despite the logical inconsistency of the hypotheses with much of our alleged knowledge. On the other hand, one might concede the presence of an epistemic threat generated by the sceptical hypotheses but conclude that, because the hypotheses are so crazy or empty or far-fetched, it is incredibly easy to defuse their threat, without even putting any thought into it. If you’re sane and breathing, well then, surprisingly enough, that’s sufﬁcient to defuse the meagre threat posed by the outrageous evil demon hypothesis. These two responses to the sceptic are variations on the fundamental anti-sceptical idea that traditional sceptical hypotheses don’t pose epistemic threats to our beliefs—either because we have all somehow automatically done something to render them non-threatening or because they were never really a threat from the start. So the traditional sceptic almost never converts us, even though she does win our admiration (because we have such a hard time ﬁguring out where she has gone wrong) and gratitude (because she makes us probe the nature of knowledge and epistemic warrant much more deeply than we would without her). This must be frustrating for the sceptic: the permanent loser whom everyone respects but very few accept. If a sceptic could present sceptical hypotheses that elude (i)–(iii) because they are real, live hypotheses akin to the supervolcano and solar ﬂare theories, her argument would be much stronger than any of the classic sceptical arguments: responsible critics could no longer start off their investigation with the presupposition that the sceptic is wrong. We would have to abandon our usual, comfortable reactions to the familiar arguments for scepticism, as the new sceptic’s arguments would be so crucially different from the traditional ones. And this is exactly what my new sceptic does: .

The sceptical hypotheses that she plugs into her sceptical argument template to generate individual sceptical arguments and

Introduction

.

.

.

.

5

conclusions are empirically testable and immune to irrelevancy objections. Further, there are compelling scientiﬁc and philosophical reasons to think that the hypotheses are actually true. Each is a ‘real, live possibility’ actually endorsed by some of today’s experts: some of them currently think that the hypotheses are true. As a consequence, for the new sceptical hypotheses, it is much harder to reject (b) or (c) in the sceptical argument given above. The sceptical hypotheses are currently live, so most of us can’t rule them out. And because they are live and are widely appreciated to be inconsistent with some enormous, cherished set of beliefs, we have to be able to rule them out in order to have those beliefs amount to knowledge. In effect, with the new sceptical hypotheses each of (i)–(iii) above is false, thereby cutting off the initial anti-sceptical reactions suggested above. In some respects the ‘live’ sceptic is more radical than most traditional sceptics. She admits that for all she can show, you may know all sorts of exotic things, such as the fact that the universe is expanding; so in that respect her scepticism is limited (but see ﬁnal point below). However, included in her scepticism are denials of knowledge claims that traditional sceptical arguments usually do not touch (e.g., you often know what you believe). The live sceptic announces that although her arguments apply to you, so that many of your seemingly most secure true beliefs fail to amount to knowledge, two friends of yours might actually escape the arguments, and so know millions of the things you don’t. If one friend is unacquainted with any philosophy or much science, then perhaps he can keep his knowledge. If the other is a genius, then she can not only retain her extensive knowledge but can also know the sad fact that you are a know-nothing loser when it comes to the beliefs covered by the live sceptical arguments. Furthermore, with some work, or appropriate help from others, one can climb out of the sceptical trap and come to know what one could not know before; so one’s time in the live sceptical snare might be brief. If that were not enough to pique your interest, the new sceptic holds that falling into the sceptical trap, and thus losing knowledge, is actually an improvement in one’s epistemic position, and does not reﬂect any defect in one’s cognitive systems, either permanent or temporary. Traditional sceptical conclusions have supporting arguments, of course, but such arguments have never generated much in the way

6

.

.

Introduction of a satisfactory explanation of why their scepticism is supposed to be true. The live sceptic has a nice, familiar, and detailed explanation of why her sceptical conclusion holds. The implications of the new scepticism are circumscribed but surprising: the new sceptical hypotheses don’t generate the mere denial of certain knowledge or ‘high standards’ knowledge; they generate the denial of the entirely modest amount of epistemic warrant we ordinarily expect. In fact, these new sceptical results remain, even if we adopt all the clever anti-sceptical ﬁxes thought up in recent years in the defence of ordinary knowledge claims: reliability, proper functioning, relevant alternatives, contextualism, the rejection of epistemic closure, etc. (more on all of these later). That is, even if one or more of those anti-sceptical strategies succeeds in defeating the traditional forms of scepticism, the new sceptical arguments go through. Finally, some epistemologists think that if you are fully aware that Q is a necessary condition on anyone’s knowing P, and you are fully aware of the fact that there’s a real, live possibility that Q is false (so Q isn’t something silly like ‘You’re a brain in a vat’), then in order for you to know P, you must be able to rule out the possibility that Q is false. If that’s right, then a new sceptical argument ﬁnishes with a conclusion truly universal: you fail to know P for any P, not just the restricted Ps—even though your friends still know loads of truths, you are in a better epistemic position than most of them are, and you can climb out of this universal sceptical trap too. As with the other live sceptical arguments, this one employs a sceptical hypothesis that some experts today actually take to be true, utterly unlike the classic sceptical hypotheses.

In this essay I will defend the outrageous thesis that there is such a set of sceptical hypotheses, each of which generates a new sceptical puzzle. Since I will be generating several new sceptical arguments from the same argument template and defending the seven bulleted claims above, you might expect the outline of my essay to be something like this: 1. Present intuitive considerations supporting the general form of the live sceptical arguments, especially regarding the main principles it uses. 2. Articulate the ﬁrst live sceptical hypothesis and defend the application of the general argument to it.

Introduction

7

3. Formulate the other live sceptical hypotheses and sketch how to apply the general argument to them. 4. Explain why the sceptical solution to the live sceptical puzzles— the position that contains the central claim that the sceptical arguments are sound—has the features described in the ﬁrst ﬁve bulleted points above. 5. Show how to construct a new argument for universal scepticism which is loosely based on the general argument template. 6. Examine the kind of sceptical conclusion common to the live sceptical arguments. 7. Examine general anti-sceptical strategies. 8. Articulate and evaluate particular anti-sceptical solutions to the live sceptical puzzles. You would be right, even regarding the order. But before we get to the live sceptical hypotheses, we must ﬁrst meet the new theory of Smith and Jones.

2

When Experts Make Mistakes Every great advance in science has issued from a new audacity of imagination. John Dewey, The Quest for Certainty There is not a single idea, however absurd and repulsive, that has not a sensible aspect and there is not a single view, however plausible and humanitarian, that does not encourage and then conceal our stupidity and our criminal tendencies. Paul Feyerabend, Three Dialogues on Knowledge

In this chapter I will introduce the main themes of the essay in a preliminary and intuitive way. Although I will be giving many arguments in this chapter, they do not, and cannot, carry the case for the live scepticisms. Again, they serve only to introduce almost all the relevant issues, illustrate the intuitive basis for the more complex sceptical arguments to follow, and generally warm us up for the live sceptical arguments. All your ‘Yes, but . . .’ reactions will, I hope, be addressed later in the essay. Suppose that some current theory is justiﬁed to a certain signiﬁcant extent and is well entrenched in some rich and vigorous intellectual community. But Smith and Jones have come across some data and lines of reasoning that challenge the status quo. With some more research, their case for a new theory is impressive. Most of the defenders of the status quo take the opposing theory seriously. Smith and Jones have marshalled several independent lines of evidence for the new theory, thereby showing that more than some ﬂuke bits of data or reasoning support it. Smith and Jones are invited to speak at important conferences. Their evidence for their theory is published in the most prestigious journals and book series. They become top members of the e´lite professional organizations. Many investigators work on the problem of seeing which theory—the old

When Experts Make Mistakes

9

one or the new one—is correct. It is a hard project, taking at least several decades of work. Smith and Jones win a signiﬁcant number of converts, themselves respected members of the professional community. Naturally, many members of the old guard take a dim view of the new alternative theory. Some think it isn’t a real contender. But it’s clear that they have offered nothing close to a conclusive objection—even by their own lights. Now, as a matter of fact, the new theory is mistaken, and the old theory is right. But there is no public refutation of the new theory. Roughly put, this means that there isn’t any widely known evidence that the well-informed experts agree shoots down the new theory. The new theory is a ‘live contender’ in a socio-epistemic sense.1 But wait. Can any hypothesis really be ruled out? Isn’t everything a live contender, ultimately speaking? That depends on what is being asked. If S has ruled out P, then maybe all that entails is that S has satisﬁed herself that she can exclude P from consideration because, according to her, P is false. But for obvious reasons this dictionary notion is not interesting here. On the other hand, it might require that S knows that P is false. Of course that is pretty strong. Better notions of ruling out, useful to sceptics and anti-sceptics alike, entail at least this weak condition: if S has ruled out P, then S has a signiﬁcant warrant for believing that P is false, and P is indeed false—but S might not quite know that P is false. (This condition will be modiﬁed later on.) On this conception, many hypotheses have actually been ruled out. For instance, the ‘famous deeds’ descriptivist theory of proper names has been ruled out, as have countless scientiﬁc and ordinary hypotheses. That phlogiston exists; that the Earth is the centre of the universe; that Babe Ruth is the all-time home run king of baseball; that the existence of evil is logically inconsistent with the existence of God; that Ronald Reagan was wise; that knowledge is true belief; that the famous Barber paradox is a real paradox; that one cannot jump higher than one’s shoulders; that logical behaviourism is true; that head lice jump or ﬂy from head to head; that the Earth is no more than a few thousand years old; that Kant solved, or provided the essentials for solving, every metaphysical problem (as he actually claimed at the beginning of the Critique of Pure Reason); that water is an atomic element; that electrons are heavier than protons; that there 1 Not a purely epistemic sense. To some ears, ‘H is an epistemic possibility’ entails ‘We don’t know anything that entails :H’. I’m not interested in that usage. I’m interested in H’s status as live in a sociological sense.

10

When Experts Make Mistakes

is no connection among causation, counterfactuals, modality, and laws; that the Turing Test provides modally sufﬁcient conditions for understanding of the appropriate sort; that Mein Kampf is a largely accurate portrayal of Hitler’s life; that the semantic paradoxes can be solved just by distinguishing sentences and propositions; that cricket is more interesting than baseball—all these have been ruled out. Return to the Smith and Jones story. Just because their new theory is a live contender in the intellectual community does not mean that there could not be someone, somewhere, working in obscurity, who has hit on some evidence that really does refute the new theory. Suppose some girl is the greatest mathematical genius ever at the tender age of 14. She even makes Gauss, who is widely considered the greatest mathematician ever, look like an amateur. She has ruled out many mathematical hypotheses that are live ones in her mathematical community. She hasn’t published them, and even if she did, it would take years, maybe decades, before her colleagues would be able to see that she’s right. At that unpublished stage, those hypotheses are not ruled out for the rest of the community, even though they are ruled out for her. A similar point holds for cognitive beings outside an intellectual community. A race of super-intelligent aliens could have ruled out all sorts of hypotheses that are live ones for twentyﬁrst-century humans. Simply because they have ruled it out for themselves, this doesn’t mean that it is ruled out for us. Now if someone in Smith and Jones’s intellectual community did come up with a decisive falsiﬁcation of the new theory such that everyone, including Smith and Jones, eventually agreed that the new theory was dead, then that theory would be ruled out for that community. Of course it would then no longer be a live contender. Perhaps it would be ruled out even for people in the intellectual community who could not understand all the reasons why the new evidence did away with the new theory. Here’s an example. Working entirely on my own, I cannot rule out the hypothesis that water is an atomic element. I would have to use books and experts if I needed to convince someone; and my only role would be to help explain some of what the books and experts are saying. Even so, it is reasonable to say that, as a matter of actual fact, the water-is-an-atomic-element hypothesis has been ruled out for me, even though it would be a bit misleading to say that I have ruled it out. It has been ruled out ‘for me’ because I happen to be a member of a scientiﬁc community that as a group has ruled it out. I am a member of that community

When Experts Make Mistakes

11

because I have a Master’s degree in physics. A less misleading locution, one that I will use later in the essay, is ‘eliminated or neutralized with respect to person X’, which would include X ruling it out and it being ruled out on X’s behalf by someone or something else. So the water-as-atomic-element hypothesis has not, I suppose, been ruled out for someone who knows next to nothing about science. But in any case no such refutation has come up in the Smith and Jones case. The new theory is still a live hypothesis in the intellectual community, a going concern if you will. Now suppose that when it comes to the new and old theories you’re no genius compared to others in the intellectual community. Maybe you are aware of the new theory, aware of what the experts think about it, but can’t evaluate it yourself. Or perhaps you’ve put quite a bit of thought into both theories and can competently hold forth on their strengths and weaknesses. Maybe you’re even a borderline expert in the ﬁeld who has done research on both theories. Suppose, further, that you have your doubts about the new theory. Even so, for any reasons or evidence you can marshal for casting doubt on the new hypothesis, if they were carefully considered by the members of that intellectual community—in particular, by the well-informed, well-respected, and highly intelligent experts who were thoroughly familiar with both theories—they would be vigorously rejected as insufﬁcient to rule out the hypothesis. This isn’t to say that if you were an expert, your colleagues would say nasty things about your research behind your back. It’s just that they don’t consider your case good enough to rule out the new theory. You’re good, but not that good. This type of situation is entirely familiar. If any of those descriptions ﬁts you, in that you’re a ‘mere mortal’ with regard to the two theories, then you would not currently be able to rule out the new theory. Your status as a mere mortal in the intellectual community, especially the fact that your reasons for doubting the new theory, if any, don’t stand up to review by the best and brightest, means that you are unable, in anything like your present state, to rule out the new theory. (You ‘could’ rule it out in the sense that you could take a genius pill and then spend the rest of your life investigating the theories, but that’s irrelevant.) And no one else has ruled it out on your behalf, so to speak. So it isn’t ruled out for you, full stop. If that is the way things are, what follows? You still believe the old theory, which is the true one. But is your true belief knowledge?

12

When Experts Make Mistakes

Surely not! You’ve got to rule out, or be able to rule out, the new theory in order to know that the old theory is right: for if it’s the new theory that’s true, then of course your belief in the old theory isn’t true. Or, someone has to rule it out and have the ruling-out status somehow transfer to you in order for it to be ruled out for you. That is, if someone came along and publicly refuted the new theory, in the sense that everyone, including Smith and Jones, eventually agreed that the new theory was dead, then that theory would be ruled out for almost everyone in that community. So perhaps you wouldn’t have to rule it out yourself in order to know that the old theory is true. But the point here is that absent some relevant kind of ruling out, you don’t know that the old theory is true. Here’s an example. You are talking with someone and make a claim that, if true, would mean that the new theory is false. Now maybe that claim really is the magic bullet for the new theory. That is, a decisive refutation of the new theory can be built up from not much more than that one innocent-looking claim you made. But that is quite unlikely, so set that possibility aside. Think of the water example again. Pretend we are back in time, and the scientiﬁc community hasn’t ﬁgured out whether water is an atomic element. The present professional stance on the matter is that water is a compound, but Smith and Jones have impressive (though ultimately misleading) evidence for water’s being an element. You say, ‘Look, water must be a compound of oxygen and something like hydrogen, because of such-and-such’; call the ﬁrst claim you made P. But then your friend gently reminds you of the new Smith and Jones theory, which of course is inconsistent with P, because the new theory says that water is an element, not a compound. Now you have to retract P, even though it’s true. You have to say something like ‘Oh, that’s right. Well, if it isn’t an element, then it’s got to be compounded as I claimed’, which may be perfectly reasonable and true. Even if your friend had not reminded you and you weren’t thinking of the new theory, you still did not know P, because it is obviously inconsistent with the new theory, the new theory is live, you’re aware of both the inconsistency and the liveness, and you are powerless to put any real pressure on the new theory. That is, even if in the conversational context neither you nor anyone else has recently been or is currently thinking of the new theory, a literal use of ‘S knows P’ would be false. Even if you don’t have a ‘friend’ tagging along with you, pointing out every one of your transgressions of the new theory, you still don’t

When Experts Make Mistakes

13

know P if P is obviously inconsistent with that theory and the new theory is a live contender as described above and you are aware of its live status and inconsistency with P and you’re no genius and you don’t have some rare evidence that refutes their theory. What if the proposition P putatively known is inconsistent with the new theory, but the connection is exceedingly complex? Or if not complex, at least highly unlikely to be noticed? You have said that P is true, where P is true, let us say, but is inconsistent with the new theory in some subtle, unnoticed fashion. The new theory is still a live hypothesis that has not been ruled out for you. Do you know P? Perhaps you do. The fact that P is inconsistent with a live, yet to be ruled out, false hypothesis might not ruin your chances to know P. Perhaps the new theory is not relevant or salient in this situation to your knowing P if you aren’t even aware of the new theory, your belief in P was well justiﬁed prior to the birth of the new theory and its evidence, and the conﬂict between P and the new theory is incredibly complex. Under those circumstances perhaps you knew P before the rise of the new theory and can retain that knowledge through the heights of the new theory’s popularity. That sounds odd, ignorance being sufﬁcient to retain knowledge. I’m not insisting upon it; I’m only being generous to the anti-sceptic in making room for it. Of course, if the inconsistency is simple and is pointed out to you, and you come to justiﬁably believe that your old belief is inconsistent with the new theory, then given that you’re unable to do anything to cast doubt on the new theory, you don’t know P. In that case you should conditionalize your belief: your belief should not be that P is true; rather, it should be that P is true if the new theory is false. For it’s plain now that you don’t know that P is true; in this context ‘S knows P’ is false. Here’s a direct argument for the claim that in this story you don’t know that water is a compound of oxygen and something like hydrogen. In this scenario you believe this fact about water due to the expertise or authority of those teachers who told you so in school. Perhaps your teachers are experts who have studied water for many years. They are recognized as experts in the scientiﬁc community. Or perhaps your teachers are not experts but have studied with the experts. Everything looks pretty good for your true belief: you have a true belief and hold it for the right reasons. But you could just as easily have gone to a school populated with teachers who subscribe to

14

When Experts Make Mistakes

the water-as-atomic-element theory. Or you could easily have gone to a school that took both theories quite seriously and did not pronounce the water-as-compound theory established. If either of those possibilities had been actual, then you wouldn’t know that water is made up of oxygen and something like hydrogen. You might not have had that belief; or you might have acquired it, but clearly would not know it. Your belief in the compoundness of water would be far too lucky. In the actual world you acquired the water-as-compound belief via a certain belief-producing method—roughly, agreeing with your professionally respected and competent teachers—that could very easily have landed you with a false belief (that water isn’t a compound). In each of the scenarios—the actual one in which the teacher defended, a little, the compound hypothesis, the possible one in which she defended the element hypothesis, and the other possible one in which she defended both hypotheses and said it was still up for grabs as to which was true—you employed the same fairly reliable beliefproducing method. It’s just an accident that in the actual world you ended up with the true belief. It’s almost as though you were in a game show presented with three doors. If you go in door 1, you’ll be told P even though loads of experts say :P; if you go in door 2, you’ll be told :P even though loads of experts say P; if you go in door 3, you’ll be told that P’s truth is currently up for grabs. In each case you encounter numerous respected experts regarding P. If you happen to choose the door that leads you to getting the true belief, based just on the word of the people in that room, that hardly means that your true belief is knowledge. Too lucky; you could just as easily have ended up going through one of the other doors and encountering contrary but still expert opinion. The fact that there are experts in each room is crucial, for otherwise we are arguing implausibly against the power of testimony. If you ask me when I was born, I may tell you the truth (door 1), but it is easy to imagine the situation in which I fool you by intentionally giving you an incorrect answer (door 2) or I fool you by falsely telling you I was adopted and no one really knows for sure when I was born (door 3). Despite the existence of doors 2 and 3 (these are possible worlds), door 1 gives you knowledge, because in world 1 there is nothing whatsoever to be said in favour of what I tell you in worlds 2 and 3. The ‘hypotheses’ I tell you in worlds 2 and 3 are not remotely close to being live in world 1 (or 2 or 3). In the water-element possible world described in this essay, all three doors are highly

When Experts Make Mistakes

15

respectable, and numerous experts (actually in that world) endorse them (e.g., in that world there are many well-respected schools ﬁtting the proﬁle of either door 2 or door 3). It is sometimes said that one need not, in order to know P, rule out (or know the falsehood of) counter-possibility Q, provided Q is appropriately ‘irrelevant’.2 Maybe so, but it is hard to see how this would apply if Q were live in the manner described above and you were aware of the inconsistency of Q and P. If what I’ve argued above is wrong, then some hypotheses do not need to be ruled out for you in order for you to know P even though such hypotheses are live, you’re a mere mortal regarding them, many believe that they need to be ruled out, many think that they are true, they are inconsistent with P, and everyone agrees that they are inconsistent with P. Maybe the counter-hypothesis Q even keeps you up at night, worrying: ‘Water has got to be made of oxygen and something else like hydrogen [that’s P], but then again what if [Q, i.e.] Smith and Jones are right that water’s an atomic element?’ According to the envisioned objection, this doesn’t matter. The objection is not very plausible when we consider the conditions that obtain in this story: i. Most people believe P. ii. Everyone believes (correctly) that Q is inconsistent with P. iii. I’ve actually put together P and Q and I’m as aware as anyone that Q is inconsistent with P. iv. Q is a real, live contender in our intellectual community. v. I’m aware that Q is a real, live possibility actually endorsed by plenty of top experts. vi. Lots of people worry about their attitude towards P in light of what they think about Q. vii. Even the experts who think Q is rubbish would pretty much laugh at any reasons I could give against Q. If just conditions (i)–(iii) hold, then perhaps I don’t have to rule out Q in order to know P. Such a situation may obtain when P is ‘I got 2 I have in mind contextualist and relevant alternative theories. In my opinion, these theories embody the most exciting and promising innovations in epistemology over the last thirty years. Contextualist theories are offered in several excellent works: e.g., Lewis 1996; Cohen 1988, 1999; and DeRose 1995. Relevant alternative theories are found in the equally superb works of Dretske (1970, 1981), Stine (1976), Goldman (1975), Heller (1999), and others. Sometimes it is difﬁcult to know whether a theory is a contextualist theory, a relevant alternative theory, or both; but DeRose offers some good advice in his 1992.

16

When Experts Make Mistakes

these blue socks today for my birthday’ and Q is ‘I became a brain in a vat last year and have remained that way since then’. We might get the same answer if just the ﬁrst four conditions hold. That is, perhaps one could know P even without ruling out Q, provided one believed P for the ‘right’ reasons and was unaware that Q was a live possibility endorsed by experts. But when all seven conditions hold, this is implausible. If all seven conditions hold, then Q needs to be ruled out, as it is now a ‘relevant alternative’ to what I believe; I am ‘epistemically responsible’ to Q. I personally think that the truth of each of (v)–(vii) makes me more responsible to Q than if just (i)–(iv) hold, but you don’t have to agree with that judgement in order to agree with my central claim: when all seven hold, the person fails to know P unless she can rule out Q. Here is one more argument. Suppose (i)–(vii) hold, but you continue to believe P even in the recognition that Q is live and P conﬂicts with Q. You know that if P is true, then Q is false; you’ve actually thought this through. And you have reﬂected on the fact that you believe P. So you believe :Q, as you can see as well as anyone that P’s truth straightforwardly rules out Q—where Q is the real, live hypothesis endorsed by experts, etc. If what I’ve argued above is wrong, so you know that P, you know that if P, then :Q, and you’ve thought this all through properly, then it seems that you know :Q, especially after you go through this reasoning. So you know that the live hypothesis is false—even though there are loads of experts better positioned than you who don’t, you’re no genius regarding P or Q, you are as aware as anyone that these experts disagree with you and are real experts, and no expert would take seriously your reasons against Q. This seems ridiculous. Let me close this chapter with not an argument but an appeal to whatever portion of vanity you happen to possess. Suppose—what might actually be true—that you have spent a decade or so formulating, defending, and revising an elaborate philosophical theory. You and many others have discussed the theory at major conferences and in major journals. Your theory is sometimes the topic of Ph.D. dissertations. You have a number of followers. Other top investigators respect you and your theory. Your theory does contradict common sense in some ways, but of course many elaborate, well-developed philosophical theories have that feature. Now suppose a ﬁrst-year graduate student in philosophy, who is nothing to get excited about with her string of B grades, says you’re

When Experts Make Mistakes

17

wrong, even though it’s perfectly plain that she is utterly incapable of even coming close to knocking your theory down. She has the common-sense belief that P, and like everyone else she knows that P is straightforwardly inconsistent with your theory, and so she believes that your theory is wrong. Unbeknownst to you and your army of supporters, your theory really is wrong. If the new scepticism is wrong, then she knows P and she knows you’re wrong. The ﬁrstyear student with no interesting philosophical or scientiﬁc ability knows that your theory is wrong even though she couldn’t write an even half-way decent paper on the topic. So why is it that you have tenure anyway? More to the point, if she knows P and deduces on that basis that your theory is wrong, if it’s really that easy, then why should we continue engaging in this silly enterprise of philosophy?

3

How Live Hypotheses Sabotage Knowledge All democrats are insane, but not one of them knows it; none but the republicans and mugwumps know it. All the republicans are insane, but only the democrats and mugwumps can perceive it. The rule is perfect: in all matters of opinion our adversaries are insane. Mark Twain, What Is Man and tOher Philosophical Writings

The considerations of the previous chapter are somewhat crude, but they are nevertheless adequate to motivate a valid sceptical argument template. In this chapter I present and attempt to clarify it. In subsequent chapters I plug various live sceptical hypotheses into the template. First I must introduce some terminology. I do this not to torture you. It’s just that the live sceptical conclusions are startling enough that we need to be precise with the premisses. I will be considering hypotheses that are live for an epistemic community, hypotheses that have not been ruled out for someone, people who are mere mortals regarding those hypotheses, and people who are about as aware as anyone of the fact that P. I will be using each of these italicized notions in what I believe to be an entirely intuitive way, but it is best to spell out some of the details involved. By a hypothesis H that is ‘live for an intellectual community’ I mean a hypothesis that satisﬁes the following ﬁve conditions: i. In your intellectual community H has been through a signiﬁcant (not to say exhaustive) evaluation by experts over many years. ii. It is judged actually true or about as likely as any relevant possibility by a signiﬁcant number of well-informed, well-respected,

How Live Hypotheses Sabotage Knowledge

19

non-crackpot, and highly intelligent genuine experts in the ﬁeld(s) to which H belongs. iii. Those experts reached their favourable opinion based on H’s merits in a familiar, epistemically responsible way (i.e., the way they reached their opinion seems about as epistemically responsible as any). iv. Those experts consider there to be several decent and independent sources of evidence for H. So it’s not the case that the only reason people pay serious attention to H is the presence of one weird experiment or line of reasoning. v. Many of those experts consider H to be a ‘real, live possibility’ (i.e., that’s what they’d say if you asked them).3 None of this says that H is true. Metaphysically speaking, it may well be nowhere in the vicinity of true. My only point is that it is currently in the intellectual community a live socio-epistemic possibility in the sense that it meets those conditions. They are intended as providing a sufﬁcient, not necessary, condition for a sociological notion of ‘live’. I suspect that weaker conditions would sufﬁce, but I won’t need to defend that claim in this essay. In saying that H isn’t ‘ruled out for you’, I mean that you have not actually ruled it out yourself; you are unable, in your current state of abilities and information, to do it yourself; and no one has actually done it for you in the sense that they did it and somehow the rulingout status was transferred to you. There are two obvious questions to ask about this notion: What does it mean to rule something out yourself? And what does it mean for ruling-out status to be transferred from one person or community to another? I can’t answer the second question, but I did provide examples of such transfers in the previous chapter. The ﬁrst question will not receive a thorough treatment until Chapter 9. For now, to rule out H is to neutralize the threat that H poses to your :H beliefs. Prima facie, this neutralization may come in two forms: I can have some warrant that defeats H head-on, so to speak; or my belief can be qualiﬁed in some way so that H doesn’t actually pose a threat to my :H belief. By ‘well-informed mere mortal regarding the issues surrounding live H’ I mean, roughly put, that you are aware of H and its live status but you are no genius about it. A little more precisely: 3 I consider adding a sixth condition in Ch. 11, sect. B, on the ‘Double Standard’ Solution.

20

How Live Hypotheses Sabotage Knowledge

a. You’re about as aware as anyone of the fact that the hypothesis H is inconsistent with the proposition P alleged to be known. b. Your intelligence, understanding, and knowledge are not extraordinary for people in your intellectual community with regard to H or the relevant issues surrounding H. (So although you may be an expert, you’re not much better than the other experts—in particular, the ones who insist that H could very well be true.)4 c. You are at least somewhat familiar with H and the issues surrounding H, including the fact that H is live in the above sense. That is, you’re about as aware as anyone of the fact that H is live— i.e., that it satisﬁes liveness conditions (i)–(v).5 d. If you have any reasons or evidence you can marshal for casting doubt on the hypothesis, and if they were carefully considered by the members of that community—in particular, by the wellinformed, well-respected, non-crackpot, and highly intelligent experts who were thoroughly familiar with the hypothesis—they would be nearly universally and conﬁdently rejected as clearly insufﬁcient to rule out the hypothesis (although they may have other merits).6 Notice that (a) and (c) are quite different from (b) and (d). The former pair says, crudely put, ‘You know what’s going on with the live hypothesis, so it’s a real threat to your knowledge, a threat that must be defused’. The latter pair says, again crudely, ‘You don’t have what it takes to defuse the threat’. 4 Matters are somewhat more favourable to the live sceptic if the mere mortal isn’t an expert. 5 Eventually I’m going to be arguing, roughly, that if H is live and you’re a mere mortal regarding H, then you don’t know things inconsistent with H. In part (c) of the above characterization of mortality I have the subject aware of the live status of H. I am not suggesting that one must be aware of the liveness of H in order for H to pose a threat to one’s knowledge of claims inconsistent with H. When I give sufﬁcient conditions, I don’t suggest that they are necessary as well. 6 It is easy to imagine borderline cases: cases in which it isn’t clear whether someone is a mere mortal, or whether the hypothesis is really live, or whether there really is a community of experts. When thinking about such cases, one quickly ﬁnds potential problems that might suggest that there is something ﬁshy with the sceptical argument. For instance, it’s clear that a live hypothesis doesn’t lose its live status if all its advocates die in a plane crash; so my liveness conditions are sufﬁcient, not necessary. (Don’t try to crawl your way out of the live sceptic’s trap by killing the advocates of the live hypotheses that go against your beliefs!) But the same potential problems occur for other, uncontroversial scientiﬁc and non-philosophical cases; and the potential problems don’t apply in many of the actual or possible cases considered in this essay. So if we like, we can conditionalize some premisses to omit the potential problems.

How Live Hypotheses Sabotage Knowledge

21

Pay particular attention to (d), as it is formulated in such a way as to block the following objection: we are always going to ﬁnd people who consider our arguments insufﬁcient, no matter how good they are. Most philosophers think that Kripke’s criticisms of the ‘famous deeds’ description theory are decisive, but we will always be able to turn up some dissenters. The problem here is that it’s virtually impossible to be an immortal. Yet surely one need not be a supergenius debater in order to rule out certain hypotheses. The last sentence is correct, but the objection is based on a misunderstanding. To be a mere mortal doesn’t just mean that we could ﬁnd some people who disagreed with your arguments against the sceptical hypotheses. It means that your reasons would be ‘nearly universally and conﬁdently rejected as clearly insufﬁcient to rule out the hypothesis’ by the relevant experts. This does not mean that your reasons would be no good or uninteresting or very inferior to anyone else’s. It just means that people would justiﬁably say that you’ve done very little, yet anyway, to put the hypotheses away. So (d) is not merely saying that you can’t convince experts that H is false. Condition (d) doesn’t just mean that you lack reasons that would be accepted as establishing beyond serious dispute that H is false. No, it’s saying that even the experts who reject H would say that your reasons for rejecting H are clearly inadequate by anyone’s lights, not just by the lights of those who accept H. For instance, you might be a graduate student in palaeontology who is aware of the rival hypotheses about the demise of the dinosaurs and who happens to believe the true meteor hypothesis. You go to see your Ph.D. supervisor and she asks you what you plan to say about the supervolcano hypothesis in your dissertation. You say that this theory isn’t very plausible, but you’re happy to throw in a brief section showing why it’s wrong. She agrees with you that the meteor hypothesis is correct, but she asks you what you plan to say against the supervolcano hypothesis. You give your spiel and she tells you ﬂat out that what you’ve said is clearly inadequate and you should either do much better with a critical section or drop it entirely and say in a footnote that you’ll merely be assuming the falsehood of the supervolcano theory. After all, Professor So-and-so right down the hall is a supervolcano advocate, he’s certainly no dope, he isn’t alone in his expert contrary opinion, and you’ve said nothing to put much pressure on his view. The graduate student is a well-informed mere mortal, ‘mere mortal’ for short. A child is not, as she fails to satisfy either (a) or (c).

22

How Live Hypotheses Sabotage Knowledge

Another kind of mere mortal is an expert in the ﬁeld whose specialization lies elsewhere. Professor Smith teaches various science classes, including palaeontology, at a small college. She is perfectly aware of the supervolcano and solar ﬂare hypotheses and wouldn’t be able to say anything interesting against them. She has the true meteor belief but, as with the graduate student, her belief is too lucky to amount to knowledge. That’s the type of person I have in mind as a mere mortal. Needless to say, there are many such individuals. Also keep in mind that the experts with whom the mere mortal disagrees are often her epistemic superiors, not her epistemic peers. The super-famous supervolcano advocate was in the best Ph.D. programme, graduated top of her class, published in the best places from the very beginning of her career, etc. By ‘epistemic superior’ I mean that such a person is generally more informed than we mere mortals are on the topics involving H; she has more raw intelligence; she has thought and investigated whether H longer and in more depth than we have; she has thought and investigated the topics surrounding H longer and in more depth than we have; she is just as, or even more, intellectually careful than we are; she is just as, or even more, intellectually honest than we are; and, crucially, we admit that she has understood and fairly and thoroughly evaluated the same evidence as we have—and usually more evidence. And yet we mere mortals are perfectly aware that we are saying that this person is wrong. We could divide the mere mortals into three groups: those who admit that many of the advocates of H are their epistemic superiors, those who deny that epistemic superiority, and those who have never given the superiority question even a passing thought. In addition, we could separate mere mortals into those who admit that they satisfy (b) and (d) and those more conﬁdent ones who deny those truths. My feeling is that the new sceptical results are most plausible when targeted at those mere mortals who admit that many of the advocates of H are their epistemic superiors, but I won’t pursue that thought in this book. By ‘about as aware as anyone of the fact that F’ I mean to convey the idea that you have as much positive epistemic warrant as just about anyone for your belief that F. For instance, say you are an average graduate student in physics, and are aware in the usual way of the liveness of the theory that says that there is lots of dark matter. Or you are in philosophy, and are aware of the liveness of a context-

How Live Hypotheses Sabotage Knowledge

23

ualist account of ‘S knows P’—again, in the usual way. In being ‘as aware as anyone’ that H entails :P, I mean that once again you are aware in the usual upstanding way that, e.g., the claim that Casper the ghost exists entails the falsehood of the claim that there are no ghosts; or the claim that water is an atomic element entails the falsehood of the theory that water is made up of the elements hydrogen and oxygen. In saying that you are as aware as anyone that F, I am not saying that you know F. I don’t want to suggest that the awareness in question amounts to knowledge, because we will be looking at some arguments that entail that many of us, but not necessarily all of us, don’t know those facts even though we have all the usual good reasons for believing them. This is getting ahead of ourselves; but, very brieﬂy, the reason you don’t know P even though you have all the usual good reasons for believing the truth P is that you are privy to some good evidence against P, evidence that you appreciate as good and well respected and that is strong enough to undermine your knowledge. Thus, you retain all your good, positive warrant for P, but no longer know P because of the evidence against P that you become aware of, appreciating its strength as good evidence against P. Condition (c) says that the mere mortal is somewhat familiar with H and the issues surrounding H, including the fact that H is live. How familiar? I want the new sceptical arguments to have fairly wide application, so that many of us count as mere mortals. So I don’t want to demand detailed knowledge of the reasons why H is held in esteem. Neither do I want to demand only that one has heard that there’s some theory out there, H, that some people think about and take seriously. The appropriate demand is that the mere mortal must be as aware as anyone that (i)–(v) hold for H. The mere mortal might not know why exactly so many experts believe H is true; she might not understand any of their reasons, although it helps if she has a passing acquaintance with some of them. What she is aware of is that these are genuine experts—as expert as just about anyone in the area—who after careful reﬂection and criticism endorse H. Keep in mind that it isn’t the mere fact that a signiﬁcant number of experts endorse H that robs the mere mortal of knowledge. Rather, it is her awareness of this endorsement that is doing the work. Perhaps this requirement will need tweaking, but I don’t think it will affect the argument negatively. Many theorists have worked with the notion of relevant alternatives (see n. 2). If you believe P, and Q is a relevant alternative to

24

How Live Hypotheses Sabotage Knowledge

P—roughly put, a :P possibility that isn’t like the BIV possibility in being either completely empty or without meaningful support— then, in order for you to know P, you must be able to rule out Q (or know Q’s falsehood). Obviously, liveness entails relevance, but relevance doesn’t entail liveness. Suppose Professor X gets a sexual harassment suit against him. He leaves his present university in order to avoid the problem and takes a post at another university in another country. Suppose, further, that you knew all about the incident in question and ﬁrmly believe that although he was stupid in what he did, it didn’t amount to sexual harassment by the legal deﬁnition, or even by the deﬁnitions employed by the persons involved in the incident. You ﬁrmly believe that P: Professor X will never do any serious sexual harassing. Then you ﬁnd out that at his new university there are two sexual harassment suits against him. So Q is now a relevant alternative by any standard of relevance, where Q is ‘The presence of these lawsuits shows that Professor X did do some serious sexual harassing at his new university’. In order for you to know P, you must deﬁnitely rule out Q. Even so, Q need not be live. We can stipulate that any expert would be able to tell you that Q is obviously false. The two new lawsuits are blatant witch-hunts with virtually no basis whatsoever, as anyone involved in the case will admit (even the people ﬁling the lawsuit know that they are merely trying to ruin Professor X and that no harassment occurred). So the notion of liveness is in a clear sense stronger than that of relevance. However, relevance is clearly highly context-dependent. What possibilities are relevant clearly depends on the individual involved. Liveness is a bit more objective, as it is generated from considerations involving one’s epistemic community. The connection between liveness, relevance, and context is delicate, however, and will not be probed until Chapters 9 and 10. Now we can present the general argument template to be used to generate most (not all) of the live sceptical puzzles. A. The Modesty Principle. If you are a mere mortal regarding some live hypothesis H, then H isn’t ruled out for you. In brief: if S is mortal, then live H isn’t ruled out. B. The Live Hypothesis Principle. There is a crucial connection between a live hypothesis and the prospect of someone knowing some truth P. If S is as aware as just about anyone that H is live (so H really is live) and that P entails :H, then if H isn’t ruled out

How Live Hypotheses Sabotage Knowledge

25

with respect to S, S doesn’t know P. If S is a mere mortal with respect to H, then the antecedent is true. So the Live Hypothesis Principle comes to this: if S is mortal and H isn’t ruled out, then S doesn’t know P (for the Ps meeting the antecedent conditions). C. So, by (A) and (B) we know two things: if S is mortal, then H isn’t ruled out (that’s (A)); and if S is mortal and H isn’t ruled out, then S doesn’t know P (that’s (B)). It follows from these two that if S is mortal, then S doesn’t know P. That is, the mere mortals among us don’t know any P provided we are as aware as anyone of the truth that H is live and inconsistent with P. D. The Mere Mortal Premiss. In the actual world or a very close possible world many members of our actual intellectual community of contemporary analytic philosophers and cognitive scientists satisfy mortality conditions (b) and (d) for some Hs (the speciﬁc philosophically interesting Hs to be given below). Furthermore, it’s obvious to nearly all of us that an enormous number of our most conﬁdently held beliefs are inconsistent with H (again, for the speciﬁc Hs given later); it’s also pretty obvious to many of us that H is live. Those are mortality conditions (a) and (c). Thus, in the actual world or a very close possible world many of us philosophers and scientists are mere mortals regarding H—especially those who read this essay. E. Thus, by (C) and (D) for each of the many Ps obviously inconsistent with H, we mere mortals don’t know P in the actual world or a very close possible world. Perhaps the ‘immortals’ among us have that knowledge, and maybe even the sufﬁciently ignorant and isolated have it, but we do not. In the classic sceptical arguments, the principle corresponding to the Live Hypothesis Principle often is the simple ‘If S knows that P is inconsistent with H, then S has to rule out or even know the falsehood of H in order to know P’ (or: ‘If S knows that P and knows that P entails :H, then S knows that :H’). Call this the simple principle. There is no requirement that S has explicitly seen the inconsistency of H and P; there is no requirement that H be live; neither, of course, is there the requirement that S be aware that H is live. These three omissions signiﬁcantly diminish the appeal of the simple principle. Indeed, some philosophers, such as Nozick (1981)

26

How Live Hypotheses Sabotage Knowledge

and Dretske (1970, 1981), have rejected the simple principle.7 However, what they say against the simple principle does not, in my opinion, give any reason at all to reject the Live Hypothesis Principle. The conclusion (E) is that all the mere mortals in an intellectual community don’t know any P provided P is inconsistent with H. Notice that it does not apply to all people (only mere mortals aware of certain facts) or all propositions (only ones competing with live hypotheses) or all times (for the hypothesis could end up ruled out). I have nothing precise in mind with my use, in characterizing mere mortality and liveness, of ‘intellectual community’, ‘relevant issues’, or some of the other terms. However, most of this sloppiness will not matter, since my uses of the argument will make it clear enough what I have in mind. Where it does matter, I will attempt to provide greater precision. Henceforth I assume that this argument template (A)–(E) is pretty solid. In it I have attempted to capture some home truths about the possibility of knowledge in the face of respected competing hypotheses. I have tried to stick to common sense as best I can. The argument is supposed to capture part of our ordinary conception of knowledge, rules we go by in ordinary life. In particular, the Live Hypothesis Principle does not presuppose that knowledge is or must be ‘high standards’ knowledge. Nothing like certainty is involved in the argument; the amount of epistemic warrant denied in the conclusion is entirely modest. The conclusion is consistent with the existence of some warrant for the beliefs of mere mortals in the truth threatened by the live hypothesis, but it does say that whatever warrant there is is not sufﬁcient for knowledge—ordinary knowledge, not certainty. Even if the conclusion (E) is true, there may be dribbles of warrant all over; they just never meet our ordinary standards for knowledge.8 Let me emphasize that when I ask, ‘Does 7 I have as well, in Frances 1999, but for completely different reasons. As far as I can see, even if I’m right about the falsehood of all versions of epistemic closure actually formulated, there are others that escape my counter-examples and that can be used in sceptical arguments. 8 Even though I will sometimes use the phrase ‘warranted enough’, I am not implying that the difference between a belief having some warrant and being warranted is a mere matter of piling on more warrant. A convenient picture would be that beliefs have units of warrant. A modest number of those units, say 233, means that your belief has some nontrivial warrant. And if you only had another 767 units, your warrant pile would reach the ‘warranted belief ’ threshold. But this is a fantasy, albeit useful. Other times I will write of high- or low-quality warrant.

How Live Hypotheses Sabotage Knowledge

27

she know that P?’ in this essay, I am asking an ordinary question. I’m not saying, ‘Does she know that P in the philosophical sense?’, whatever that means. Eventually I’ll be asking questions such as ‘Does she know that ﬁre-engines are red?’, ‘Does she know that she has a pain in her chest?’, and ‘Does he know that Moore believed that scepticism is false?’ I will be trying to determine whether by ordinary standards the answer is ‘yes’. Think of the standards that are in play when one asks of a foreigner the question about ﬁre-engines, or the standards in force when we ask of a student the question about Moore’s beliefs. Those are the perfectly ordinary standards and contexts that I’m interested in. In my judgement, if an enormous and interesting set of perfectly ordinary and usually considered highly secure knowledge claims are in fact false, then we have a philosophically interesting kind of scepticism. The questions ‘Does this belief have any epistemic warrant at all?’ and ‘Does this belief amount to a certainty?’ are not the only two interesting sceptical questions, and are not our focus. What is at issue here, when in Chapters 4 and 5 we apply the above argument schema to the live sceptical hypotheses, is whether a belief that almost every sensible person would insist has more than enough warrant to turn a true belief into knowledge really is so warranted. Some epistemologists give the impression that scepticism is boring if it challenges just ‘S knows P’ and not also ‘S has a true belief with some justiﬁcation or warrant’, but if it can be shown that you, the intelligent philosopher, don’t know any of the following, despite being in what are universally considered to be excellent positions to know each of them, then something of considerable philosophical signiﬁcance will have been shown. Your socks are blue. Moore believed that scepticism is false. You have a pain in your knee. John Rawls was kind. In short, if you don’t know those ordinary facts, then your approach toward scepticism, if not epistemology generally, must be altered. At certain points in the essay we will have reason to ask, ‘Even though she knows P, does this knowledge amount to any kind of epistemically signiﬁcant accomplishment?’, which is a very different question and one justly termed ‘philosophical’. We will return to an

28

How Live Hypotheses Sabotage Knowledge

investigation of the kind of scepticism generated by the live arguments in Chapters 6 and 8. No one today can argue for any kind of scepticism without directly confronting theories that insist on the epistemic import of certain contextual factors: e.g., relevant alternative and contextualist theories.9 If you are interested in any of those approaches to scepticism, as you should be, then you will want to examine versions of the above principles where uses of ‘S knows P’ are relativized to certain contexts. (In effect, the restrictions to live hypotheses and mere mortals already carry a great deal of information regarding relevance and context.) In Chapters 9 and 10 I will pay due, probably excessive, attention to contextual matters. Until then we can keep with the version given above. The argument is pretty detailed, so on inductive grounds I am all but certain that there are problems with (A)–(E) and with my characterizations of ‘mere mortal’, ‘live’, and ‘as aware as anyone’ even when applied to ordinary, philosophically uninteresting cases. But whatever these problems are, I suspect that (A)–(E) can be repaired in such a way that the argument is not undermined. Philosophers revel in the construction of counter-examples, and rightly so, but the point of the enterprise is lost if no effort is made to see if the claim thus falsiﬁed can be altered and retain its philosophical importance. If you’re not satisﬁed with the principles given above, repair them as you see ﬁt; why should I do all the work? However, there is a more serious point here. Philosophers rarely make their arguments complete and valid, despite their verbal support for argumentative rigor. I have tried to do so, and so I expose myself to a familiar problem of using universal generalizations: no matter how intuitive a universal generalization is, if it isn’t a logical truth, then there is the possibility that a counter-example is waiting to be found, provided we come up with an instance that is strange enough. I will try to defend my instances of the generalizations as I go. In this chapter I have done four things. First, I tried to indicate what I mean by certain key terms: ‘live’, ‘mere mortal’, ‘ruled out’, and ‘about as aware as anyone of the fact that P’. Eventually, in Chapter 9, the notion of ruling out will be replaced with something a bit less 9 Proponents of such theories will most likely be wary of the Live Hypothesis Principle. See Pritchard 2002 for many references and a nearly up-to-date presentation of the state of play regarding traditional scepticism.

How Live Hypotheses Sabotage Knowledge

29

vague. I then brieﬂy compared live hypotheses with relevant alternatives, again leaving more sophisticated discussion for Chapter 9. Then I presented the all-important live sceptical argument template. Finally, I offered some initial discussion of the nature of the live sceptic’s conclusion, which will be further discussed in Chapters 6 and 8. Now we proceed with the ﬁrst application of the template.

4

The First Live Sceptical Hypothesis It is a sign of the immature state of psychology that we can scarcely utter a single sentence about mental phenomena which will not be disputed by many people. Franz Brentano, Psychology from an Empirical Standpoint It is genius, and not the want of it, that adulterates philosophy, and ﬁlls it with error and false theory. Thomas Reid, An Inquiry into the Human Mind

In this chapter I will articulate the ﬁrst of the live sceptical hypotheses, set out the corresponding sceptical argument generated from that hypothesis, and offer some initial defence of the argument. This is not the entire defence; the rest of the essay constitutes the whole case. In particular, Chapter 6 gives the sceptical solution to the philosophical puzzle generated by the live sceptic’s argument. But what appears here should clarify the basics of the ﬁrst argument as well as the template. A The First Hypothesis and Argument Now, ﬁnally, we reach the ﬁrst of the many live sceptical hypotheses. Near enough, it’s eliminativism from the philosophy of mind: all occurrences of ‘S believes P’ are false. ‘Eliminativism is highly dubious,’ you say. I agree! ‘So’, you continue, ‘whatever the ﬁrst new sceptical argument is, it will be highly dubious as well. Now I can stop reading this essay and ask for my money back.’ Unfortunately not, because the sceptical argument does not employ eliminativism as a premiss. How it employs eliminativism will be made explicit below, but it needs to be said straight off that in this essay I assume that all versions of eliminativism are false and that

The First Live Sceptical Hypothesis

31

we have millions of propositional beliefs. Odd as it may sound, I’m even willing to admit that some readers may have actually refuted eliminativism outright—and they even know, in the best way possible, that it is false. I know this sounds strange in an essay presenting an argument for scepticism that employs premisses concerning eliminativism; all will be revealed as we go. The eliminativist sceptic—who accepts not eliminativism but the soundness of the eliminativist sceptical argument given below and generated from the argument template of the previous chapter—can accept all these assumptions as well. The point is that even on these assumptions the new sceptical argument strongly suggests that we—or a large group of intellectuals very likely including you—actually have no propositional knowledge of either knowledge or belief; that is, for humans all uses of ‘S knows that R knows/believes that P’, where S and R are among us, are false. The eliminativist sceptic need not endorse or even take seriously any version of eliminativism; she may have even refuted it and know its falsehood. So you can’t ignore the arguments just because you have a very dim view of eliminativism. I know that many excellent philosophers simply turn a deaf ear to eliminativism, which is why I vigorously emphasize this point before getting to the sceptical argument itself. There will be a bit more on versions of eliminativism below. Throughout every paragraph of this essay I’m assuming that eliminativism is false; the eliminativist sceptic can do this as well, with perfect consistency. For instance, some of the conditions for liveness/ mortality entail the falsehood of eliminativism, but that is not a problem, because the eliminativist sceptic is not an eliminativist. Even an arch-anti-eliminativist such as Jerry Fodor could be an eliminativist sceptic. I must drone on emphasizing this point because several commentators continually miss it. For instance, one might think that if my sceptical argument fails, then we have a reductio of eliminativism. But, once again, eliminativism is not a premiss and simply cannot be affected by my argument, regardless of its quality. The core of the eliminativist sceptical argument is that although there is no convincing case for the truth of eliminativism, there is a compelling argument for its status as a live possibility that must be ruled out in order to have knowledge of belief. The eliminativist sceptical argument trades on that fact. As you have guessed, the eliminativist argument is little more than (A)–(E) with the eliminativist hypothesis plugged in.

32

The First Live Sceptical Hypothesis

We begin with a two-part deﬁnition of the eliminativist hypothesis. (i) There are cognitive states in our minds/brains that do the work in successfully navigating us around the world, but they aren’t propositional beliefs (or ‘acceptances’; see below). Our physical cognitive architecture shows that none of our cognitive states satisﬁes the necessary conditions for being a belief. So all uses of ‘S believes P’ are false for us. Furthermore, (ii) propositional knowledge requires propositional belief (acceptance): if ‘S knows P’ is true, then ‘S believes (accepts) P’ is true. Now for the argument. 1. The Modesty Principle. From before: if S is mortal, then live H isn’t ruled out. 2. The Live Hypothesis Principle. From before: if S is mortal and H isn’t ruled out, then S doesn’t know P (where P is obviously inconsistent with H, S is aware of that fact, etc.). 3. So by (1) and (2), if S is mortal, then S doesn’t know P. 4. The Mere Mortal Premiss. In the actual world or a very close possible world many members of our actual intellectual community of contemporary analytic philosophers and cognitive scientists satisfy mortality conditions (b) and (d) for the eliminativist hypothesis. Furthermore, it’s obvious to most of us that the contents of second-order beliefs—i.e., truths of the form ‘R believes that Q’ in beliefs of the form ‘S believes that R believes that Q’—are inconsistent with the eliminativist hypothesis. It’s also pretty obvious to many of us that the eliminativist hypothesis is live, especially if you keep reading this essay.10 Those are mortality conditions (a) and (c). So all four conditions (a)–(d) for mortality are satisﬁed, and the eliminativist hypothesis is live. That is, many philosophers and cognitive scientists are, in the actual world or a world next door, mere mortals when it comes to eliminativism (i.e., all nine conditions for liveness and mortality given in the previous chapter are satisﬁed). 5. Thus, by (3) and (4), for each claim P of the form ‘R believes Q’, we mere mortals don’t know P. That is, we don’t know what anyone believes, ourselves included. The eliminativist sceptical argument is complicated, but the governing intuitions can be put into simple and roughly accurate terms as follows. Suppose there were a live, scientiﬁc, philosophical, and 10

Almost. See subsequent discussion.

The First Live Sceptical Hypothesis

33

explanatory hypothesis about a class of entities E. It said that all the Es don’t have property P. If we were aware of all that, then until we could rule out that hypothesis, surely we mortals would not know that some of the Es do have P. That seems straightforward. The average person on the street who has never heard of the live hypothesis might be able to get away with knowing that some Es have P, but we who know about the hypothesis and who appreciate its respected live status do not. Now let the Es be our cognitive states and the property P be that of being a belief state. So the corresponding hypothesis is that the Es don’t have P; that is, our cognitive states don’t amount to belief. Since the hypothesis is live, by the reasoning in the previous paragraph, we don’t know that the Es are P; that is, we don’t know that our cognitive states (Es) are states of belief (P). We don’t know that we believe. Thus, even if all the traditional sceptical arguments fail for some set of beliefs, and many of those beliefs amount to knowledge, the principle ‘If you know P, then you (implicitly or dispositionally or explicitly or whatever) know that you believe or know P’ fails for each element of that set—no matter what P is and no matter how hard you try to know that you believe or know that P (short of refuting eliminativism outright). Thus, for all that this eliminativist sceptical argument shows, you may know that you have two hands, but you don’t know that that’s what you believe. You do believe that you have two hands; and you do believe that you believe that you have two hands. And you know that you have two hands (at least, that possibility is left open by this sceptical argument). But your second-order belief, although true, isn’t warranted enough for knowledge; you don’t know what you believe. And of course you don’t know what anyone else believes either—even though you have loads of true beliefs about their beliefs. That’s bizarre. I don’t have to defend the signiﬁcance of such a sceptical thesis.11

11 If this argument is sound, then it’s clearly very important and should be published. On the other hand, publishing it will only lead to more people, epistemologists mainly, being appropriately acquainted with eliminativism (and the other sceptical hypotheses to be articulated below), thereby increasing the scope of people left without certain large and interesting kinds of knowledge. So if the argument is sound, then it shouldn’t be published: it will lead to large masses of philosophers losing knowledge. Thus, if the argument is sound, it both should and should not be published. So . . .

34

The First Live Sceptical Hypothesis

The argument nowhere assumes that anyone actually knows any of (1)–(5). Neither does it assume that anyone actually believes or goes through the reasoning in (1)–(5). All it relies on is the truth of (1)–(5). If the argument is sound, then we mere mortals don’t know premiss (4), that many of us are mere mortals, because in order to know that fact, we would have to have knowledge of each other’s beliefs. Perhaps the immortals among us with respect to eliminativism could know each of (1)–(5), but the rest of us can’t. The mere mortals don’t have to be aware of or believe or know any of (1)–(5) in order for them to be true. There are several ways of elaborating eliminativism. What I have in mind will be elaborated below. The other new sceptical hypotheses, which generate the other new sceptical arguments, are reserved for Chapter 5. Often I will use ‘eliminativism’ instead of ‘eliminativist hypothesis’, since the eliminativist part of the hypothesis is the key. Although I have structured the argument with the sceptical conclusion (5), this is an arbitrary choice. The argument presents us with a puzzle about knowledge and warrant: a set of individually plausible yet apparently mutually inconsistent claims. The claims consist of the premisses, the common-sense idea that we mere mortals have loads of knowledge of belief and knowledge, and the claim that the eliminativist sceptical argument is valid. Any adequate solution to the puzzle, sceptical or otherwise, must identify the false claims in this set, explain why they are false, articulate the truths they are mistaken for (if any there be), and explain how it is that we made those mistakes—anything less is inadequate. Those are four constraints on an acceptable solution to the puzzle. (This holds for all such puzzles, not just sceptical or epistemological ones.) In this book I sometimes write as if the conclusions of all sceptical arguments are mistaken, and we know all sorts of things, but that’s due to force of habit. I will also often proceed as an advocate of each of the live sceptical arguments, someone endorsing the soundness of those arguments, but almost nothing of what follows represents my own opinion. I’m not the live sceptic; I’m the sceptic’s lawyer, prosecuting knowledge and defending the importance of the live sceptical arguments. There is no point in endorsing proposed solutions to the live sceptical puzzles unless one is conﬁdent that some proposal meets the four constraints, and I have no such conﬁdence in any proposal I have been able to think of (although I consider the sceptical proposal given in Chapter 6 to be quite promising).

The First Live Sceptical Hypothesis

35

There are three premisses: (1), (2), and (4). I have already given some arguments for (1) and (2). The only other premiss is (4), the Mere Mortal Premiss that many of us are mere mortals regarding the eliminativist hypothesis—which means that all nine conditions in the characterizations of ‘live’ and ‘mortal’ (ﬁve for ‘live’, four for ‘mortal’; see previous chapter) are satisﬁed for many of us. I will argue for (4) in this chapter. Notice that that premiss isn’t that eliminativism is true ; neither will I ever assume that knowledge requires belief.12 I want to make a couple of comments on part (ii) of the eliminativist hypothesis—‘propositional knowledge requires propositional belief (acceptance)’. There have, of course, always been relatively minor objections to this claim. But these won’t matter here. Surely knowledge requires something along the lines of ‘acceptance of the truth of ’: at the very least, it’s true in both our world and nearby worlds that we don’t get knowledge without belief/acceptance. And the eliminativist sceptic can successfully utilize this weak notion of ‘acceptance’. Furthermore, it won’t do to object that while propositional beliefs or acceptances don’t exist, propositional knowledge does. That is, it is unpersuasive to say that although some kind of cognitive brain state can, in conjunction with something like truth and epistemic entitlement, sufﬁce for knowing that P, no such state sufﬁces for believing (accepting) that P. For if some state successfully helps to make ‘S knows that P’ true (when conjoined with P’s truth and some appropriate epistemic warrant), then surely some state— probably the same one!—is good enough for helping make ‘S believes or accepts that P’ true. So one cannot be an eliminativist about belief/ acceptance and a realist about propositional knowledge. Maybe that’s wrong. In any case, the live sceptic does not have to take a hard line here. The hypothesis that says that there is no (or little) propositional knowledge without propositional belief (acceptance)—that’s part (ii) again—is surely more plausible than the hypothesis that says that there is plenty of propositional knowledge but no propositional beliefs (acceptances). This essay focuses on the former hypothesis, but for the sake of familiarity, I will use ‘belief ’ instead of ‘acceptance’. 12 What if eliminativism is true? Then no one knows anything, but no one has any true beliefs either—or even false beliefs. It isn’t clear how either scepticism or anti-scepticism applies to such a bizarre situation. Cf. n. 20 below.

36

The First Live Sceptical Hypothesis

Now for a partial clariﬁcation of part (i) of the eliminativist hypothesis. Some philosophers think that instances of ‘S believes that P’ can be true even if there are no beliefs in the head or soul or anywhere else. According to them, the truth of ‘S believes that P’ is consistent with eliminativism for belief tokens, which are supposed to be (in humans) neurophysiological particulars—not properties or other sharable items—that satisfy ‘S’s belief that P’, and are partially individuated by and literally in believers. John has a tremendous lack of skills; the average family has 2.4 children; there is a beef shortage in Albania. In order for those claims to be true, it seems that there need not be particular things that are lacks of skill, average families, or beef shortages. In order for those claims to be true, the world has to be some way, for sure; but we cannot just read off what that way is from the surface structure of the sentence. Similarly, it is sometimes thought that we can have beliefs without belief tokens in the above sense. Contrary to this view, the eliminativist hypothesis as it functions in the live sceptical argument is not just a metaphysical denial of human belief tokens, but includes the semantic claim that ‘S believes that P’ is never true for humans. This is where we are so far. The eliminativist hypothesis seems to be a live one (much more on that claim below). Given its status as live, we mere mortals have not ruled it out either together or separately. It is not ruled out for us (or at least the eliminativism-loving society). If it has to be ruled out in order for us to have second-order knowledge, as it certainly seems, then it looks as though we are in trouble. However, what is the scope of the eliminativist hypothesis? We know it includes the claim that no one has ever believed anything, but what about all the other mental states? Usually eliminativism is generally characterized as the view that there is something radically wrong with ‘folk psychology’, at least when compared with other theories. But there are several ways to make this unambiguous and precise. And how we do so is determined in large part by which proeliminativist arguments we focus on.13 13 Stich offers a very helpful summary of eliminativist arguments (eleven by my count!) in his 1996: 16–29. Since the details of these positions and arguments are both distracting and irrelevant to my purposes, I will not examine them here. However, it is worth noting that some eliminativist positions are put forward but go unendorsed. For instance, I don’t know anyone who actually thinks that all instances of ‘Linguistic string X is meaningless’ are true, despite Kripkenstein (i.e., neither Kripke, Wittgenstein, nor even Wittgenstein as interpreted by Kripke believed that; see Kripke 1982).

The First Live Sceptical Hypothesis

37

One might try to articulate an eliminativist theory that is consistent with the truth of ‘S knows that P’. One might do this by arguing that although ‘S knows that P’ does require the satisfaction of some doxastic-cognitive condition, it isn’t that of belief or propositional acceptance: even though ‘S believes that P’ and ‘S accepts the truth of P’ are false, ‘S knows that P’ is true. True, one could try to work this out; I have nothing against the existence of such theories. My point is this: eliminativism as I have characterized it, as entailing the falsehood of ‘S believes that P’ and ‘S accepts the truth of P’ as well as including the truth of ‘If ‘S believes that P’ and ‘S accepts the truth of P’ are false, then ‘S knows that P’ is false’, is live—or damn close to being live, and certainly live in very close possible worlds (worlds in which eliminativism is false but people take it more seriously than we do). So alternative eliminativist theories do not affect the argument; I don’t care about them. I will, however, present some relevant parts of a theory I call propositional eliminativism. My only reason for doing so lies in the extreme yet typical reactions to eliminativism. Many have viewed eliminativism with utter disbelief. ‘How can people take it seriously? Why should I waste my time giving it a moment’s thought?’ Brutally put, most of this disbelief is due to mere fashion: nowadays anticommonsensical theories are all the rage in metaphysics, but not in epistemology or in most quarters of the philosophy of mind. That is part of the reason people might think that it has been refuted, even though many lost souls have been unable to appreciate the ‘refutations’. I think that eliminativism can be made more plausible than it is usually viewed. Unquestionably, some of this dismissive attitude toward eliminativism is the fault of critics and incredulous bystanders not sufﬁciently understanding the theory as it has already been developed. For instance, one still encounters otherwise respectable philosophers who assert that eliminativists deny that we have cognitive processes or mental states; but of course no eliminativist holds any such thing. But the theory can be moderated as well. In the rest of this section, all I attempt is to indicate ways in which eliminativism need not do as much violence to common sense as is usually portrayed. This is meant to increase our conﬁdence in the truth of the Mere Mortal Premiss. But I do this with great hesitation, since I don’t want to rest the live sceptical argument on any one version of eliminativism. In order for the sceptical argument to work, we need just one version of eliminativism that is live. Thus, if

38

The First Live Sceptical Hypothesis

propositional eliminativism doesn’t do the trick, this doesn’t mean that we’ve eluded scepticism’s traps; Quine, the Churchlands, the former Stich, the 1970s Dennett, and others, real or merely possible, are waiting in the wings.14 The eliminativist might in the end admit the reality of a greater number and greater variety of mental states than Fodor—the alleged industrial-strength realist about the mind. It is just that, according to this propositional eliminativist, none of those mental states are of the kind that Fodor wants; none of them is a propositional belief. It would be similar to the case of phlogiston: there’s something there all right (for phlogiston it’s oxygen; for propositional beliefs it’s some non-propositional mental states), but the notion in question is such that its extension is empty. The notion of mentality in general is so vague and relatively theory-independent that the propositional eliminativist will allow it a non-empty extension. This holds for cognitive processes, feelings, emotions, and perhaps even thoughts; the propositional eliminativist will accept non-empty extensions for them all.15 Often in eliminativist discussions a philosopher says that the entities of folk psychology are the target, but there is often no speciﬁcation of the set of entities meant. One cannot just say that belief, desire, and the rest all stand or fall together. One has to indicate the precise scope of the folk ‘theory’, and one has to show that the pro-eliminativism arguments cover everything that falls within that scope; otherwise all we have is loose talk. I happen to be afraid of spiders, and I don’t know of any decent eliminativist considerations that put pressure on that folk-psychological claim. It seems to me that the pro-eliminativist arguments can be interpreted as requiring that all beliefs and assumptions, for instance, go while many folk psychological episodes and states can stay: Fear, want, anticipation, terror, remorse, hope, pride, joy, jealousy, annoyance, cheer, boredom, hesitation, love, attention, shock, hate, grief, regret, notice, disgust, sorrow, relief, exultation, disgust, wit, courage, discretion, expectance, pride, humility, vanity, modesty, 14 Paul Churchland 1989; Patricia Churchland 1986; Stephen Stich 1983; Richard Rorty 1970; Willard Quine 1960, 1985; Paul Feyerabend 1963; Daniel Dennett 1978. 15 The propositional eliminativist can also acknowledge the existence of instances of non-propositional kinds of knowledge such as objectual and ability knowledge. Indeed, she will probably argue that these kinds of knowledge do a lot of the work in our getting about in the world.

The First Live Sceptical Hypothesis

39

sympathy, fanaticism, cordiality, pathos, enthusiasm, aggression, deﬁance, scorn, earnestness, inspiration, ecstasy, callousness, nonchalance, fascination, patience, impatience, intolerance, equanimity, smugness, passion, arrogance, and shame. Armed with a thesaurus (as I was), we could continue the list for hours, producing a list of hundreds of folk-psychological terms that have non-empty extensions compatible with eliminativism. The eliminativist can admit that many statements from folk psychology (using the above vocabulary) are true, and genuinely explanatory to boot: she kicked her little brother because she was annoyed with him; he cried because he wanted her ice-cream. And these folk-psychological states have intentionality: her annoyance is with her little brother; his desire is for her ice-cream. This is not to say that everything you think about ‘aboutness’ will be saved by the propositional eliminativist; but it is to say that she can admit the existence of some of it, though impoverished. Furthermore, the eliminativist can allow the deliverances of introspection in a sense: you can ‘observe’ your cognitive states while having them. And you can know by introspection those cognitive states—it’s just that such knowledge is objectual (as in ‘I know Paris like the back of my hand’), not propositional, and objectual knowledge is not what you thought it was (because it need not require any corresponding propositional knowledge). Why draw the line at propositional beliefs? If we’re going to accept so much folk psychology (see list above), then why not go all the way? Basically, it’s a fear and loathing of ‘that’-clauses (see n. 13): strictly speaking, the theory goes, everyone fails to X that P, for any mental attitude X and proposition P. There are (in some very close possible worlds anyway) well-respected, evidentially based, thoroughly tested, and not knocked-down—live—arguments for the idea that if propositional beliefs exist, then they have to satisfy certain conditions (modularity conditions, for example). But science suggests that nothing comes even close to satisfying those conditions; and just as importantly it also suggests that other things—in the place of beliefs, as it were—constitute our cognitive processes and are the things that make us so successful in navigating the world. Not just anything can be a belief. Conditions such as modularity are required for propositional thought content, not impoverished and mere causally based intentionality. If the science turns out differently, then we might see propositional beliefs scientiﬁcally vindicated.

40

The First Live Sceptical Hypothesis

I know, I know. None of the philosophical-scientiﬁc evidence in favour of eliminativism is anywhere near conclusive. My sketch is not intended to purchase your assent to its truth! Like you, I don’t buy propositional eliminativism or the story I just told. And it just seems wrong when people say that we can have so many mental states successfully guiding our actions but that none of them qualify as beliefs. One is tempted to reply to the propositional eliminativist as follows: Yes, but surely in order to believe that P one does not need to satisfy any really tough conditions involving physical cognitive architecture. What the scientists discover about action guiding mental states just are discoveries about the things—beliefs—that make belief sentences true. And how can we construct sentences, which presumably can be intentional objects, but remain forbidden to accept them in a semantic way? You’ve practically admitted that we can as it were gaze approvingly upon the truth that Mary married Mark. Why isn’t that sufﬁcient for believing it? And how on earth could we even have meaningful sentences without beliefs—how do you propose to ground linguistic meaning without any propositional attitudes?! One also thinks of many other objections and challenges. But I won’t go into them because while they show that eliminativism needs development and its supporting arguments face difﬁculties, they do not defeat them.16 Isn’t that right? Think about how philosophical fashions and conventional wisdom change over the years. ‘It’s obvious now that theory T is hopeless’; then ﬁfty or a hundred years later, ‘Boy oh boy: were they wrong about T! Thank goodness we’ve got it right now.’ Browse the journal articles from a hundred years ago; witness the conﬁdence in the presentation of utterly inconclusive arguments. Again, this is not to say that virtually nothing gets ruled out; in Chapter 2 I gave over a dozen examples of claims that have been ruled out at one point or another, some of which are philosophical. Consider Kant’s ironically adolescent proclamation in the ﬁrst preface of the Critique of Pure Reason noted in Chapter 2; could he really have had much appreciation of the history of philosophy? The sorry fact is that we haven’t ruled out the possibility that we are navigating the world successfully with intentional but non-propositional cognitive 16

Okay, maybe my version is no good. But there are others.

The First Live Sceptical Hypothesis

41

processes—and propositional cognitive processes are precisely what are needed for the truth of instances of ‘S believes that P’. Perhaps you have intentional mental states that are successfully guiding your actions, but they just aren’t beliefs. Many distinguished philosophers and cognitive scientists have recently believed or suspected just that. You, a mere mortal, haven’t ruled it out. B Initial Defence of the First Sceptical Argument In this section I will address many objections to the live sceptic’s reasoning. I treat them in no particular order. I consider many other objections in Chapters 10–12. I choose to treat these here because most of them are not difﬁcult to rebut, some are based on misconceptions of the live sceptic’s reasoning and so should be cleared up right away, and some are objections that are immediately obvious and so deserve to be treated at least in part right away. The Mere Mortal Premiss says that in our intellectual community the eliminativist hypothesis is a live epistemic possibility in the sense given in Chapter 2. Obviously, this is true for part (ii) of the hypothesis, about knowledge requiring belief (or acceptance). The only question is whether it is true for part (i), the hypothesis that we have no beliefs and ‘S believes that P’ is always false for us. The sceptical argument says that part (i) of the eliminativist hypothesis is currently a live socio-epistemic possibility in our intellectual community, in the sense that it has been through a signiﬁcant scientiﬁc and philosophical evaluation, it is judged actually true or about as likely as any relevant possibility by a signiﬁcant number of well-informed, well-respected, and highly intelligent experts in the philosophy of mind and cognitive science, those experts reached that opinion based on its merits in a familiar, epistemically responsible way, those experts consider there to be several decent and independent arguments that support it, many of those experts consider it live, etc.17 In short, the hypothesis is a going concern among the experts. But that is just a brute fact about our profession. You might not like it, but that’s just the way it is. The Churchlands, Feyerabend, Stich, 17 Don’t try to reduce my sufﬁcient conditions for liveness to something else. Although I sometimes use the phrase ‘epistemic possibility’, what I’m really after is sociological possibility, where the latter is partly epistemic.

42

The First Live Sceptical Hypothesis

Rorty, Quine, Dennett, and others have at various points in their distinguished careers concluded upon careful reﬂection that it’s very probably true or as serious a contender as any hypothesis about folk psychology. They arrived at these views in the familiar and epistemically responsible way of looking hard at data and lines of reasoning, and then thinking about them in highly intelligent, relatively unbiased, and open-minded ways. And it goes without saying that no magic bullet has been found for it, and people still take it virtually as seriously as ever. Now who are you to tell me that these distinguished philosophers were or are a bunch of bozos, and that eliminativism isn’t live? None of this shows that eliminativism is true, of course; ‘live’ doesn’t mean true. What this also means is that the live status of a hypothesis is a sociological fact, not a matter for philosophical argument. One does not mount a philosophical argument for (or against) the live status of a hypothesis. Instead, roughly put, one just examines how the relevant people in the intellectual community actually treat the hypothesis. That is part of the reason why I refuse to provide a discussion here of eliminativism’s truth or the formulation of eliminativism that best ﬁts the best arguments of the Churchlands or the others. Even someone with almost no awareness of the particular content of eliminativism could ﬁgure out that it’s live by studying how professional philosophers and cognitive scientists treat it, compared to competing claims about the mind. Okay, I lied. Eliminativism isn’t live today, and has probably never been live. Too few adherents. When I argued above that the eliminativist hypothesis is actually live, I did it just for dramatic effect, to get you to feel as though I’m robbing you of your own current knowledge of belief and knowledge. For the sake of making things vivid, I argued that the eliminativist hypothesis is actually live. Surprisingly enough, its actual current status doesn’t matter too much for the new sceptical arguments. Suppose every version of eliminativism really is completely dead in the water, and there are just a few fools like me who haven’t noticed. So whenever one hears people take eliminativism seriously, one can safely ignore them. It is absolutely not live. The Churchlands, Stich, and the rest aren’t really serious and have been playing us for fools. Even so, there could quite easily be a society just like ours where eliminativism is false but a live view in philosophy and cognitive science. The same holds for the ‘knowledge requires belief/acceptance’ view. And there could easily be no geniuses

The First Live Sceptical Hypothesis

43

around with their own refutations of eliminativism. Just imagine a world in which the reaction to the works of Stich, Rorty, Dennett, and the Churchlands, for instance, had been a bit more positive in the 1970s, 1980s, and 1990s (and pretend that Stich didn’t offer a partial recantation in his 1996 book). That’s easy to imagine; the range of possible widespread human intellectual depravity is virtually without limits. Continue the fantasy—or nightmare—appropriately, and you will see that in an easily possible eliminativism-serious society the hypothesis is live. Thus, the live sceptical argument shows that a society virtually just like ours but with sufﬁcient philosophical confusion—imagine that! a philosophically confused society!—could have endured for a signiﬁcant period of time (decades, even centuries) during which virtually all philosophers and cognitive scientists knew nothing about their beliefs or knowledge. That’s shocking enough. Yet, in such a society you, dear reader, could feel the very same way about your own mind as you actually feel. That is all the Mere Mortal Premiss requires. In addition, for all we know, our society will soon enter into philosophical wickedness and whole-heartedly vote eliminativism as the theory of cognition most likely to be true. So although you have loads of knowledge and have been doing your best to educate your children and students, they will soon be full-ﬂedged members of an eliminativism-live society. The eliminativism-live society need not exist in another possible world; it may be in our future. You might be prepared to admit that if eliminativism really were live, then the live sceptic’s argument would be pretty compelling— but since eliminativism is live only in other possible worlds, you say, we in the actual world have nothing to worry about. But no. There is a good argument that suggests that if the eliminativist sceptical argument works as applied to a realistically possible society, then it works for (large parts of) our current, actual, philosophical community. For even if eliminativism never reaches live status in the actual world, contrary to what I argue, surely it is just a tiny accident that it never reaches that status. A few more eliminativists and eliminativism-sympathizers publishing in the right places at the right times over the past few decades, coupled with silence from or nonexistence of a few chosen realists, and eliminativism would have been live now. Given the extreme proximity of the eliminativismlive world, it really is just a tiny accident that eliminativism isn’t actually live. We in this world are on an epistemic par with the

44

The First Live Sceptical Hypothesis

eliminativism-live world. This makes one think that we actually know that Moore believed that scepticism is false only if we know that fact in the eliminativism-live world. Thus, if the liveness of eliminativism takes away our knowledge in that nearby world, then it takes it away in this world too. So the success of the live sceptic’s argument when applied to that world has the result that we lack the same knowledge in this world, whether or not eliminativism is ‘fully’ live in this world. This line of argument will hold for the other new sceptical hypotheses, to be introduced in the next chapter. But even if it fails, the idea that the argument works for the nearby world remains and is my primary focus in this essay. So I will proceed on the assumption that there is a large group of really intelligent and well-informed people whose opinion we respect, who judge eliminativism probably true or about as serious a contender as any, and who have good, well-respected, thoroughly tested, yet surviving arguments for their view.18 In this circumstance, who are we to say with such conﬁdence that we have ruled eliminativism out? That given our argument against it, it is no more than remotely possibly true? That it is based entirely on a confusion or misunderstanding? We mustn’t confuse ‘Theory T is ruled out’ with ‘Given our epistemic situation, we’d be better off spending our research time and effort assuming T is false’. The latter is reasonable in the case of eliminativism; the former is not. Remember: philosophical modesty. Surely philosophical arrogance cannot sufﬁce to purchase or retain knowledge. You just can’t be intellectually respectable and insist that, if you felt like it, you could write up your argument that rules out eliminativism—or worse yet, insist that your argument that has failed to convince the experts is perfectly ﬁne as it is. Get real. The same holds for hiking up the ﬂagpole your favourite famous argument that some genius gave decades or centuries ago; it isn’t a refutation because we’ve already examined it and found serious problems with it. It may have a lot of merit, but its merit is not that of being a public refutation.

18 How large does the group have to be? Surely precision is out of place here. There is a group of well-informed, etc. philosophers who take seriously the possibility that there are no chairs. But there aren’t enough of them to make it live, at least in my judgement. More to the point, that hypothesis does not meet all ﬁve of the liveness conditions.

The First Live Sceptical Hypothesis

45

At least, that’s the way it is for us mere mortals. I wouldn’t be surprised to learn that Kripke has among his voluminous unpublished writings the deﬁnitive refutation of eliminativism. None of what appears above entails that philosophical hypotheses are never or only rarely ruled out; none of it even entails that eliminativism hasn’t actually been ruled out (by a few chosen philosophers). But Kripke isn’t helping us. As remarked earlier, it is hard to see how the question of whether someone, somewhere, working entirely alone—or for that matter a small group of such individuals—has indeed ruled out eliminativism could have any bearing on whether the live sceptical argument works for the rest of us philosophers and cognitive scientists who are mere mortals and haven’t been able to rule it out, as a group or as individuals. So even if someone, somewhere, working in obscurity, has ruled it out with an exceedingly subtle proof that escapes many bright philosophers, that potential fact does nothing to help me, the dense one, rule it out. (Recall from Chapter 2 the 14-year-old girl who is a mathematical genius.) Even if a small portion of the best and brightest have refuted it, not just Kripke but Dretske and Burge, say, this still does not, it seems to me, mean that it is dead for me, Quine, Stich, the Churchlands, Dennett, Lycan, and just about everyone else. (Please don’t take the lists too seriously!) Even if eliminativism can be ruled out by introspection and a priori reasoning alone, which strikes me as possible but pretty unlikely for any serious version of eliminativism, this is nothing remotely like easy reasoning. It does not mean that anyone has actually done it or that any of us is bright enough to do it—even if we have tried. Perhaps you could, if you were a genius; but you aren’t, so you can’t. In addition, perhaps you could rule it out if you just worked at it for a few years and got lucky; but you haven’t, so you’re in no position to rule it out. Perhaps next year someone will publish a knock-down refutation of eliminativism—even the Churchlands and every other eliminativist will be completely convinced. Eliminativism will be as dead as any hypothesis has ever been. That would defuse the sceptical argument, for premiss (4) would then be false. This puts a legitimate limitation on the eliminativist sceptical argument (as applied to our community, not all possible communities). Roughly put, today it works; tomorrow it doesn’t. I’ll grant the point while insisting that no such refutation is around yet. Here is an objection to the above line of reasoning. Just because A is live and B is live does not mean that (A & B) is live. Suppose

46

The First Live Sceptical Hypothesis

theories T1 and T2 are live rivals. They are mutually inconsistent at many points. Clearly, their conjunction isn’t live. So just because both parts of the eliminativist hypothesis are live gives no reason to think that their conjunction is live. That is correct. Even so, the eliminativist conjunction is live. There is no inconsistency in the two parts; and plenty of philosophers hold both live. Furthermore, many of the reasons put forward on behalf of the ﬁrst conjunct about the non-existence of belief ﬁt well with the second conjunct about knowledge entailing belief. Eliminativists tend to think that it’s at least a contingent truth that believing P requires that one satisfy certain non-trivial causal-functional conditions. That is, a love of substantive necessary conditions on believing is often found among eliminativists. That attitude breeds acceptance of a substantive necessary condition on knowing: believing. In any case, if you’re willing to consider the contravention of common sense via the rejection of belief, then surely you’ll be willing to consider the rejection of propositional knowledge. I, for one, am happy with the idea that if propositional belief goes, then propositional knowledge goes as well; in the semantic mode: if ‘S believes that P’ is false, then ‘S knows that P’ is false as well. I’m trying to sell you on the signiﬁcance of the eliminativist sceptical argument (as well as the other live sceptical arguments to be presented below), partly on the basis that it is importantly different from the BIV argument. The eliminativist hypothesis is live; since the BIV one is absolutely dead, this makes the eliminativist argument stronger than the BIV argument. But it might be thought that there is really no interesting difference between the eliminativist and BIV hypotheses. They are equally outrageous. The philosophers and cognitive scientists who take eliminativism seriously don’t see that; but let’s call a spade a spade: they’re just horribly confused. It’s a crackpot theory, full stop. There’s no reason to let their confusion infect the rest of us. Surely we need not heed every outrageous philosophical fad that happens to ﬁnd willing victims among our contemporaries. I sympathize with some of that line of reasoning, but it doesn’t do the trick against the eliminativist sceptical argument. Perhaps the eliminativist hypothesis is just as outrageous as the BIV hypothesis— perhaps according to some metaphysical measure. I disagree, but I’ll grant it. Even so, only the former is live, and that’s the crucial epistemic difference. Eliminativism may be a fad, but it’s a fad

The First Live Sceptical Hypothesis

47

taken seriously by several of the best and brightest in the philosophy of mind and cognitive science. If you think you are best off ignoring it in your research, that’s ﬁne, but until you can rise to the level of an immortal with respect to it, it ruins your knowledge of belief because you are epistemically responsible to it. We have an epistemic responsibility to the live hypotheses in our intellectual community competing with our own beliefs, a responsibility reﬂected in the Modesty and Live Hypothesis Principles. No one is saying that you have to pay it any attention in your research or other reﬂections, but it does affect your knowledge just the same. Am I claiming that if some group of nuts takes over philosophy or cognitive science, just due to the fact that they are clever enough to publish often, but not wise enough to have any sense regarding anything, then whatever crackpot idea they latch on to automatically becomes live—something the rest of us level-headed folk are now epistemically responsible to, even though we can see how silly it is? No. Let it be stipulated that the experts in question—the ones taking eliminativism, for instance, so seriously—must have distinguished themselves in other areas of philosophy. And remember that the hypothesis has to have survived a thorough evaluation by such experts. Further, the experts have to think that the hypothesis is actually true, not just interesting. Contemporary epistemologists take the BIV hypotheses very seriously, but not because they think they have any chance at truth. The traditional sceptic doesn’t think that there are any evil demons, or BIVs, or mass hallucinations; she just thinks that those possibilities undermine our knowledge. Sceptics don’t think the sceptical hypotheses are anywhere near true; they just think that the hypotheses aren’t sufﬁciently neutralized for us to know things inconsistent with them. Throw in other conditions on liveness and mortality as you feel appropriate, adding to the nine I gave in the previous chapter. I trust that when you’re ﬁnished, our society still comes pretty close to satisfying the conditions and in very nearby possible worlds easily satisﬁes them. Am I claiming that we don’t have a successful refutation of a position if the advocate of the view we are trying to refute is intelligent and doesn’t accept it? No. If that were the case, virtually nothing would ever be refuted. As mentioned earlier, I’m allowing that some of us may have actually refuted eliminativism. All the eliminativist sceptic is saying is that many of us haven’t neutralized eliminativism enough to know things obviously inconsistent with it

48

The First Live Sceptical Hypothesis

if we’re mere mortals. Let me remind you what it means to be a (wellinformed) mere mortal: a. You’re about as aware as anyone of the fact that the hypothesis H is inconsistent with the proposition P alleged to be known. b. Your intelligence, understanding, and knowledge are not extraordinary for people in your intellectual community with regard to H or the issues surrounding H. c. You are at least somewhat familiar with H and the issues surrounding H, including the fact that H is live. d. If you have any reasons or evidence you can marshal for casting doubt on the hypothesis, and if they were carefully considered by the members of that community—in particular, by the wellinformed, well-respected, non-crackpot, and highly intelligent experts who were thoroughly familiar with the hypothesis—they would be nearly universally and conﬁdently rejected as clearly insufﬁcient to rule out the hypothesis. I’m claiming that all of these apply to most philosophers of mind and cognitive scientists, as well as to many other philosophers (if not today, in the future; and if not even in the future, at least in some possible worlds next door to ours). You might complain that many of us satisfy each of those conditions for the hypothesis that there are no artefacts, for instance, since there are some brave metaphysicians who endorse that wild hypothesis. But surely we don’t want to infer that we can’t know that there are electric cars just because we can’t defeat the artefacteliminativist arguments. That is correct, but (a)–(d) aren’t satisﬁed for artefact-eliminativism, mainly because that thesis is nowhere near live: i. In your intellectual community H has been through a signiﬁcant evaluation by experts over many years. ii. It is judged actually true or about as likely as any relevant possibility by a signiﬁcant number of well-informed, wellrespected, non-crackpot, and highly intelligent genuine experts in the ﬁeld(s) to which H belongs. iii. Those experts reached their favourable opinion based on H’s merits in a familiar, epistemically responsible way. iv. Those experts consider there to be several decent and independent arguments that support H. So it’s not the case that the only

The First Live Sceptical Hypothesis

49

reason people pay serious attention to H is the presence of one weird experiment or line of reasoning. v. Many of those experts consider H to be a ‘real, live possibility’ (i.e., that’s what they’d say if you asked them). In addition, I will later be adding a sixth condition: H has to be scientiﬁcally live in the sense of being endorsed as true or very likely true by scientists and for scientiﬁc reasons. But that discussion is reserved for section B of Chapter 11. Am I claiming that if we mere mortals in philosophy have the nerve to continue to believe that there are beliefs, then we are being irrational or arrogant? No, but I can see how someone might get excited and jump to this conclusion. Think of it this way. The Churchlands come to visit you, and you say to them, ‘There are loads of beliefs. I have lots of them. You’re wrong. So there. Losers.’ They then proceed to destroy every argument that you can give on your behalf or against their theory or evidence. They point out that many excellent philosophers and cognitive scientists take eliminativism very seriously, etc., etc. You sit there grinning with your arms folded across your chest: ‘Nope. I still don’t buy it. You’re wrong. Yeah, I can’t say one thing to put a big dent in your theory or evidence, but you’re wrong and there are lots of beliefs.’ There does seem to be something epistemically irresponsible going on here; but if we omit the grin, some insults, and the admission of one’s argumentative impotence, then it seems to me that the person is not guilty of anything very bad in continuing to believe in beliefs. We might even say that they aren’t being arrogant, irrational, unreasonable, or otherwise epistemically irresponsible. But they no longer have, on balance, enough (or sufﬁciently high-quality) warrant to know that there are beliefs—that is the minimal claim I’m making. We can elaborate on the point about irrationality and arrogance. It certainly seems overwhelmingly certain that we have beliefs. Lots of truly excellent philosophers have thought long, hard, and expertly about the epistemic status of beliefs in beliefs. But they don’t feel much pressure on the status of those beliefs from the threat of the live eliminativist hypothesis. And it’s not like they have their heads in the sand; they have confronted virtually every contrary argument levelled at them—surely they have carefully investigated all the good contrary arguments anyone has been able to think of for eliminativism. Well, if that’s the way things are, then one would think that our beliefs about

50

The First Live Sceptical Hypothesis

our beliefs can hardly be epistemically impoverished, given that those experts have concluded after thorough investigation that they are not impoverished. There is epistemic comfort in numbers—especially when the numbers include experts who have thoroughly and expertly investigated the matter for decades. After all, how bad can my beliefs about beliefs be when so many expert philosophers who have worked on the issues so well say that my second-order beliefs are just ﬁne? I understand how I need to neutralize the live contrary eliminativist hypothesis to some extent, but I can do so: I can point to the able research and opinions of expert philosophers who expertly and thoroughly reject eliminativism for my neutralization. It might be true that all of us common-sense philosophers have some epistemically vicious bit of reasoning underlying our second-order beliefs, but it can’t be so bad as to rob them of their status as warranted! This response might be sufﬁcient to rescue the ‘more than anaemically warranted’ status for a large portion of our beliefs about beliefs, but I don’t see much reason for thinking that it supplies enough neutralization for the beliefs to amount to knowledge. That is, if the best response to my argument lies in this comfortin-numbers objection, then although you are rescued from charges of irrationality and arrogance, for instance, you have done nothing that rescues you from the charge that your second-order beliefs don’t amount to knowledge. Am I claiming that you cannot know that your mother believes that your brother shouldn’t have bought that motorcycle? Neither your mother nor your brother know anything at all about eliminativism or the philosophy of mind, and when you came to believe that fact about your mother’s beliefs, you were certainly not thinking about eliminativism at any level. So how much of a threat, really, could the eliminativist hypothesis hold for that situation? Isn’t it really irrelevant in the sense that you don’t need to eliminate it in order to know the fact about your mother’s beliefs? Sure, eliminativism is an epistemic threat on occasion, in some select contexts, perhaps when we are working through theories of cognition, but that doesn’t mean that it is live in every context in which we form beliefs about beliefs. For instance, perhaps when we are discussing highly theoretical issues regarding belief the eliminativist hypothesis is relevant, but in ordinary contexts it isn’t. Here we see the appeal of a contextualist or relevant alternatives approach to knowledge, to be discussed in detail below (especially Chapter 10). For now, however,

The First Live Sceptical Hypothesis

51

I need to set out and clarify the live sceptical arguments before defending their soundness. One might think that the live sceptic is implicitly assuming externalism. After all, consider a world just like the eliminativism-live world except that eliminativism isn’t live. You exist in both worlds and possess the very same evidence for your belief that Moore believed that scepticism is false: in both worlds you have read some of what he wrote about scepticism, you have read what Barry Stroud wrote about Moore in his 1984 book, etc. You are intrinsically identical in the two worlds. The only difference between the two worlds is in how favourably philosophers of mind and cognitive scientists view eliminativism. Clearly, in the world in which eliminativism is not even remotely live—no Stich, no Churchlands, no Rorty, no Feyerabend, etc.—you are free to know that Moore believed scepticism to be false. The live sceptic has allowed that knowledge, since there is no live hypothesis ruining your belief. Well, if we’re internalists, then since you are intrinsically physically identical across worlds, your true belief is warranted enough for knowledge in both worlds. Thus, even in the eliminativism-live world you know that Moore believed scepticism to be false; so live scepticism is false. This is confused, and matters won’t improve if we ﬁddle with deﬁnitions of internalism. Take another look at the conditions on being a mere mortal. a. You’re about as aware as anyone of the fact that the hypothesis H is inconsistent with the proposition P alleged to be known. b. Your intelligence, understanding, and knowledge are not extraordinary for people in your intellectual community with regard to H or the issues surrounding H. c. You are at least somewhat familiar with H and the issues surrounding H, including the fact that H is live. d. If you have any reasons or evidence you can marshal for casting doubt on the hypothesis, and if they were carefully considered by the members of that community—in particular, by the wellinformed, well-respected, non-crackpot, and highly intelligent experts who were thoroughly familiar with the hypothesis—they would be nearly universally and conﬁdently rejected as clearly insufﬁcient to rule out the hypothesis. If eliminativism hasn’t even been thought of in your possible world, then it’s hard to see what those conditions amount to. For one thing,

52

The First Live Sceptical Hypothesis

(c) says that you are aware that eliminativism is live and you are familiar with H and the issues surrounding H, including some of the evidence for H. But how could that be the case in the world in which eliminativism isn’t even on the radar screen in philosophy or cognitive science? If eliminativism has never even crossed anyone’s mind, then (c) and (d) are odd, to say the least. Let’s back up. The internalist wants an objection that runs like this. A. In the eliminativism-dead world (the world in which eliminativism hasn’t crossed virtually anyone’s mind) the subject has internal access to all the same considerations (as to what Moore believed) as the mere mortal in the eliminativism-live world. She has read the same literature about Moore, etc. B. In the eliminativism-dead world she knows that Moore believed that scepticism is false. C. Internalism is true. And it entails that if the warrant she has internal access to in the ﬁrst world is sufﬁcient for knowledge, then, since she has access to all and only that warrant in the eliminativism-live world, she has the same knowledge in that other world. D. So, she knows she’s a believer in the eliminativism-live world. In order to make (A) true, the subject needs to be just like us, roughly put, vis-a`-vis eliminativism. Pluck a mere mortal from an eliminativism-live world and set her down in the eliminativism-dead world with exactly the same attitudes and philosophical abilities; that’s what it takes to make (A) true. Thus, the truth of (B) depends on the answer to the following question. Suppose you sincerely think that the liveness conditions are all true of your community of philosophers of mind and cognitive scientists; you know that your belief P (e.g., that Moore believed scepticism to be false) is inconsistent with eliminativism; your intelligence, understanding, and knowledge are not extraordinary for people in our eliminativism-live intellectual community with regard to eliminativism or the issues surrounding eliminativism; and if you have any reasons or evidence you can marshal for casting doubt on eliminativism, and if they were carefully considered by the members of our eliminativism-live community—in particular, by the well-informed, well-respected, non-crackpot, and highly intelligent experts who are thoroughly familiar with eliminativism—they would be nearly

The First Live Sceptical Hypothesis

53

universally and conﬁdently rejected as clearly insufﬁcient to rule out eliminativism. (That’s what it takes to make (A) true.) If that’s all true of you, then do you know P? I don’t see any reason to answer afﬁrmatively, even if internalism is true. I admit that someone utterly unacquainted with eliminativism could know many truths inconsistent with eliminativism; that issue was explored in detail above. But (B) in the internalist’s argument is just plain dubious. So the internalist objection fails. And frankly, even if it succeeded, we would still have the interesting conclusion that on externalist theories live scepticism is true. My sympathies are with externalism, and I constructed my argument under the inﬂuence of externalist thinking. So it’s fair to say that I was assuming externalism when I came up with the live sceptic’s argument. But I think that the above argument shows that live scepticism is a problem for internalists as well as externalists. I may be an externalist, but the live sceptic need not be. On to another objection. Some philosophers have thought that eliminativism is self-refuting in some sense: e.g., if one could accept eliminativism, then one could believe it; if eliminativism were a truth, then because there are truths, we could have intentional attitudes towards them—propositional attitudes. If there is such a refutation, and if the refutation is widely appreciated, then perhaps eliminativism isn’t live. The most satisfying way to deal fully with this objection is to demonstrate that every self-refutation argument fails (e.g., by noting that the eliminativist does not admit to accepting anything). For reasons of space I won’t do that here, although I suspect that it can be done for every self-refutation argument given thus far (many anti-eliminativists agree). However, all that really needs to be done to defuse the self-refutation objection to the eliminativist sceptical argument—although not to eliminativism itself!—is to point out that there is no well-known convincing argument that eliminativism is self-refuting or otherwise undermined. No such argument exists; as before, that is a sociological fact. So even if there is an unknown successful objection ﬂoating around waiting to be discovered—or even found by a few lucky souls—it does not threaten the Mere Mortal Premiss in the least. (I will be considering what I suspect is a new but still inadequate objection to eliminativism in section B of Chapter 11.) How could such an argument help the vast majority of us who don’t know about it and are unconvinced by

54

The First Live Sceptical Hypothesis

the arguments we have encountered? Maybe you know of one. Good for you; perhaps you are now an immortal with respect to eliminativism. I’m not. But wait: surely we mere mortals can rule out eliminativism! Consider the following analogous scenario. You have agonizing pain in one of your teeth. You go to several dentists and they can’t ﬁnd anything wrong with you, so you turn to doctors. Eventually you generate enough interest that the best doctors in the world spend all their time examining you. They still can’t understand what’s wrong with you. They say, ‘You can’t be feeling pain! You’re just faking it!’ That is the considered expert opinion. You know only so much about pain and nerve endings, and you certainly can’t show them where they have gone wrong in their investigations. Even so, you do know that you’re feeling pain. On the doctors’ advice you have given up your belief that it’s your tooth that’s in pain; and you might be convinced that one could feel pain even if there wasn’t any ‘objective’ pain, but you insist—and know—that you are feeling pain. The hypothesis that you aren’t feeling pain but are just faking it is live, and you need to rule out this hypothesis to know you’re feeling pain, but you have ruled it out—via some special experiences (the painful ones, naturally). Whether or not you are a mere mortal with respect to the live hypothesis that you aren’t feeling pain but are merely faking it depends on how we understand condition (d) of mere mortality. d. If you have any reasons or evidence you can marshal for casting doubt on the hypothesis, and if they were carefully considered by the members of that community—in particular, by the wellinformed, well-respected, non-crackpot, and highly intelligent experts who were thoroughly familiar with the hypothesis—they would be nearly universally and conﬁdently rejected as clearly insufﬁcient to rule out the hypothesis (although they may have other merits). You don’t have reasons that come anywhere near to defeating the justfaking-it hypothesis, at least under some natural construals of what reasons are. However, you do possess evidence that defeats the hypothesis: the painful experiences count as evidence (when the latter is of course being construed broadly). But you can’t use or marshal that evidence against the doctors, at least not in effective argument. They have tried to ﬁgure out what sensations you have, and they have

The First Live Sceptical Hypothesis

55

failed; the only evidence you can present them with is your sincere belief that you’re in pain. This scenario isn’t metaphysically impossible: it’s not that you have a sensation that can never be revealed through physical investigation; rather, the point is that you have a sensation that the current state of medical science can’t discover. Is the pain scenario similar to the eliminativism scenario? Pretend that all cognitive scientists and philosophers become eliminativists (but eliminativism is false). They’re all screaming at you with one voice: ‘You have no beliefs!’ This is just like the doctors screaming, ‘You aren’t feeling pain!’ Since in this latter scenario you could know that you’re feeling pain, shouldn’t we conclude that you can also know that you have beliefs—isn’t that a solid privileged access principle? You have some kind of privileged access to your painful feelings and beliefs; so you can rule out eliminativism just as you can rule out the just-faking-it hypothesis about your painful feelings. I don’t think so; the cases are importantly different. The doctors are missing out on something that you have: your experiences of pain. You are indeed epistemically privileged compared to the doctors. But in the belief case the philosophers and cognitive scientists aren’t claiming that you don’t have the particular belief that P although you have lots of other beliefs. They are saying that no one has any beliefs at all. Furthermore, you don’t have some special epistemic access to your beliefs that they don’t have to their beliefs. You both have beliefs, and you both have the same epistemic access to those beliefs. They think that the ‘introspective’ evidence for beliefs stinks; you think it’s great; but you have it in equal amounts. To get a sensation situation analogous to the belief situation we need to have the philosophers and cognitive scientists proclaim that there are no feelings at all, and not just that you didn’t have that tooth-related painful feeling. These eliminativists about feelings would argue just as the eliminativists about belief argue: No one has any beliefs/feelings. I realize that there are all sorts of cognitive/experiential processes occurring in our heads and bodies, but none of them have what it takes to be a belief/feeling. The notion of a belief/feeling is a muddled and primitive folk notion that has been constructed in such a way that anything that is a belief/feeling must satisfy certain conditions. But in all probability nothing satisﬁes those conditions, which is why there are no beliefs/feelings.

56

The First Live Sceptical Hypothesis

I assume that eliminativism about feelings really is crazy. Even if all the philosophers of mind and cognitive scientists were screaming at my door, ‘No one has ever had a painful feeling! There are no feelings! No one has ever been in pain!’, I would still know that I have had loads of feelings, painful ones included. If it really seems to me as though I have a perfectly deﬁnite, paradigmatic painful feeling, then I have a painful feeling. My ‘access’ to my painful feeling is so reliable and direct that I can rule out, without even thinking about it, any hypothesis that says I don’t have that feeling. The experts can scream all they want. Now it might be thought that we have given the anti-sceptic all she needs. Since I can know without any effort that the eliminativist about feeling is wrong, since I automatically have evidence that rules out her being right, then similarly I can know without effort that the eliminativist about belief is wrong, since I automatically have evidence that rules it out. I don’t think so. For one thing, I simply can’t imagine any remotely plausible reasons for thinking that no one is ever in pain. The nopain hypothesis just could not be live in a rational philosophical community. One can of course marshal good reasons for thinking that a certain kind of ‘phenomenal pain’ doesn’t exist, but that is another matter entirely. More to the point, belief is just more theoretically loaded than feeling. For belief isn’t some subjective state like feeling. Perhaps a painful feeling just is something that seems painful; but the analogous claim for belief doesn’t work. Sure, there is a feeling of subjective certainty towards a claim; to that extent belief can involve something akin to sensation. I can know, no matter what the experts say, that I have a feeling of certainty that is directed towards a claim. But there is more to belief than that. Beliefs aren’t subjective feelings, of certainty or anything else. They might include such feelings, but they are more than that. At the very least, it is pretty unreasonable to reject this view about belief going beyond feeling. In fact, I don’t know anyone who rejects this view. So even though I do have ‘subjective evidence’ that neutralizes the hypothesis that I don’t have any feelings of certainty or at least approval of a claim, I don’t have subjective evidence that rules out eliminativism about belief. For those reasons I don’t think that the anti-sceptic can reasonably hold that we all automatically, or upon easy reﬂection, possess warrant that is, on balance, sufﬁcient to neutralize eliminativism.

The First Live Sceptical Hypothesis

57

As we’ll see in Chapter 6, we possess positive warrant that when left alone does the trick, but the liveness and mere mortality conditions generate enough negative warrant to sabotage our knowledge. Earlier I mentioned the possibility that we have neutralized the epistemic threat posed by the BIV or evil demon hypotheses if we are merely ‘sane and breathing’. Well, one might think, why isn’t the same sufﬁcient for neutralizing eliminativism? Answer: because (a) the latter but not the former is live, and (b) you are a mere mortal with respect to the latter but not the former. Those are enormous epistemically relevant differences. I assume that the only reason to balk at the Mere Mortal Premiss lies in mere mortal condition (d). One might think that even if eliminativism is live in the sense of (i)–(v), one could still rule it out easily by merely reﬂecting on the fact that one has beliefs. But (d) is just saying that virtually all experts would judge the evidence you have against H to be clearly insufﬁcient to rule it out. It doesn’t say that the experts are right about that. Perhaps every one of us can rule out eliminativism, for instance, merely by saying to ourselves, ‘It sure seems to me that I believe that 2 þ 2 ¼ 4’ (although I just argued against that move). But in the worlds in which eliminativism is live, philosophers sufﬁciently sceptical about the truth of common sense will not consider such a line of reasoning as sufﬁcient to rule out eliminativism (so there aren’t too many John Searles there!). One might suspect that in the actual world a large majority of philosophers of mind really do think that one can refute eliminativism merely by breezily reﬂecting on one’s own beliefs. (Recall that we are assuming that eliminativism is false.) If so, then perhaps worlds that make (i)–(v) and (a)–(d) true are quite distant from actuality, contrary to the Mere Mortality Premiss. I think that the objector’s claim about the actual world is mistaken, but it will help the live sceptic’s cause if we get clearer on just how a world could make (i)–(v) and (a)–(d) true and yet remain very similar to actuality. A quick answer is that just as colour error theory is actually live today, as anyone familiar with the philosophy and science of colour can attest (more on that theory in Chapter 5), so belief error theory could have the same respected status. Both error theories are radically and comparably opposed to common sense; so that hurdle can be met. Perhaps more convincingly, one additional feature, mentioned above, easily makes (i)–(v) and (a)–(d) true at a world: more people

58

The First Live Sceptical Hypothesis

being sceptical about the strength of the connection between common sense and truth. As it stands today, in the philosophy of mind and epistemology communities there is a strong conﬁdence that statements that are conversationally appropriate in ordinary contexts are true. Epistemologists and philosophers of mind are typically loath to claim that large parts of common sense are radically mistaken. In metaphysics and the philosophy of language and logic, however, one does not ﬁnd this faith in the inference from common sense to truth. (Think of theories of vagueness, truth (the semantic paradoxes), material constitution, and identity through time—not to mention the philosophy of physics!) Indeed, it is hard to ﬁnd any contemporary metaphysician who doesn’t hold claims that are about as outrageous, from the perspective of common sense, as eliminativism. An acquaintance of mine who is sympathetic to dialethism (and with a very impressive CV) thought live scepticism was pretty obviously true. In my experience, those who work only in epistemology or the philosophy of mind often ﬁnd this attitude highly dubious or even hard to fathom, but there are good reasons why it is prominent in other areas of philosophy. Perhaps the paradoxes that engage the metaphysicians and philosophers of language, logic, and physics are much harder than those in epistemology or the philosophy of mind, which would causally account for the difference in attitude; but given that metaphysics, the philosophy of language, and the philosophy of logic have shown us that even some of the most entrenched parts of common sense simply have to be rejected, one wonders whether ﬁdelity to common sense in epistemology and the philosophy of mind is a good constraint on theorizing. I don’t want to pass judgement on faith in the truth of common sense. I personally suspect that it really is mere faith. But the live sceptic needs only a much weaker claim, one operative in her rebuttal to the objection that worlds in which (i)–(v) and (a)–(d) are true are remote from actuality: in a single, close possible world, the sceptical attitude we ﬁnd actually in contemporary metaphysics, philosophy of language, and philosophy of logic regarding even the most wellentrenched and universal common sense has permeated the philosophy of mind, while eliminativism has many more expert supporters and fewer detractors. I’m not claiming that such an attitude shift would be correct or fully justiﬁed; I’m just saying that it could easily happen. In those worlds, which I claim are highly realistic in being very similar to the current time of the actual world, eliminativism is

The First Live Sceptical Hypothesis

59

live, and many of us are mere mortals regarding it. And since we inhabit such possible worlds, the use of the second person in motivating live scepticism is appropriate. So we have justiﬁed the live sceptic’s claim that worlds satisfying (i)–(v) and (a)–(d) are next door to actuality. Now I want to further clarify my sceptical argument employing eliminativism by comparing it to another kind of argument, the kind concerning ‘evidence one does not possess’. Here is an example of Gilbert Harman’s: In case one, Mary comes to know that Norman is in Italy when she calls his ofﬁce and is told he is spending the summer in Rome. In case two [but not in case one], Norman seeks to give Mary the impression that he is in San Francisco by writing her a letter saying so, a letter he mails to San Francisco where a friend then mails it to Mary. This letter is in the pile of unopened mail on Mary’s desk before her when [exactly as in case one] she calls Norman’s ofﬁce and is told he is spending the summer in Rome. In this case (case two), Mary does not come to know that Norman is in Italy. (1980: 164)

Three points must be noted regarding case two. (a) Mary isn’t aware of the misleading claims made in the letter on her desk (she hasn’t opened it yet); she isn’t even aware that there are any claims in the letter that go against her belief that Norman is in Italy. (b) If she read the letter, she would probably give up her justiﬁed true belief that Norman is in Italy, as the letter would explain Norman’s ﬁctional but completely believable change of plans, etc. (c) If she read the letter, yet failed to give up her belief that Norman is in Italy, then she would probably be irrational in doing so provided we complete the story in the obvious way. The evidence in the letter is ‘evidence one does not possess’; it is not possessed because of point (a). It is controversial whether in case two, in which she does not yet know about the letters, Mary knows the fact that Norman is in Italy. She has a justiﬁed true belief formed in the ‘right’ way, but because of the easy accessibility of the misleading but powerful counter-evidence in the unopened letter, it is difﬁcult to say whether her justiﬁed true belief amounts to knowledge. After all, once she reads the letter, she will no longer know that Norman is in Italy; in fact, she’ll give up that belief once she reads the letter. Harman says that in case two she doesn’t know that Norman is in Italy, even though she has yet to see the letter—but of course he argues for this; he doesn’t leave it to untutored intuition.

60

The First Live Sceptical Hypothesis

In the scenarios that the live sceptic employs, the analogies of (a), (b), and (c) are false. The protagonist for the live sceptic—the wellinformed mere mortal with regard to the live sceptical hypothesis— does possess at least some of the available misleading evidence (from cognitive science and the philosophy of mind), yet continues to have beliefs that contradict the contrary hypothesis. Harman’s Mary is not a mere mortal, because she is unaware of the contrary evidence; the live sceptic’s argument concerns mere mortals only; so the analogy for (a) is false. The misleading evidence for eliminativism has convinced some people of eliminativism, but not most people, including our protagonist; so the analogy for (b) is false. Finally, the analogy for (c) is false as well, as there is nothing irrational about continuing to have beliefs in conﬂict with the live sceptical hypotheses, even when one is a mere mortal. For these reasons, then, the live sceptic is obviously not using an ‘evidence one does not possess’ type of argument. The conclusion (c) from Harman’s example can be watered down: if Mary hung on to her Norman-is-in-Italy belief even after reading his letter, then although she might not quite be guilty of irrationality, at least she no longer knows that Norman is in Italy, even if she knew it while the letter was unopened on her desk. That conclusion seems perfectly safe, especially if she then encounters some further, independent evidence that he is San Francisco—another letter perhaps. Now suppose further that after receiving the two letters from San Francisco she gets more evidence that he is in Rome: she sees what certainly looks like Norman in the fuzzy background of a picture in an appropriate newspaper. Now she has conﬂicting evidence and probably shouldn’t believe anything regarding his whereabouts. If she sticks with her Norman-is-in-Italy belief, then surely her true and partially justiﬁed belief doesn’t amount to knowledge, even if she isn’t guilty of irrationality. But that is what the live sceptic is saying about mere mortals with regard to eliminativism! If we happen to go along with antieliminativism, even when appraised of the misleading evidence and admitting both the strength of that evidence as expertly endorsed and the liveness of the thesis, then even though we manage to have true, partially justiﬁed, and rational beliefs, they aren’t knowledge. Here is an interesting objection centred on degrees of belief and epistemic vices.19 Professors Conﬁdent and Cautious are both mere 19

Thanks to J. C. Beall for convincing me of the need to address this topic.

The First Live Sceptical Hypothesis

61

mortals with respect to the live eliminativist theory. Yet both correctly think that eliminativism is mistaken. Both believe that Moore believed that scepticism is false. But as you may have guessed from their names, whereas Conﬁdent remains as conﬁdent as ever regarding the truth that Moore believed that scepticism is false, Cautious has lowered her degree of belief. Cautious just isn’t as conﬁdent as she was before she found out about eliminativism. The following principle seems pretty reasonable: if two people believe that P on the very same basis but one has a lower degree of belief, then even if the more conﬁdent one is unreasonable in her belief, the less conﬁdent one might be reasonable in her belief. After all, we say things like ‘The evidence justiﬁes P to only a relatively low degree’. Presumably, if you adjust your conﬁdence in P accordingly, you’re epistemically upstanding. Thus, perhaps Cautious is reasonable even though Conﬁdent is not. This is not sufﬁcient for an objection to live scepticism, though, as the latter doesn’t deny reasonableness or rationality to mere mortals. It denies only knowledge. But the objection can continue in either of two ways to conﬂict with live scepticism. First, just as ‘lower degree of belief ’ entails ‘less warrant needed for reasonableness’, ‘lower degree of belief ’ also entails ‘less warrant needed for knowledge’. So even though the live sceptic is right that Conﬁdent doesn’t know P, Cautious does know P. Second, the objector might again grant to the live sceptic that Conﬁdent doesn’t know that Moore believed that scepticism is false and then offer an explanation of that fact: Conﬁdent doesn’t know because she has been epistemically naughty in not lowering her degree of belief. You can’t know that P if your belief in P commits signiﬁcant epistemic sin. But Cautious is still free to know that Moore believed that scepticism is false because she is without epistemic sin, as she did lower her degree of belief. I have several comments on these two arguments. First, the live sceptic could easily live with them. That is, she could just accept them. Live scepticism would still be interesting because, as a matter of brute fact, zillions of philosophers and cognitive scientists don’t lower their degrees of belief in claims obviously inconsistent with live sceptical hypotheses. Furthermore, the qualiﬁed result, ‘You may know that Moore believed that scepticism is false but only if you signiﬁcantly lower your degree of belief ’, is still pretty interesting, at least by my lights.

62

The First Live Sceptical Hypothesis

Second, I don’t see why we should accept the premiss that just as ‘lower degree of belief ’ entails ‘less warrant needed for reasonableness’, so ‘lower degree of belief ’ also entails ‘less warrant needed for knowledge’. Maybe I’m being daft here, but I don’t see any reason at all to accept it. Third, I agree that Cautious is less naughty than Conﬁdent. But (a) she still looks a bit naughty in not going agnostic vis-a`-vis the Moore belief, and (b) even if she is without epistemic sin in continuing to believe that Moore believed that scepticism is false, I still think that she doesn’t know that Moore believed that scepticism is false. There are loads of examples that show that being without sin doesn’t make one knowledgeable, even with a true belief. So those arguments fail. Thus far I have focused on the epistemic position of the mere mortals who aren’t experts. But can the full-ﬂedged experts know which theory is true—the experts who are as aware as anyone of the reasons the contrary hypotheses are endorsed? I think so, depending on the details of the situation. It might be the case in the dinosaur story that if the advocates of the solar ﬂare and supervolcano hypotheses saw all the available evidence and understood it, they would come to know that the meteor hypothesis is true on the basis of their understanding and appreciation of that evidence. This might be the case if the only reason there are any solar ﬂare and supervolcano advocates has to do with people who are not informed of all the available evidence. Once the relevant evidence is published and digested, anyone who knows their stuff will be a meteor theorist— the only exceptions being boring ones (e.g., those who have a professional stake in one of the other hypotheses and can’t give them up for that reason). But this is not the scenario envisioned in this essay. Instead, there are many theorists informed of pretty much the same evidence, lines of reasoning, and data, who are equally sharp and informed, who have distinguished themselves elsewhere, etc. They just disagree. Many plump for theory A; many others plump for theory B, obviously inconsistent with A. The theorists on both sides are top-notch. As I wrote earlier, perhaps some of the best and brightest will know which theory is correct, but this will be the exception, not the rule. We might even admit that some of their best students know which theory is correct as well—even though those students aren’t experts! That’s one of the reasons we want our children to study with the best experts. I don’t know if these concessions to the anti-sceptic are correct, but even if they are, the vast

The First Live Sceptical Hypothesis

63

majority of people will not know which theory is correct. That is what the live sceptic is claiming. The goals of this chapter were ﬁve. First, I had to set out very carefully the ﬁrst sceptical hypothesis, distinguishing it from related theses with which it could easily be confused. Second, I argued that it is live—either in our current, actual epistemic community or in highly realistic possible or perhaps future communities. Third, I argued that many of us are mere mortals with regard to the hypothesis. Fourth, I addressed a fairly random series of objections. Fifth, I showed how the live sceptic’s argument is importantly different from arguments concerning ‘evidence one does not possess’. Now we take a look at some other applications of the live sceptic’s argument template.

5

Additional Sceptical Hypotheses It is impossible to calculate the moral mischief, if I may so express it, that mental lying has produced in society. When a man has so corrupted and prostituted the chastity of his mind as to subscribe his professional belief to things he does not believe, he has prepared himself for the commission of every other crime. Thomas Paine, The Age of Reason

Thus far I have presented the sceptical argument template with the eliminativist hypothesis plugged in, but other sceptical hypotheses work just as well or even better. In this chapter I show how to use the sceptical argument template to generate roughly fourteen new sceptical arguments. The familiar BIV hypothesis is a sceptical hypothesis, as it is employed in a ruling-out kind of argument for a kind of scepticism. The hypothesis that we don’t know anything about the external world (roughly put) because of the BIV hypothesis is what I will call a meta-sceptical hypothesis. We use the sceptical hypothesis to argue for the meta-sceptical hypothesis. Similarly, corresponding to the familiar evil demon sceptical hypothesis is the meta-sceptical hypothesis that we know nothing at all because of the evil demon hypothesis. In general, there is a whole family of meta-sceptical hypotheses generated from the family of sceptical hypotheses. I take those meta-sceptical hypotheses to have at least three components. First, they each have a central argument for the conclusion that set S of people do not, and perhaps cannot, know any of the true propositions in set P. (It is assumed that such a conclusion has philosophical interest; so it won’t do to announce that 3-year-olds don’t know any of the non-logical truths of quantum theory.) Second, each metasceptical hypothesis contains an interesting explanation of why the conclusion holds (this explanation might be partly given by the ﬁrst component, the argument). Third, each meta-sceptical hypothesis

Additional Sceptical Hypotheses

65

will have an interesting explanation of why we tend to think that the sceptical conclusion is false. Thus, each meta-sceptical hypothesis looks like this: Such-and-such sceptical theory is true: its argument is sound and its explanations are true. Thus, the conclusion, ‘None of the people in S knows any member of set P’, is true. The Ss don’t know these Ps, even though they have lots of true beliefs in them (depending on the sceptical argument). Many if not most of these true beliefs may (again depending on the sceptical argument) have dribbles of warrant. The only problem is that none of those beliefs is warranted enough for knowledge. Switching to a semantic construal: all occurrences of ‘S knows P’ (for these Ss and Ps) are false. Now we can take these meta-sceptical hypotheses and plug them into the sceptical argument template, just as we plugged in the eliminativist hypothesis. When we do so, we obtain the following argument. You believe that S knows P. But you’re a mere mortal fully aware of the live status of meta-sceptical hypothesis H, which says that S doesn’t know P. You can’t rule out that meta-sceptical hypothesis. So you don’t know that S knows P; you lack that bit of second-order knowledge. That is, you don’t know that S knows P no matter how hard you try, short of ruling out H. Take care here: the idea isn’t that if meta-sceptical hypothesis H is live, then you don’t know any of the Ps. Instead, the conclusion is that since H is live, the mere mortals don’t know that anyone knows any of the Ps. For instance, the BIV sceptical hypothesis generates the metasceptical hypothesis H that we don’t (or can’t) know any external world proposition (roughly put). Plugging H into the live sceptic’s argument template from chapter 3 generates this conclusion: Mere mortals with respect to the live H don’t know that anyone knows any external world proposition. It does not conclude with any of these claims: . . .

No one knows any external world proposition. Mere mortals with respect to the live BIV hypothesis don’t know any external world proposition. Mere mortals with respect to the live BIV hypothesis don’t know that anyone knows any external world proposition.

66 .

Additional Sceptical Hypotheses

Mere mortals with respect to the live H don’t know any external world proposition.

Thus, when I plug traditional external world scepticism in for H, the resulting scepticism isn’t external world scepticism but a scepticism of the form, ‘No mere mortal regarding H knows that anyone knows anything about the external world’. This is not a boring theory. I think that I know perfectly well that my daughter Julia knows lots of facts about the external world, and any theory that says otherwise is pretty radical, even if not as radical as external world scepticism, which says that my daughter doesn’t know those facts. Are any radical meta-sceptical hypotheses actually live in our current philosophical community? For instance, is external world scepticism based on the evil demon argument a live theory? Opinions will differ, but consider this. Most of the greatest geniuses in history have tried hard but failed utterly to defeat sceptical arguments. Centuries of exceedingly careful and ingenious investigation into these arguments have yet to put them to rest. If anything, we currently investigate them more closely than our predecessors did. The sceptical conclusions are counter-intuitive, but after so much brilliant but failed refutation it would be a bit foolish to deny that they have much of a chance at being right. Isn’t it high time to admit that there is a good chance that the reason we have failed to block them even with our greatest efforts over hundreds of years is that they are sound? If you spend twenty years thoroughly searching your house for your keys and fail to ﬁnd them, isn’t it time to take very seriously the possibility that the keys simply aren’t there to be found? I personally ﬁnd this line of reasoning good enough to convince me that scepticism is a real, live possibility despite the fact that not many philosophers today are sceptics. Traditional scepticisms do not satisfy the sufﬁcient conditions for liveness given in Chapter 3, but those were sufﬁcient conditions, not necessary ones. It is sometimes said that no one is a sceptic. I disagree; I know philosophers who endorse scepticism. They aren’t afraid of admitting to others that their considered view is that no one knows anything, or practically anything. If one probes them, one ﬁnds that they are not confused—or at any rate, they aren’t much more confused than antisceptics. They conﬁrm that what they hold is that when in ordinary life we say things like ‘I know she thinks you shouldn’t have bought that motorcycle’ or ‘Taz knows that John is hiding in the tree’, what

Additional Sceptical Hypotheses

67

we say is false. Like anti-sceptics, they are impressed with the usefulness of these knowledge attributions, and so they persist in them with the attitude that in everyday life falsehood isn’t much of a drag on usefulness. No one denies that there’s something about our knowledge claims that is entirely ﬁrm and right, in some sense. They think that these attributions are indeed quite useful, at least in the limited ways we demand, but they are false anyway (there will be much more on this issue in Chapter 6). Knowledge attributions have at least two interesting qualities: usefulness and truth-value, and when antisceptics get hot and bothered about the ‘outrageous’ nature of scepticism, they are either fearful of looking foolish (if that’s your problem, you’d better either be a genius or stop expressing your philosophical opinions) or they are over-impressed by the reliability of the inference from usefulness to truth. Other philosophers, such as myself, don’t quite endorse scepticism, but hold that it’s a real, live possibility. These sceptics and sceptic-respecters usually don’t publish in epistemology, they aren’t epistemologists, and they are probably not abreast of the state-of-the-art anti-sceptical moves, but that’s to be expected, given their views. You don’t have to be an epistemologist to be a sceptic; neither need one be in tune with the last thirty years of epistemology in order to be an informed and intelligent sceptic! There is of course the idea, familiar in the history of philosophy, that even the most avowed radical sceptic cannot maintain her scepticism in ordinary life. But this is a mistake. There is no reason why I cannot continue giving knowledge claims without being resolutely committed to their falsity the whole time. Just because I continue employing claims of the form ‘S knows P’ does not mean that I think they are true—even when I say ‘They are true’. I can endorse them for their qualities other than truth, and I can call them ‘true’, intending just that they are true enough or true for all practical purposes or acceptable. When the sceptic utters ‘S knows P’ in what looks like an assertive manner, all she is committing herself to is the appropriateness of saying or using or assuming the truth of ‘S knows P’. I can do similar things in church even though I’m a goingstraight-to-hell atheist. The idea that the sceptic has to remain in her study because she claims not to know whether there is a monster outside the door is nonsense. Neither need the sceptic be at all in doubt of the truth of various propositions. She just thinks that knowledge is a peculiarly demanding concept that isn’t satisﬁed even when there is no room for any doubt whatsoever.

68

Additional Sceptical Hypotheses

One can quarrel with this interpretation of people who seem to profess scepticism or take it very seriously. Instead of taking them at their word, one can insist that they don’t really believe that. I ﬁnd this move dubious, especially after talking with these people to clarify what position they hold. It’s also a bit insulting, given my own views. Historians will probably be able to point to philosophically sophisticated intellectual communities in which scepticism was pretty clearly live. G. E. Moore certainly thought that many philosophers of his time were genuine sceptics. And for some reason I can’t fathom, no one has noticed that eliminativists who haven’t articulated a highly revisionary theory of knowledge that says that the belief condition on propositional knowledge is an illusion—and that’s all the eliminativists—are obviously committed to radical scepticism: you can’t know that P if you don’t believe that P. So it looks pretty clear that some philosophers today are committed to radical scepticism.20 I personally consider many forms of scepticism to be real, live possibilities; they really, really, could be true, but I don’t endorse any. Philosophers defend anti-commonsensical views, especially error theories, all the time;21 what is to prevent us from doing so in epistemology?22 In any case, I insist that it is trivial to imagine our society becoming in the near future a meta-scepticism-live society, if it isn’t already, so that the nine conditions alluded to in the Mere Mortal Premiss are or will actually be satisﬁed for many, many philosophers and appropriately educated non-philosophers. The other claims in the argument are independent of the nature of the sceptical hypothesis being considered, so they are as compelling as ever. Thus, we have a class of rather peculiar new arguments for various sceptical conclusions. Because these arguments use the liveness of 20

One way out of this argument regarding eliminativists: scepticism isn’t the claim that we don’t know; it’s the claim that our beliefs don’t amount to knowledge (or that our true beliefs don’t amount to knowledge). Eliminativists don’t buy that kind of scepticism, as they think that we have no beliefs. Against this idea is the thought that if there are huge communities of people with zillions of cognitive states milling about their environment successfully but not having any propositional knowledge, which is what eliminativists believe about the actual world, then scepticism is true for them. So it seems to me that eliminativists are sceptics. 21 We have already seen one of these error theories, eliminativism. About a couple of dozen comparable theses are mentioned near the end of sect. B of Ch. 11. 22 I will return to this matter in Ch. 6 when I explore the live sceptic’s position further, especially regarding how she accounts for the way in which ‘Joe knows that P’ is superior to ‘Joe doesn’t know that P’, even though the ﬁrst is false and the second true.

Additional Sceptical Hypotheses

69

kinds of scepticism to generate the truth of kinds of scepticism, we can call them bootstrap sceptical arguments, as the sceptic is pulling herself up by her own bootstraps. In some close possible worlds, the fact that it’s a live possibility that our beliefs are not sufﬁciently warranted makes many of them not sufﬁciently warranted. Every real philosopher hates the inference ‘People think it’s true; so it is true’; but perhaps we have stumbled across a bizarre case where the inference works.23 Let’s set aside meta-sceptical hypotheses and consider some sceptical hypotheses that we can plug into the live sceptic’s argument template from Chapter 3. The live sceptical argument gets off the ground whenever we ﬁnd an appropriately sceptical hypothesis that satisﬁes the nine conditions for liveness and mortality. As one might expect, those conditions are pretty demanding. But here’s a hypothesis that satisﬁes them in spades in some very close possible worlds: No ordinary physical objects (such as tomatoes, socks, etc.) are coloured; so all occurrences of ‘X is red’, for instance, for any ordinary physical object X, are false. Call this view colour error theory. Clearly, many people actually believe this thesis, especially those with a bit of scientiﬁc awareness. In fact, it just might be the dominant view among colour scientists these days. And many of the best philosophers ever have thought that it was a real, live possibility— at least, given undisputed historical facts, it’s child’s play to imagine this happening many times over in a realistic possible world next door to ours. It is trivial to imagine a robust intellectual community in which this hypothesis is live in the best circles. Not only that, the causes of the liveness of the error theory could be almost entirely scientiﬁc. This kind of error theory could have begun with scientists reﬂecting on experimental results, not philosophers armed with a priori considerations. Four clariﬁcations are needed in order to prevent misunderstandings. First, one can easily imagine a live theory according to which, in everyday contexts, ‘My socks are brown, but I’m going to wear black trousers anyway’ is true even though when the scientist or philosopher utters ‘No ordinary physical objects such as socks are coloured’, 23 Again, the bootstrap arguments do not quite go from the liveness of H to the truth of H. By the way, there are boring cases where the inference ‘People think it’s true, so it’s true’ works: e.g., if people think that there are people thinking thoughts, then it’s true that there are people thinking thoughts.

70

Additional Sceptical Hypotheses

she utters a truth. Such a theory may be defensible, although one must rely on some semantic tricks to pull it off; but that’s not the error hypothesis I’m considering. Colour error theory, as I deﬁne it, suggests that the right response to the disbelieving anti-sceptic is something along the lines of, ‘Well, even though strictly speaking ﬁre-engines aren’t red and those beliefs of ours are all false, it doesn’t matter too much because ﬁre-engines look red in virtually all contexts, and that’s enough to ground the practicality of colour realism’. Many philosophers adopt as a constraint the truth of common sense when construing truth conditions. They bend over backwards to make sure that according to their theories common folk aren’t speaking falsehoods a hundred times a day when uttering bits of common sense. Thus, no matter what the facts about colour and colour perception, they implicitly take it as their task to construe the truth conditions of ‘My sock is blue’ so that it comes out true when uttered in everyday contexts. These philosophers will not judge colour error theory to be live; instead, we must appeal to semantic complexities to characterize the live and semantically sophisticated theory hiding among the scientist’s pronouncements that ‘Nothing is really coloured’. This is a contentious approach in my judgement (I know I have unfortunately followed it from time to time). The proper constraint is practicality, not truth: ‘My sock is blue’ cannot fail to be completely practical, something the truth of which we can rely on in virtually any ordinary context—even if it isn’t true (‘S relies on X’ doesn’t entail ‘X exists’). Truth is another matter. Complete practicality is of course pretty good evidence of truth, but many examples from the history of science show that practicality evidence is easily defeated. Given that colour error theory is rooted in scientiﬁc considerations, we need not bend over backwards to save the truth of ordinary colour claims. (There is more on the nature of common sense in section B of Chapter 11.) Second, please keep in mind that I’m not making the trivial point that if ‘X is red’ is false, then ‘S knows X is red’ is false. I’m making the interesting point that if ‘X is red’ is true, but ‘X isn’t red’ is live, then ‘S knows X is red’ can be false even though ‘S has a true and perfectly ordinary belief that X is red’ is true. Third, Larry Hardin, who knows about as much of the science and philosophy of colour as anyone who has ever lived (e.g., Hardin 1988), comes very close to being a colour error theorist in my sense. (See Byrne and Hilbert 2003 for references to other colour error theorists.)

Additional Sceptical Hypotheses

71

Fourth, in a possible world ‘next door’ in which colour theory is live, it is still false. The world is exactly like ours except that people take colour error theory more seriously than we do; that’s the only difference. (The same holds for neighbouring worlds in which the eliminativist hypothesis is live: eliminativism is false over there, just as it’s false here.) So one can’t complain that the world in which colour error theory is live is so bizarre that it has no implications for us. Each of these four points holds, mutatis mutandis, for ‘No ordinary physical objects or events are ﬂat, solid, circular, triangular, straight, empty, parallel, unchanging, inert, permanent, instantaneous, perpetual, etc.’, ‘There are no character traits’, and ‘Pain occurs just about as frequently as people say it does, but it only occurs in the brain [or brain stem; I don’t know], so no one has pains in the rest of their body’ as well.24 Each of those fourteen claims serves to generate another live sceptical argument via the template in Chapter 3. I’m not suggesting that the sceptical arguments that these claims generate are equally good; I’m merely pointing out possible ways to ﬁll in the live sceptical argument template. Clearly, some of those hypotheses are live or close to live in our current analytic philosophy community; others are not. Interestingly enough, there are some live hypotheses whose negations could easily have had virtually the same sociological status among the intellectuals who investigate them (but not the general populace!): lying is/isn’t wrong; there is/isn’t a God; Michelangelo’s David is/isn’t beautiful; we have/lack free will. In these cases, the corresponding pair of sceptical arguments (generated from the Chapter 3 template) amounts to something like a positive case for agnosticism with regard to the corresponding issue. There is plenty of room for intelligent disagreement on this point regarding the scope of the live sceptical argument. We can ask, that is, what hypotheses can be plugged in to generate a decent sceptical argument for an easily possible intellectual community much like ours? Ethical, aesthetic, religious, and otherwise non-scientiﬁc hypotheses are so divorced from fruitful scientiﬁc investigation that one can easily suspect that they don’t work well in the live sceptical argument template. For instance, the quality of being philosophically live 24 For character trait error theory see Harman 1999, 2000; Doris 1998, 2002. For error theory regarding pain location see Hyman 2003.

72

Additional Sceptical Hypotheses

might be thought so weak, compared to that of being scientiﬁcally live, that having your belief that P contradicted by a philosophically live but not scientiﬁcally live hypothesis H is no real threat to your belief that P. In section B of Chapter 11 I will claim that the live sceptic is best off severely restricting the use of the live sceptical argument template. Although I have presented the sceptical problems in the form of pro-sceptical arguments, it is best to keep in mind that what we have are puzzles in the traditional philosophical sense: an inconsistent set of individually plausible claims. In most of the remainder of the essay I formulate and evaluate proposed solutions to the puzzles. In my judgement, defending the eliminativist sceptical argument is the hardest and most interesting of the lot, as it is most tempting to think that if you believe that P, you believe that you believe that P, and you formed both of these beliefs in the entirely ordinary way, then your belief about your belief amounts to knowledge. The colour and pain-location arguments are on somewhat more solid ground. What I personally ﬁnd fascinating is that even if some anti-sceptical solution succeeds, contrary to what I’ll be arguing, this does nothing to take away from the interest of the live sceptical problems, for every reasonable solution must embrace some interesting epistemological position that we have never had much good reason to endorse heretofore.

6

The Sceptical Solution One should respect public opinion in so far as is necessary to avoid starvation and to keep out of prison, but anything that goes beyond this is voluntary submission to an unnecessary tyranny. Bertrand Russell, The Conquest of Happiness

Odds are that regardless of what I have written thus far, you aren’t convinced of the soundness of any of the live sceptical arguments, even restricted to philosophers of mind, cognitive or colour scientists, or perhaps even merely possible or future yet realistic societies. For you still operate under the impression that any argument for any version of radical scepticism is outrageous. In all probability, you didn’t really give my pro-sceptical arguments an entirely fair hearing. You started out your evaluation of my arguments thinking that since the conclusion is totally unacceptable, there is certainly a mistake in the premisses or the inferences. This is just philosophy, after all. Before I even began defending myself, you ‘just knew’ I was wrong. Besides, who the hell is Bryan Frances anyway? Why should we listen to him? I hope I can say this without placing too much blame because, given what I know about my own philosophical prejudices, I suspect that I would have reacted that way, if not for what comes next. Although you have no doubt been thinking up anti-sceptical solutions to the live sceptical puzzles, I ask you ﬁrst to consider the consequences of the soundness of the sceptical arguments. Later in this essay I will use about 30,000 words presenting and evaluating a dozen or so anti-sceptical solutions. It is a brute historical fact that we are confronted with three seemingly outrageous propositions when considering traditional sceptical arguments: (a) the sceptical hypothesis itself, (b) the conclusion that we know next to nothing (or nothing of a certain form), and (c) the claim that some outrageous sceptical hypothesis dashes

74

The Sceptical Solution

our hopes for ordinary knowledge. In this chapter I present the bulk of the remaining positive case for the live sceptical arguments by examining these three alleged ‘outrages’, (a)–(c), applied to the live sceptical arguments. I will argue that the live sceptical arguments signiﬁcantly diminish all three. In particular, (b) and (c) are wiped out almost completely. So the nearly instinctive reaction ‘This is an argument for a kind of scepticism; so there’s a mistake in the premisses or the reasoning’ is deﬁnitely not intellectually responsible (if it ever was). This is not to deny that the sceptical conclusions and arguments generated from eliminativism and colour error theory (the two sceptical hypotheses I focus on most in the remainder of the essay) are less than wholly innocent; they have their own counterintuitive elements. Even so, the considerations to follow make me think that the option of simply accepting the soundness of the live sceptical arguments is at least as good as any of the anti-sceptical options. A Outrage: Sceptical Hypotheses Traditional sceptical hypotheses are completely outrageous. The hypothesis that you’re a BIV is just about the most outrageous thing any philosopher has come up with. This is part of the problem in convincing some people to take the sceptical arguments seriously. But now we are seeing that sceptical hypotheses need not be nearly as outrageous as they have traditionally been. Suppose for a moment that the eliminativist and colour sceptical hypotheses are true. . . . . . . . . .

Is there a physical world? Yes. Is there a planet Earth? Yes. Do you have two hands? Yes. Were the New York Yankees the best team in baseball in the 1950s? Yes. Are you a cognitive being functioning perfectly well in navigating the world? Yes. Are there any devious mad scientists or evil demons manipulating your mind? No. Are you in a coma or an extraordinary dream state? No. Was the universe created just ﬁve minutes ago? No. Are there any brains in vats enjoying sensory lives akin to yours? No.

The Sceptical Solution

75

Practically everything is the same in the true error theory world—a world in which all the live sceptical hypotheses are true—as it is in our world with the error theories all false. Ask a group of nonphilosophers to make utterly random lists of claims about the world they think are deﬁnitely, expertly, and widely acknowledged to be true. Even if the live sceptical hypotheses are true, virtually everything on the lists can be perfectly correct—that is, true even though we have no beliefs, tomatoes aren’t coloured, etc. These features of the live sceptical hypotheses demonstrate how utterly different they are from classic sceptical hypotheses. ‘Yes,’ you may counter, ‘but although eliminativism isn’t an outrageous hypothesis about the non-mental, physical world, so in that way it’s different from traditional sceptical hypotheses, it is an outrageous hypothesis about the mind. So don’t try to fool us into thinking that they aren’t wild.’ That’s right; I was overstating the case. But are they really that wild? Only someone literally insane would think that any of the traditional sceptical hypotheses is actually true today; but clearly sane, intelligent, sober, and informed individuals support the live sceptical hypotheses. Again: even if they are true, you are still a cognitive being functioning perfectly well in navigating the world with mental states. Don’t ever be suckered into accepting the wholly inaccurate idea that if eliminativism is true, then you have no mind or cognition or perception or emotions. A:

Fred is a real live person but not a child, a bachelor, a married man, a divorced man, or a widower. B: What? That’s impossible! A: Fred is a woman with an odd name. B: Oh. That explains it. A:

Jan says that her basket has thirty-two apples in it, an even number. But she’s not right. The number of apples in her basket is neither odd nor even. B: What? That’s impossible! A: She has no basket. She’s hallucinating. (And I meant to have the negation have wide scope!) B: Oh. That explains it. A: You have no beliefs. Perhaps better yet: no tomatoes are red. B: What? That’s impossible!

76

The Sceptical Solution

A: For all the philosophers and cognitive scientists have been able to determine, you have a whole slew of mental states that roughly correspond to things in the world and successfully guide your actions, but they aren’t technically beliefs. Not just any cognitive state has what it takes to be a belief, and it turns out that although you have a very rich mental life, just as you have always thought and experienced, the category of belief just doesn’t apply. Instead, a different set of categories applies to your mental states. The category of belief is a persistent and oddly useful illusion from our primitive conceptual scheme. The real cognitive categories will come from science. Further, although you perceive things as being coloured, they aren’t really coloured; the colour is just in your perception for suchand-such reasons. B: Oh. That explains it (a bit anyway). B Outrage: Sceptical Conclusions Suppose that even though eliminativism is false and we have as many true beliefs as you like, the eliminativist sceptical conclusion is true. In this situation most philosophers and cognitive scientists (the obvious candidates for mere mortals) know nothing about our beliefs. Or suppose that even though ﬁre-engines are red, we fail to know that ﬁre-engines are red. Conclusions like the ﬁrst one have seemed so outrageous that even the most radical sceptics usually won’t ask our indulgence in considering them for even a moment: e.g., how could you not know that you believe some things? But just why is it that sceptical conclusions, restricted or not, are judged to be so bad, so utterly shocking, even horrifying? Many things might be said concerning the outrage over the conclusions, but we need to keep our perspective. The great majority of contemporary philosophers judge radical forms of scepticism, restricted or not, to be just that: radical. It cannot have gone unnoticed that many undergraduates don’t share that judgement. For many of them, the conclusion is not so radical, not so bizarre, and not so weighty. Some even take it as pedestrian. Do they just not understand the thesis and its consequences? Certainly some are missing out, especially the ones who take scepticism to be a relatively pedestrian claim. But whereas the pedestrian-judging

The Sceptical Solution

77

students go too far one way, philosophers go too far in the other direction. One reason is simple: it’s almost an entrance requirement into contemporary analytic philosophy that one feel charged by the sceptical conclusion. Another, more interesting, reason is that nearly every philosopher spends a great deal of her working time and effort examining beliefs: beliefs about morals, works of art, physical objects, knowledge, truth, theories, abstract objects, induction, causation, time, necessity, and mental items themselves. It doesn’t matter what your areas of specialization are; you spend a lot of time thinking about people’s beliefs about the subject-matter of those areas. And we spend our professional lives trying to justify our philosophical views and beliefs and ask for the justiﬁcatory status of other beliefs (philosophical and non-philosophical). In other words, our philosophical lives revolve around beliefs and warrant. So, naturally, the sceptic, who forces us to consider the possible truth that knowledge is almost never found, is going to cause us signiﬁcant psychological discomfort. Frankly, what is a philosopher to do if there is no knowledge, or knowledge about the mind? What is left other than science and stamp collecting? This might be why we lose patience with the error theorist for knowledge and not for the error theorists for modality, mathematics, and colour.25 This also partly accounts for some of the more impassioned criticisms of eliminativism. If your life were devoted to ghost stories, you would be uncomfortable with people who thought that there were no ghosts whatsoever. For a professional philosopher, scepticism generates the type of psychological discomfort that you feel when a discussion participant at a conference claims that your talk is irrelevant because even if everything you said were true, it is philosophically and even scientifically without any interest. Philistine. That is why many of your students don’t take you too seriously when you blather on about the outrageous nature of scepticism.26 It’s easy to be melodramatic about sceptical conclusions. ‘No one knows anything ! We don’t even know our names or whether I’m over four 25 I confess to some blindness here. Error theorists are ubiquitous in contemporary philosophy, and their views are taken seriously. But philosophers tend to go to great lengths to exaggerate the counter-intuitiveness of the traditional sceptical conclusions. On my accounting, the arguments for error theory in epistemology are superior to those found in most other areas of philosophy. I confess to once being irrationally anti-sceptical myself. 26 The ‘good’ students, the ones who have a chance of becoming philosophers, are of course completely taken in.

78

The Sceptical Solution

feet tall or who is President of the USA or how many arms we have!’ But that isn’t much different from brainwashing. I don’t suppose that there is much we philosophers can do about this fact about ourselves, but it should sufﬁce to remind us that we might be signiﬁcantly exaggerating the counter-intuitiveness of the sceptical conclusion. Live sceptical views are not only easier to swallow than the traditional forms of scepticism but aren’t that hard to swallow, full stop. I will now present four arguments for this view regarding the degree of counter-intuitiveness of the live sceptic’s conclusions. First, what initially makes the live sceptical conclusions seem implausible, even if not shocking, is that although we have zillions of true beliefs about beliefs and colours formed through evolutionarily governed processes, for some reason these beliefs are never, or almost never, warranted enough to count for knowledge. How odd! How could we naturally form all these beliefs all the time, use them to navigate the world with enormous success, and yet a large portion (perhaps all) of them not be warranted enough to count as knowledge—even though part of our incredibly successful practice includes relying on distinctions between belief, true belief, and knowledge? How could we end up with so many happy but unwarranted beliefs? This does seem like a miracle. Thus, the conclusion does seem outrageous. However, if we keep eliminativism and colour error theory in mind, then although oddity remains, there is no miracle. Like other sceptical arguments, if the eliminativist sceptical argument is sound—but eliminativism is false!—then although you have millions of true beliefs, they aren’t warranted enough for knowledge. Why not? Not because you haven’t ruled out some insane scenario in which everything is completely different from the way you actually think it is. Not because knowledge requires certainty or infallibility or anything nearly that strong. No, the reason is that you haven’t ruled out a live, partially warranted, scientiﬁc hypothesis about the nature of our cognitive processes. The reason our beliefs aren’t sufﬁciently warranted is that we are aware of, but haven’t ruled out, the possibility that we are making our way through the world successfully with just non-propositional but intentional cognitive processes. There are cognitive and intentional things in your mind that successfully guide your actions; they just aren’t beliefs. That’s a possibility we need to rule out in order to know anything about our beliefs, because if it’s true, then we can’t have such knowledge (because you can’t know what’s false). Further, your

The Sceptical Solution

79

intuitions about the truth-values of epistemic claims are pretty close to right: your intuitions have been reliably tracking the appropriateness of epistemic claims. Your only mistake was to think that this appropriateness is, or required, truth (more on that below). A:

You have no knowledge about anyone’s beliefs—not even your own. B: What? That’s impossible! A: For all the philosophers and cognitive scientists have been able to determine, you have a whole slew of mental states that pick out things in the world and successfully guide your actions, but they aren’t technically beliefs. And of course without beliefs there’s no knowledge of beliefs. In order for us to know anything about our beliefs, we need to rule out this possibility that our mental states don’t count as beliefs. But you haven’t done so, and neither have we. You’ve got loads of know-how and acquaintance knowledge, which are absolutely crucial to getting around in the world, but that’s it. B: Oh. That explains it (a bit anyway). Now it doesn’t seem quite so reasonable to get hot and bothered about the eliminativist sceptical conclusion. Neither does the question ‘But how on earth could our successful beliefs fail to be warranted?’ have the bite it used to. As a matter of historical fact, we have quite naturally fallen into having one view about the mind. A prescientiﬁc view if you will. Now a competing view with scientiﬁc credentials—but no proof of course—has come on the scene. It promises a comprehensive theory of mentality—a theory that in many ways is much richer than the naı¨ve theory. You can hardly expect to just know that the naı¨ve view is true without having an impressive refutation of the scientiﬁc theory. All of this holds for the colour hypothesis as well. In other words, the live sceptic is claiming that we lose our knowledge by means of a perfectly familiar mechanism.27 Pretend that it had been common sense for hundreds of years that whales are ﬁsh; pretend it’s true too. Then scientists came along to challenge that bit of common sense. They admit that there are many ﬁsh and 27 Here I’m assuming that before we become mere mortals, say as children, the liveness of the sceptical hypotheses did not sabotage our knowledge. More on this point later in this section.

80

The Sceptical Solution

many whales, but they have some impressive arguments for the shocking claim that whales aren’t ﬁsh—they are mammals instead. If you take a quiz listing kinds of ﬁsh and you write ‘whale’, scientists will mark your answer as incorrect. If you say to your child, ‘Look at the whales, Julia; they’re the biggest ﬁsh there are,’ what you say is ﬁne in some respects but really false. For the purposes of this essay pretend that the scientists are wrong: whales are ﬁsh. So one familiar and even highly reliable method for ﬁnding out that something is a ﬁsh—ﬁnd out it’s a whale—would clearly be question-begging against the group of scientists. In such a scenario, if you were a mere mortal regarding the whale–ﬁsh controversy, then you would be familiar enough regarding the issues that, in order to know that Keiko (the whale) is a ﬁsh, you would need to have some way of neutralizing the whales-aren’t-ﬁsh hypothesis. Consider the advocate of the pain-only-in-the-brain hypothesis and compare what she says with what the whales-aren’t-ﬁsh advocate says (keep in mind that we’re pretending that all the sceptical error theories we plug into the argument template are false, so whales are ﬁsh and toes sometimes throb). . . . . .

. .

. . .

I realize that it’s common sense that whales are ﬁsh. I grant that there are loads of ﬁsh and loads of whales and to a certain extent there’s nothing wrong with classifying whales as ﬁsh. But they aren’t ﬁsh, not really. Whales are actually mammals, not ﬁsh, for various technical reasons. There are zillions of ﬁsh to be found, but they aren’t to be found among the whales. I realize that it’s common sense that toes sometimes throb. I grant that there are loads of throbbing feelings and loads of toes and to a certain extent there’s nothing wrong with classifying some toes as throbbing. But they aren’t throbbing, not really. The throbbing feeling is actually in your brain, not your toe, for various technical reasons. There are zillions of throbbing feelings to be found, but they aren’t to be found among the toes.

I take it as intuitive that in the envisioned whale–ﬁsh scenario mere mortals wouldn’t know that Keiko the whale is a ﬁsh—even though

The Sceptical Solution

81

they used the familiar, commonsensical, and highly reliable ‘If you ﬁnd a whale, then you’ve found a ﬁsh’ method to form the true belief that Keiko is a ﬁsh. By analogy, just because you (a mere mortal) have a true belief that Mary’s toe is throbbing and you formed it in a common way—see her stub her toe, curse, and tell you it’s throbbing painfully—this method, as familiar, commonsensical, and reliable as it indeed is (in that world and our world), is not sufﬁcient to make you know that Mary’s toe is throbbing. Not when you’re a mere mortal with respect to the live pain-only-in-the-brain hypothesis. Again: the reason for the lack of knowledge is exactly the same as in the whales-aren’t-ﬁsh case. The live sceptic is saying that our knowledge is sabotaged due to the operation of a perfectly familiar epistemic mechanism, the one governing the whales–ﬁsh or dinosaur cases. Now for the second and most interesting reason that the new sceptical conclusions are not so bad. Suppose once again that live scepticism is true. Suppose that tomorrow the deﬁnitive refutations of colour error theory and eliminativism are published and widely digested. It’s reasonable to think that we then know that ﬁre-engines are red and that Moore believed scepticism false, for at that point the error theories have been ruled out on our behalf. In addition, it is plausible to hold that since my father knows nothing of eliminativism, colour error theory, professional philosophy, or cognitive science, he currently knows that ﬁre-engines are red and that my mother believes that motorcycles are a menace. It’s also reasonable and consistent with the new sceptic’s position to think that my father knew those facts years ago, before eliminativism (and, let us pretend, colour error theory) was even on the radar screen. Finally, we should hold that before we were mere mortals, say when we were children, we had loads of knowledge inconsistent with the sceptical hypotheses even if the latter were already live. This is just to say that the reach of the epistemic threat posed by the live sceptical hypotheses isn’t sufﬁcient to affect those people for whom it isn’t live. In a way, this makes the new scepticism modest: by saying it rules out knowledge only for mere mortals, we are restricting its scope severely. In other ways, the new scepticism is now unlike anything we have ever encountered. Here’s why. Scepticism has always been considered a permanent, blanket, and highly negative condition. If you’re in it, then everyone else is in it too (that’s ‘blanket’), for ever (‘permanent’); and this

82

The Sceptical Solution

marks a signiﬁcant epistemic deﬁciency on our part (‘highly negative’). If scepticism is right, then it must reﬂect an everlasting, perhaps absolutely necessary, failure of our cognitive systems to achieve a certain result. The idea that some kind of radical even if restricted scepticism could be the result of a purely accidental, contingent, and temporary conﬂuence of chance social events affecting a fully rational, cognitively well-off individual seems absurd. And it is absurd, at least for classic versions of scepticism. But what we have seen is that scepticism can be thus accidentally generated. The scepticisms attributed in this essay are relatively ﬂeeting and are no indictment of our cognitive systems or evolutionary progress. Most surprisingly, the new scepticisms are the odd result of cognitive systems working well, not poorly: it is through the epistemically beneﬁcial practices embodied in the premisses of the new sceptical argument template that we have temporarily fallen into our regions of restricted scepticism. By being part of a community that pays due heed to expertly produced contrary evidence, we have temporarily robbed ourselves of large portions of knowledge, but we have done so as the result of following epistemic practices that almost always actually buy us knowledge. Once we understand how we got into the mess of not having knowledge of colour, belief, or pain location, for instance, we see that it’s not a bad thing. In fact, we are epistemically better off than we were before! Yesterday you knew; today you don’t; but today you’re smarter than you were yesterday regarding colour, belief, and pain. People unacquainted with philosophy or colour science or cognitive science may know more than we do, but this just shows that we need a new and improved measure of epistemic standing. Suppose I think that on balance colour error theory is very likely mistaken, and so even in the full awareness of the live status of that theory, I continue to hold ordinary colour beliefs. My mother’s belief that ﬁre-engines are red is warranted enough for knowledge; mine is not, as I remain a mere mortal; but in some sense I’m in the superior epistemic situation with regard to the colour of ﬁre-engines despite my lack of knowledge and lack of immortality vis-a`-vis colour error theory. That may sound contradictory: if token belief A (my mother’s) is better warranted than token belief B (mine), and A and B have the same content (e.g., that ﬁre-engines are red), then of course A is epistemically better than B. In the previous paragraph I had the live sceptic deny this conditional, but she need not. Perhaps the collection

The Sceptical Solution

83

of my attitudes regarding colour is epistemically superior to hers, even though when it comes to the particular ﬁre-engine belief her position is superior to mine. The details of how best to describe the differences in the two believers are interesting, but in any case the end result is that a typical reaction to scepticism—the kind of epistemic failure attributed to us by the sceptic just couldn’t be built into our cognitive systems—does not touch the live sceptical theses. Similarly for the objection that runs: according to the sceptic, we are cut off from reality, unable to know it, and that’s just implausible. On the contrary, by becoming a mere mortal and thereby falling into the live sceptical trap, I am in a better epistemic position than I was before; scepticism is an improvement, something to be bragged about, not ashamed of. Falling ‘victim’ to the live sceptical snare does not cut me off from the reality of colour, belief, pain, and knowledge compared to those non-philosophical folk who in a real sense know more than I do. Instead, I’m the one better in tune with the facts; my opinions dig deeper into the nature of reality than theirs do. This is not too surprising: by knowing more about the possibilities regarding colour and belief, I have a better appreciation of the ‘whole colour thing’, or the ‘whole story regarding cognition’. The ordinary person may know that ﬁre-engines are red, while I do not, but her belief has an accidental quality that mine will never have. If she had just been privy to the intelligent and sophisticated ruminations of some philosophers and scientists, then she would no longer know. It’s just an accident that she does know.28 The nature of the accidental quality is not so great as to rob her of knowledge, or so the anti-sceptic says and the live sceptic may admit (more on this point in a few pages), but coupled with my expanded knowledge of possibilities and evidence for and against those possibilities, it is enough to render her overall epistemic situation less secure than mine. Furthermore, I’ve got all the positive warrant that she does when it comes to our beliefs that ﬁre-engines are red. This point is so important I’ll put it in a box, as if it were in a science textbook. The live sceptic is not denying us any of the warrant we think we have; neither is she denying its quality. 28 It’s also just an accident that I don’t know, as I could very easily have been in a world in which colour error theory had never even been formulated. Whereas knowledge can’t exist if it is too accidental, not knowing can so exist.

84

The Sceptical Solution

Engage in the useful ﬁction that warrant comes in units (we can make the same point without the ﬁction). She has 1,200 positive warrant units for her ﬁre-engine belief; I have just 800; one needs 1,000 for the true belief to be knowledge. I have had all the same ﬁre-engine and colour experiences that she has had; I have each of her 1,200 units. My only problem is that I also have 400 negative warrant units, coming from my foray into colour science and philosophy.29 Just because she knows and I don’t does not mean that I lack some positive warrant that she has. Perhaps we need to judge the epistemic standing of beliefs by additional measures, such as positive warrants (I have all the good reasons, reliability, facts, etc. that she has) and absolute value warrants (since the absolute value of my warrants is much greater than hers, I’m much more familiar with the relevant issues regarding the content of the belief). We get the same result without thinking of competing warrants in quantitative terms. It seems plain that although as a mere mortal all my previous positive warrant for believing that ﬁre-engines are red is still in place and is still sufﬁcient, if left alone, for knowledge (that was the 1,200 ‘units’), there is negative warrant sufﬁcient to offset or cancel or ‘veto’ the positive warrant. It seems to me that some sources of negative warrant produce a veto of almost any positive warrant for the belief in question. If H is live, S is a mere mortal with respect to H, and H obviously entails :P, then H vetoes the warrant that S has for her belief in P, where ‘vetoes’ means ‘makes insufﬁcient for knowledge’. Thus, if you still believe that ﬁre-engines are red and you worry, at least a bit, about the possibility that you have fallen into the sceptical trap, be assured that this trap is quite unlike the traditional ones. You can hold your head high and boastingly declare your arrival in the pit of scepticism. If you can then actually crawl out of it under your own power, by ruling out the sceptical hypotheses in question, then you

29 It might be thought that we mere mortals are ‘too clever by half’ as a result of our familiarity with the sceptical hypotheses. However, we are imagining mere mortals who reject the sceptical hypotheses and continue to assert things that they understand to contradict the hypotheses; otherwise the new sceptical argument never even starts. So they aren’t too clever by half in the standard sense that they end up on the wrong path due to mistakes by the experts. The ones who are really too clever by half are the ones fooled into accepting the sceptical hypotheses. They are not my target group.

The Sceptical Solution

85

should get a distinguished chair in the philosophy department of your choice.30 These points on epistemic standing within the sceptical snare are bolstered by the third reason that live scepticism is palatable. Suppose that live scepticism is true, so you, the mere mortal, don’t know that the table is brown; neither do you know that you believe that. Still, no one said that you didn’t know perfectly well that the table looks brown and is for all practical purposes brown.31 Further, you still know that red is darker than yellow. In addition, you still know that you act like you believe P, that it seems to you that you believe P, that you feel like you believe P, that it seems as though Mary believes Q, that she behaves as if she believes Q, etc. You also know that for all practical purposes Moore believed scepticism false. And you know that belief and truth are required for knowledge. If that’s the case, then what have you lost, really, in losing your knowledge that ﬁre-engines are red? Before becoming a mere mortal, you knew that ﬁre-engines are red; now you don’t; but you’re now familiar with much more colour theory than you were, and you still know that ﬁre-engines look red and are red for all practical purposes. Now who is in the better epistemic position vis-a`-vis the redness of the ﬁre-engine, the old you or the new you? Still, one might ﬁnd it odd that the uninformed know, while we mere mortals fail to know. Gilbert Harman (1980) and others have investigated so-called social Gettier cases that are somewhat similar to the cases I’m focusing on and that may seem to put pressure on my suggestion that non-mere mortals know while mere mortals don’t know. Jane and Jon read in the newspaper that Dr Kirby has been assassinated. The story is a true account by an eyewitness. On the basis of that report both Jon and Jane come to believe that Dr Kirby was assassinated. However, starting the next day there is widespread publicity of credible (but ultimately misleading) denials of the 30 The contextualist who admits that in certain sceptical contexts ‘I know I have hands’ is false (e.g., DeRose 1995) can use some of this reasoning to block the objection that she is giving in too much to the sceptic. In fact, the scepticism that the sceptic-friendly contextualist accepts is quite similar to that of the new sceptic. It affects only a few people (philosophers mainly); it is temporary (only in some very specialized contexts is it correct to say ‘She doesn’t know P’); and even while in those contexts in which knowledge claims are false, the subject has all the usual positive warrant for her beliefs—the positive warrant had by people who do know the facts in question. However, the new sceptic’s scepticism is much grander than the contextualist’s, for reasons the reader will be able to list herself. 31 Here of course I’m assuming that the traditional sceptical arguments fail.

86

The Sceptical Solution

assassination by well-placed public ofﬁcials (who are lying). Jane is unaware of those denials, as she didn’t read the newspapers that day; Jon did read them. Does Jane still know the assassination occurred? Jon is a mere mortal with respect to the no-assassination idea; Jane is the non-mere mortal as she is totally uninformed of that hypothesis (and we can pretend that she has no special knowledge that could show the denials to be mistaken). If we tell the story right, Jon doesn’t know the assassination took place even if he retains his true belief. Presumably, Jane doesn’t know either (although many people’s intuitions are not so strong on this matter). If so, then the live hypothesis rules out knowledge for the non-mere mortal as well as for the mere mortal, contrary to my suggestion. However, the live sceptic’s case is different in at least one crucial respect, one that Harman himself notes. Jane is a member of the epistemic community in which the denials are public. She read about the assassination in the New York Times on April 5th; the denials were published in the same paper on April 6th. She is epistemically responsible to the April 6th information in the sense that it affects her belief ’s epistemic status. However, she doesn’t read any philosophy or science. She knows absolutely nothing of the philosophy of colour or colour science. She isn’t part of those epistemic communities. So there is less reason to think that the colour error hypothesis poses a threat to her belief that her socks are blue as there is reason for thinking that the no-assassination hypothesis poses a threat to her belief that Dr Kirby was assassinated.32 Furthermore, the method she used to obtain her belief in the assassination is the very same method for obtaining the defeater for that belief: relying on the newspaper’s reports. This isn’t true in the colour case: she relies on a relatively primitive perceptual experience to obtain her socksare-blue belief and would have to endorse a completely different method for obtaining evidence against that belief (i.e., relying on the testimony of expert colour scientists or philosophers). In any case, I’m not claiming that live scepticism does not apply to people who know nothing of the error theories. I’m just allowing for that relatively likely possibility. If it does apply to them, then live scepticism is even stronger, covering a much wider range of beliefs, although it still doesn’t affect the millions of beliefs not targeted by live contrary hypotheses. 32

I say a bit more about cases like this one later in this section.

The Sceptical Solution

87

The fourth and last reason for not getting too upset over the sceptical conclusion lies in the live sceptic’s response to an intriguing anti-sceptical argument. Even a sceptic has to admit that we divide up knowledge (and justiﬁcation) claims into two huge piles, ‘Appropriate’ and ‘Inappropriate’, according to relatively stable, crosscultural, intersubjective criteria. Further, we agree on what goes in the piles, at least for an important class of ‘commonsensical’ claims; call those claims ‘core’. Common sense says that for those core claims—and there are lots of these claims—the ones that show up in the pile marked ‘Appropriate’ are virtually all true, and the ones that show up in the other pile are virtually all false. Now revisionist theory T comes along, and it disagrees wildly with common sense, at least in one respect. It agrees with common sense that we sort the knowledge claims into the two enormous piles, ‘Appropriate’ and ‘Inappropriate’. It agrees that we do this according to some stable criteria. It agrees with common sense regarding what core claims are in fact sorted into which piles. That is, when common sense says that core claim C gets sorted, as a matter of brute fact, into ‘Appropriate’ or that core claim D gets sorted into ‘Inappropriate’, theory T concurs. What T vigorously disputes is common sense’s claim that virtually all the core ‘Appropriate’ knowledge claims are true, and virtually all the core ‘Inappropriate’ knowledge claims are false. Regarding a signiﬁcant and uniﬁed portion of the core claims that common sense puts in the ‘Appropriate’ pile, the sceptical revisionist says that they are false. To take a concrete example that targets the live sceptic, it’s just plain weird that Sam the mere mortal regarding colour is just as competent regarding colour as Samantha the non-mere mortal, and yet Samantha but not Sam knows that some ﬁsh are purple—as they personally encountered the ﬁsh while scuba diving together, and both believe it in the usual way based on those veridical perceptions. The two attributions ‘Sam knows that some ﬁsh are purple’ and ‘Samantha knows that some ﬁsh are purple’ seem equally good in some important and epistemically central sense. They are ‘warrantedly assertable’ to the same extent. That Sam knows that some ﬁsh are purple is the judgement of common sense and common language. These judgements are not to be dismissed frivolously. We end up with several questions the live sceptic needs to address:

88 . .

.

The Sceptical Solution

Why is it that ‘Sam knows that some ﬁsh are purple’ is false, as the live sceptic claims? In what sense are ‘Sam knows that some ﬁsh are purple’ and ‘Samantha knows that some ﬁsh are purple’ equally good, even though, according to the live sceptic, only the latter is true? Why is it that ‘Sam doesn’t know that some ﬁsh are purple’ seems inferior to ‘Sam knows that some ﬁsh are purple’, even though, according to the live sceptic, the former is true and the latter is false?

These are good challenges. The traditional sceptic has problems answering the analogous questions for, e.g., ‘Sam knows he has hands’. But I think that the live sceptic can answer them. Regarding the ﬁrst question, the live sceptic’s sceptical argument provides the explanation, as I was trying to indicate in the ﬁrst of these four reasons why live scepticism is palatable. That argument says that Sam doesn’t have knowledge because of a wholly familiar mechanism: losing knowledge in the face of recognized expert contrary evidence and opinion. The phenomenon is perfectly ordinary. The other two questions require answers that relate truth and the conversational appropriateness of knowledge attributions. We begin to answer them by noting that the phrase ‘knowledge attribution’ is ambiguous in the following, relevant way. It may stand for the truth or falsehood attributed, the proposition that S knows P, or it can stand for the action of attributing knowledge: the attributing of a piece of knowledge. Whether an attribution is conversationally appropriate is obviously dependent on whether it serves certain purposes, especially the ones peculiar to the context of conversation. That’s pretty much what ‘conversationally appropriate’ amounts to. Now one purpose is to have the proposition expressed by the use of ‘S knows P’ true. But there are other purposes, and often those are more important. (Think of the idiosyncratic ways in which you converse with someone you’ve known for many years.) So a use of ‘S knows P’ is conversationally appropriate if and only if it serves the purposes relevant to that use. We cannot complain that the right-hand side of the biconditional in the previous sentence is vague: it’s that vague because the left-hand side is that vague. The live sceptic can employ the action–proposition distinction to help answer the two questions raised above (where we assume that the scuba-diving subject Sam correctly believes, in the usual

The Sceptical Solution

89

experiential way, that some ﬁsh are purple). For convenience, let ‘practicality’ be short for ‘conversational appropriateness’. 1. Any use U of ‘Sam knows that some ﬁsh are purple’ expresses a false proposition when uttered by another mere mortal. This is a sufﬁcient condition for live scepticism for colour (given the quality of Sam’s epistemic situation relative to that of other people). This says nothing regarding the practicality of the action U. 2. A use U of ‘Sam knows that some ﬁsh are purple’ is an action that is practical if and only if U serves the purposes relevant in U’s context. 3. As a matter of brute fact, for zillions of potential uses U of the sentence ‘Sam knows that some ﬁsh are purple’, U is practical. That is, those knowledge claims succeed at being practical. These practicality facts are the truths that common sense is tracking. The sceptic can say that a use of ‘Sam knows that some ﬁsh are purple’ is ‘warrantedly assertable’ if and only if practical, if and only if it’s conversationally appropriate, if and only if it serves the purposes relevant in U’s context. 4. Why is (3) true? Why do all those false knowledge claims succeed in being practical? The answer has three parts: (a) Sam has all the usual good reasons for believing it that people who know it have. More generally, Sam has all the positive warrant that Samantha has. (b) Sam has all the usual abilities you would expect in someone who knows it. For instance, he can point out the purple ﬁsh, distinguish them from red ones, tell you that they look a different colour when exposed to sufﬁciently different light, etc. (c) The attributor of the knowledge claim ‘Sam knows that some ﬁsh are purple’ most likely isn’t currently thinking about colour error theory, and so does not currently and consciously appreciate how it threatens Sam’s belief. The same very likely holds for her audience. Since that threat is the only signiﬁcant threat to Sam’s belief, it follows that the only signiﬁcant threat to Sam’s belief is very likely to be currently and consciously out of everyone’s mind—thus contributing to the appropriateness of the knowledge attributor’s action. 5. A use U of ‘Sam knows that some ﬁsh are purple’ seems to express a true proposition (even though it doesn’t) because: (a) U expresses an action that is practical (as explained in (4)).

90

The Sceptical Solution (b) We don’t usually distinguish the practicality of the action from the truth of the proposition, especially when (4c) is true, which it is in virtually all ordinary contexts. (c) The practicality of the action is often dependent, but only in part, on the truth of the proposition. (d) A use of ‘Sam doesn’t know that some ﬁsh are purple’ seems to express a false proposition.

6. Why does a use U* of ‘Sam doesn’t know that some ﬁsh are purple’ seem to express a false proposition, as noted in (5d) (even though it expresses a true proposition)? Because: (a) U* expresses an action that is impractical. (b) We don’t usually distinguish the impracticality of the action from the falsehood of the proposition, especially when (4c) is true, which it is in virtually all ordinary contexts. (c) The impracticality of the action is often dependent, but only in part, on the falsity of the proposition. 7. Why is a use U* of ‘Sam doesn’t know that some ﬁsh are purple’ impractical, as noted in (6a)? The answer has three parts: (a) It strongly suggests to us that Sam doesn’t have all the usual good reasons (more generally, warrant) for believing it that average people who know it perfectly well have. But that claim is false; Sam has the usual good reasons. Thus, since U* strongly suggests something that is false, and recognizably so, we think U* is false; and that makes U* impractical. (b) It strongly suggests to us that Sam doesn’t have all the usual abilities you would expect in the average person who knows perfectly well that some ﬁsh are purple. But that claim is false; Sam has all those abilities. Thus, since U* strongly suggests something that is false, and recognizably so, we think that U* is false; and that makes U* impractical. (c) If someone did employ a use U* of ‘Sam doesn’t know that some ﬁsh are purple’, the audience would most likely not be currently thinking about colour error theory, and so would not consciously appreciate how it threatens Sam’s belief. Since that threat is the only signiﬁcant threat to Sam’s belief, it follows that the only signiﬁcant threat to Sam’s belief is very likely to be absent from everyone’s mind—thus contributing to the inappropriateness of the use of U*.

The Sceptical Solution

91

8. A use U of ‘Sam knows that some ﬁsh are purple’ seems to express a false proposition—but only to those of us who have thought hard about the new scepticism, and only in some of our moods— because, well, U does express a false proposition and we can easily come to suspect that fact by thinking hard about the new sceptical arguments. What we have here in (1)–(8) is a good part of an explanation of how the sceptic can stick to her scepticism while admitting the undeniable practical irrelevance of her result as well as the undeniable utility of the false knowledge attributions. It’s important for us to be clear that the explanation does not rely on some controversial distinction between semantic content and pragmatic content—in fact, it relies on no such distinction, controversial or not. Neither does it rely on any controversial claims about implicature. It does employ implicature-like claims as described in (7a) and (7b), but these are entirely commonplace as well as highly intuitive. Obviously, ‘S doesn’t know P even though P is true and S believes it’ often strongly suggests that S doesn’t have the warrant ordinarily sufﬁcient for knowing P; and just as clearly it suggests that S doesn’t have the abilities you’d expect someone to have who does know that P. There is nothing mysterious in those suggestions. Further, it is clear that the connections in (7a) and (7b) are defeasible, in that it is easy to imagine circumstances in which their suggestions prove false (i.e., ‘S doesn’t know P’ is true even though S has all the usual warrant and abilities of someone who does know P). Many ‘warranted assertability manœuvres’ are pretty suspect because they rely on controversial semantic–pragmatic content distinctions and controversial implicatures. Fortunately, my explanation avoids those problems. Here is an uncontroversial and analogous use of (7), loosely based on examples present throughout the literature. Suppose that Senator Smith is in Bigtown, Smith’s aides have told the media that Smith is in Bigtown (the media aren’t in Bigtown, so they don’t have direct evidence), and the media have dutifully told the masses that Smith is in Bigtown. George has seen the newscasts and read the relevant newspaper articles saying that Smith is in Bigtown. But while he was seeing those newscasts and reading those newspapers, his sister, who is one of Senator Smith’s aides, told George that although Smith’s people were going to tell everyone that Smith would be in Bigtown

92

The Sceptical Solution

that weekend, in reality Smith was going to be in Bribetown on secret business. George’s sister has let him in on secrets like this before, and she has always been proved right later, when the media found out about the deception. This time, however, George doesn’t believe his sister. He believes that Smith is in Bigtown, based on what the media have reported. He disregards his sister’s assertions because, perhaps, he unwisely thinks that Smith wouldn’t take bribes. He knows that most senators take bribes all the time, but for some bad reason he suspects that Smith wouldn’t do it (which is false). It turns out that Smith is in Bigtown, because the secret business in Bribetown fell through at the last minute, so Smith didn’t go there at all (George’s sister didn’t know of these last-minute developments). It’s pretty clear that George doesn’t know that Smith is in Bigtown even though it’s true, he believes it, and he has all the positive warrant for his belief that the average person does—the average person who knows that Smith is in Bigtown based on the warrant that George has. He doesn’t know because he has the expert testimony of his sister, her perfect track record on this sort of thing, his knowledge that most senators take bribes, etc. And this isn’t a case of someone ﬁrst knowing P and then losing knowledge because of the emergence of good but ultimately misleading contrary evidence that some people may have but the subject knows absolutely nothing about. That kind of case, an instance of which is Harman’s assassination story given earlier, is controversial, as that subject is unlike George, in that she knows nothing about the contrary evidence.33 Since George got his sister’s advice at the same time as he was watching television and looking at the newspapers (his sister was there in the room with him), he was never in any position to come to know that Smith is in Bigtown. The assertion ‘George doesn’t know that Smith is in Bigtown’ is true but a bit misleading. And it’s misleading because it suggests that he doesn’t have the warrant that people typically have for knowing that Smith is in Bigtown—seeing the television and newspaper reports, for instance. That is, that sentence suggests that he didn’t see those news reports, or if he did see them, he didn’t believe them. But he did see them, and he understood them perfectly well, and even came to believe them. So ‘George doesn’t know that Smith is in Bigtown’ is quite misleading in many ordinary contexts, even though it is uncontroversially true. That’s (7a). (The other claims in (1)–(8) 33

See Lycan 1977; Harman 1980 (especially n. 7); Ginet 1980; and Conee 2001.

The Sceptical Solution

93

apply to George as well.) The same holds for ‘Sam doesn’t know that those ﬁsh were purple’: he did see the ﬁsh, he understood what was going on when he saw them, and he came to believe that the ﬁsh were purple. So ‘Sam doesn’t know that the ﬁsh were purple’ is quite misleading but true anyway—just as in the case of George. We could tell an analogous story for (7b). I showed how (1)–(8) work in a one-off scenario: the Smith-inBigtown story. Let me show how (1)–(8) work in a different sceptical scenario, one in which they have very wide application (as the live sceptic wants). I do this in order to prove that they can serve as an explanation of a widespread practice of using false knowledge claims. I also do it because it’s based on an interesting philosophical question that has not been asked before. It’s about brains in vats. Many philosophers have explored the question ‘What would you know if you were a BIV?’, but the question I want to investigate is ‘What would you know if your friends were BIVs?’ Not all your friends, but a few anyway. The scenario is this. BIVs become technologically and ﬁscally possible; a team of mad scientists ends up ruling the world; certain top secret documents detailing plans for massive kidnapping and BIVing are discovered and made public; impenetrable 100,000 square mile complexes of laboratories are set up in the wilds of Canada, Brazil, and elsewhere; thousands of people start mysteriously disappearing in the middle of the night to become BIV’d. . . . Well, you can see how the BIV hypothesis—‘I have become a BIV as the result of these crazy mad scientists’—could become something close to a real, live possibility for most any person (not the mad scientists themselves, provided we tell the story right). One morning you wake up and although pretty much everything seems the same, you have a nagging suspicion that something is not right. And many of the ‘people’ around you start to lose interest in the theory that the mad scientists are BIVing people. Years go by—or so it seems. Do you know that you have two hands? Do you know how long you have been married or how many bicycles you own? You might be the kind of epistemically cautious person who at least tries to refrain from believing most empirical things. It’s a good question how much success such a person could enjoy. But, for the sake of clarity, suppose that you conclude that the whole BIV thing is just some silly conspiracy theory or at least incredibly improbable. You’re wrong, and if you were clever, you would not have that headin-the-sand view, but you’re a bit pigheaded to begin with. So you

94

The Sceptical Solution

have many standard empirical beliefs, many of which are true and formed in the same way in which we form our beliefs in our BIV-less world. You are not currently a BIV, have never been a BIV, and will never be a BIV. Do any of your empirical beliefs such as ‘I have two bicycles’ amount to knowledge? Many of your friends have been BIV’d. You recall your neighbour disappearing one night; the same happened to your former boss. For all you know, it has happened to you as well. It hasn’t, but given the particulars of the story, including the fact that it has happened to many people, including some of your friends and relations, it’s a live possibility that it has happened to you. Under these conditions what do you know about your world? In that world you believe that you have two cars. And you’re right: you do have two cars, and you’re not a BIV. Even so, if we tell the story appropriately (use your science ﬁction skills here), it seems that in such a situation you wouldn’t know that you have two cars. How could you? You’d have to rule out the live possibility that you are one of the unlucky ones who has been BIV’d (and stripped of possessions). A successful ruling-out procedure need not be impossible. The BIV process isn’t perfected yet, and the scientists are only so clever in deciding what imaginary world to electronically ‘feed’ you. The secret documents correctly reveal that the mad scientists cannot fake certain activities, such as sophisticated philosophizing, and so the philosophy test, although quite difﬁcult, is total: if you learn enough philosophy, say by getting a Ph.D. and going to Rutgers University, then you may engage in a foolproof test: you are not envatted if and only if your experiences at Rutgers are of talking with their philosophers in a coherent and conversational manner about relevant alternative theories, for instance. Other tests could work too; e.g., talking conversationally with great poets or mystics (i.e., whatever the scientists can’t fake). So it is deﬁnitely in some people’s power to test ‘The mad scientists have envatted me’. In addition, you could seek out key people from your past who know things about you that no one else knows. By having the right conversations with them, one could determine that one is not envatted, as it is incredibly unlikely that the scientists would have ﬁgured out a way to fake that, given how slowly BIV technology has progressed in the past. (You recall the early days, thirty-ﬁve years ago, when BIV technology offered only ﬂeeting amusement for the very rich in the form of a small palate of programmed evil pleasures.) And if no one seems to know much of anything about your past, even your mother, then you

The Sceptical Solution

95

can safely conclude that you are envatted. Learning that you are envatted seems hard, but given the right circumstances, you could prove it beyond a reasonable doubt. No perfect guarantees are present for proving that you aren’t envatted, only good probabilities, but that’s par for the course in an empirical matter such as ‘The mad scientists have envatted me’. Here it won’t do to complain that the BIV possibility is crazy, remote, empty, unprovable, and therefore irrelevant—i.e., an alternative you don’t have to rule out in order to know ordinary things. It isn’t crazy and remote: it’s true for millions of people who have been envatted already and who are no different from you in any way that is relevant to why they were chosen for envatment. The BIV possibility is not empty: there is good evidence for it, in that you’ve seen the secret documents published everywhere, heard the stories of the populations of whole cities disappearing overnight, seen the initial, public BIVs twenty years ago, witnessed the sudden and unsolved disappearance of friends and relatives and even whole neighbourhoods, now feel as though your own life is strangely different, your philosophy thesis adviser doesn’t seem able to discuss philosophy in anything except full paragraphs as if from a book (that’s the best the scientists can do to mimic real philosophizing), etc. And as we saw earlier, it is empirically testable. You could rule out the BIV possibility if you learned enough philosophy and went to the right place and talked with the right people, but you don’t have the philosophical ability just yet. Perhaps you can be justiﬁed in thinking that you haven’t been envatted in the last few seconds, since that is so unlikely; but that is another matter. Do people in that BIV-live society know things like the fact that Pete Sampras won the men’s singles title at Wimbledon that year? Presumably mad scientist Max ruling over everything could know that Sampras won Wimbledon simply by watching the match on television, just as he can rule out the possibility that he is envatted. But could head-in-the-sand conspiracy-scoffer Jack, watching the match on television, know it? Max is in a much better position than Jack: Max can rule out the possibility that he is a BIV. And Jazmine, who isn’t one of the ruling mad scientists but has Ph.D.s in both philosophy and religious studies, has actually ruled out the BIV possibility for herself by conversing with many top philosophers, poets, and mystics, as well as interrogating many people from her past. Jack has not done any of those things. As a mere mortal

96

The Sceptical Solution

in the BIV-live society, Jack has not ruled out the possibility that he has become envatted. He could if he put in the effort and had some luck, but the task is formidable. Jazmine and Max have neutralized their BIV possibilities; Jack has not.34 There are many interesting questions to ask about such a scenario. But the one I want to consider regards how we would talk about the people in that society, observing them from our spaceships. When we imagine ourselves commenting on that society, we place our ‘knowledge’ claims about them into two piles just as they do, according to the same principles or rules of thumb they use; call those piles ‘acceptable’ and ‘unacceptable’. Watching television, Mary has seen the whole Sampras match plus the award ceremony; Jack saw just the one ace that was match point, with the volume turned off and no words on the television screen, and then immediately turned off the television and concluded, quite rashly, that Sampras had won (for all Jack knows, what he saw was a replay of another Sampras match, or just a set point but not match point of the actual match). When describing what is happening, we would say, ‘Jack doesn’t know that Sampras won but Mary does’, but we wouldn’t actually believe that Mary knows that Sampras won. She doesn’t, as she has done nothing to rule out the real, live, testable BIV possibility. When we say to each other, ‘Mary knows that Sampras won, but Jack doesn’t’, we don’t intend the ﬁrst part of the claim to be taken literally. Instead, we have either an acceptability judgement—Mary’s but not Jack’s belief is acceptable in virtually all the ways that knowledge claims count—or a non-literal use of ‘S knows P’, really meaning something like ‘For all practical purposes, S knows P’ (more on the latter in Chapter 10). Neither Jack nor Mary knows that Sampras won, as Mary is just like Jack in not having done anything to rule out the BIV hypothesis. Jazmine, the super philosopher-mystic-autobiographer, knows that Sampras won, as does the mad scientist Max, but virtually no one else in the society does, even though Sampras really did win, and the television broadcast was entirely legitimate. 34 I just asserted that Jazmine and Max have done things—complex empirical investigations in Jazmine’s case—in order to rule out the BIV possibility. This makes it sound as though Jazmine has single-handedly defeated traditional scepticism—she ruled out the BIV hypothesis!—which of course is false. What she has done is rule out some BIV hypotheses, but not the really tough ones, the ones in which the agent behind the deception isn’t an actually existing group of scientists but a god who can make the deception perfect. Descartes was right: we need to appeal to supernatural powers in the classic sceptical hypotheses (evil demon, dreaming), in order to make them do the work they are created for.

The Sceptical Solution

97

‘Mary knows that Sampras won’ is just like ‘Sam knows that some ﬁsh are purple’. In particular, in both cases the attribution-action is practical, but the attribution-proposition is false. With only slight modiﬁcations, (1)–(8) apply to our ‘Mary knows that Sampras won’ quite easily, and explain just why it is that we think the knowledge claim is false but continue to employ it in ordinary life (when talking about the BIV-live society). I don’t pretend to have told the whole story here, but I hope I have said enough to assuage fears that the live sceptic has nothing to say to the Sam–Samantha challenge articulated a few pages back. Now the live sceptical results look odd, but not debilitating; strange, but not indicative of any signiﬁcant epistemic failure. Does that make the results philosophically uninteresting? Is live scepticism dull? Of course not: you don’t know that you believe that Moore thought scepticism false; neither do you know that ﬁre-engines are red, pains occur in chests, or that John Rawls was kind (recall character trait scepticism). Despite all that, you are in a better epistemic position than those who aren’t in the sceptical traps (other than the immortals of course). You would never have accepted any of those claims without ﬁghting tooth and nail; you would have laughed at them. So the live sceptical claims are philosophically interesting. There is more on this point in the ﬁnal chapter. C Outrage: Sceptical Hypotheses Sabotaging Knowledge What seems outrageous, and probably correctly so, is the claim that the bare existence of some crazy hypothesis ruins your warrant for your beliefs. How could such a nutty, irresolvable, and scientiﬁcally baseless hypothesis such as the BIV hypothesis take away your knowledge? Indeed, that would be shocking. In fact, it’s a good reason for thinking that contextualism—or any other theory—is mistaken in giving in to the classic sceptical arguments in any actual context, even ones having to do with sceptical epistemology. But none of this holds for the live sceptical hypotheses. Those hypotheses are live, scientiﬁc hypotheses. So it is entirely reasonable that their status as live and obviously inconsistent with P means that you need to rule them out in order to know P. In the remainder of this section I would like to show that sceptical arguments akin to the new ones can overthrow knowledge of colour,

98

The Sceptical Solution

belief, and pain location. This is important, I think, because it is reasonable to doubt that any set of circumstances could overthrow such fundamental and effortless knowledge. Overthrowing George’s knowledge of Senator Smith’s whereabouts or the student’s knowledge of the cause of the demise of the dinosaurs is one thing; overthrowing knowledge of the colour of your socks is something else entirely, as the latter seems much more secure and immediate, given that we’re assuming that the socks are coloured. If I come to a true belief regarding simple colour, belief, and pain location facts and do so for the ‘right’ reasons or experiences—the reasons or experiences that in ordinary circumstances are abundantly sufﬁcient for knowledge—then how could I not have knowledge? It’s obvious that this can happen in other cases (e.g., the cases of George and the student palaeontologist), but can it happen in the case of one’s own beliefs or sock colours? Are there any clear and realistic examples in which a live hypothesis overthrows one’s knowledge of one’s own beliefs or sock colours? It would be nice to see some examples before we subscribe to the live sceptic’s idea that there are worlds in which vast amounts of such knowledge are overthrown due to live sceptical hypotheses. Before you convince me that there are dozens of people over ten feet tall, I’d like to see at least one. You see a sock in the usual excellent viewing conditions. It looks, and is, blue. But it’s your colleague’s sock, and his wife is a colour scientist, and he insists that he is wearing some of her ‘trick’ socks she uses in her experiments, in that although they look blue and normal, they’re actually very weird and really green. We can suppose that he’s made an innocent mistake, in that the socks he is wearing are entirely normal and blue. You mistakenly think that he is trying to fool you, even though he’s actually a pillar of honesty, so you stubbornly persist in your belief that the socks are blue. Suppose his wife comes in and says, ‘Well there are those trick socks! We were looking for them all morning in the lab! What are you doing with them on?’ Other people concur with her (her lab assistants and children, say). She and other colour theorists have created various other strange objects, strange in ways having to do with their colour appearances. You are somewhat aware of these objects, involving rotating disks, unusual materials, and the like. So you know of the existence of such objects. Your blue-socks belief is true and reliably produced in the entirely ordinary way, but does this belief constitute knowledge once you’ve encountered the weird-socks story, especially given that

The Sceptical Solution

99

you’ve heard and understood loads of intelligent, sincere, and honest experts saying that the socks are really green—not just his wife, but her assistants, other professors, etc.? Don’t you have to rule out the weird-socks hypothesis to retain your knowledge? I think you would be in violation of some signiﬁcant epistemic duties if you retained your belief. Do you keep your knowledge in these circumstances, in spite of your epistemic violations? It seems not. You believe the truth, and you do so for the right reasons, but in circumstances such as these, that isn’t enough. We can tell analogous stories for some of the other sceptical hypotheses. During surgery you have a particular pain that seems to be, and really is, in your spine, but the neurologist asserts that this illusion happens all the time with this operation, and the pain is really in the brain as your spine is perfectly sound and not even being operated on. For each of these ‘one-off ’ cases we get the result that persisting in your belief is in serious violation of some epistemic duties. Further, it seems as though you no longer know that the socks are blue or that the throbbing pain is in your spine. If this is right, then we have one-off sceptical results, but sceptical results all the same, each based on highly realistic sceptical hypotheses and each denying instances of the knowledge denied by the live sceptic. We’re all aware of the clever devices made by colour scientists, so the trick socks are easily understandable; we’re aware that the spine and brain are intimately connected and that people feel pains ‘in’ phantom limbs. There’s nothing bizarre about any of these one-off sceptical hypotheses—limited to just one pain or one pair of socks. So it seems that we should accept these one-off sceptical arguments. And if we’re going to accept them, thereby accepting that one could have come to have the belief that P in the perfectly ordinary, upstanding manner and yet later not know that P—where these Ps are exactly the targets of the live sceptical arguments—then why should we balk at the conclusions of the more encompassing live sceptical arguments? For we now have precedents for the live sceptical results: relatively uncontroversial cases in which you once knew, in the ordinary upstanding way, that P, and then, due to expert contrary evidence, you no longer know that P—where the P in question is targeted by the live sceptical arguments. The eliminativist case is harder. That is, it is difﬁcult to ﬁnd a oneoff case in which you start out knowing that you believe P but then you clearly no longer know that you believe P because of a live,

100

The Sceptical Solution

familiar, realistic, and not philosophical hypothesis to the effect that you don’t have that belief. The problem lies in ﬁnding a case that satisﬁes the italics. Here are some potential cases. You believe that Rutherfordium is an element, but some authorities tell you that your concept corresponding to ‘Rutherfordium’ is a complete and hopeless muddle. ‘You’re so confused, you haven’t got any belief at all!,’ they all tell you. You don’t believe these expert critics—authorities in the sense that they are your chemistry teachers from your youth, so they should know whether your concepts are muddled, as they are the ones who introduced them to you. They tell you that when they taught you, they themselves were highly confused, and passed on all sorts of nonsense to their students, confusing the element with some plastic as well as with some mythical substance supposed to give eternal life. They even show you their lecture notes and the attendance records that show you were in those classes. Some classmates of yours testify in support of the claims of your teachers. It turns out that, unbeknownst to both them and you, you were not present during those lectures (you were ill and stayed home), and you received a perfectly proper education regarding Rutherfordium the next year in college. You stubbornly persist in your true belief that you believe that Rutherfordium is an element and ignore their excellent evidence to the contrary. You agree that you were in those classes when they claim that their instruction was so horribly confused. You also agree that if your conception were so muddled, you wouldn’t have Rutherfordium beliefs. (You agree with them, and perhaps with your philosophy teacher, that if someone has a completely muddled conception, then she can’t have beliefs involving that conception. It doesn’t matter whether you’re right about this.) But you simply refuse to believe that, as a result, you acquired a muddled concept. Don’t you have to rule out the muddledconception hypothesis in order to know that you believe that Rutherfordium is an element? You correctly admit that you have good reason to think that your Rutherfordium conception is a complete muddle; and you correctly admit that you have good reason, let us suppose, for thinking that one cannot believe something in such a muddle. Perhaps you can still know that you confusedly grasp or at least cognitively aim at the idea that Rutherfordium is an element, but in these circumstances how could you know that you believe it—cognitively grasp it in that special, competent way? You would need to neutralize the muddled-conception hypothesis.

The Sceptical Solution

101

But wait: can’t you rule out by introspection alone the hypothesis that your current Rutherfordium conception is muddled? Your teachers say that they taught all these silly things about Rutherfordium, but you don’t remember any of them. You consider your current Rutherfordium ‘attitudes’, and ﬁnd no traces of the confusion they passed on to their students. You conclude that you must not have been listening in class when the confusion was being spread, or you were absent those days, or perhaps your conception started out muddled but somehow got straightened out soon afterwards. No amount of interrogation now can show that your current conception is muddled. A philosophy student who has a muddled conception of supervenience may insist that her conception is ﬁne, which shows that just because someone judges that their conception isn’t muddled doesn’t mean that they’re right. But that’s a case of someone not knowing she has a coherent conception because she doesn’t have such a conception. In the Rutherfordium case we are looking for a situation in which someone doesn’t know she has a coherent conception even though she does have a coherent conception and believes that she has such a conception. I’m not sure about this case. I think the eliminativist sceptic is advocating the strongest of the new sceptical arguments, so it isn’t surprising that it’s tougher to ﬁnd supporting arguments for it. The Rutherfordium story does establish that you can at time T know that at T you believe that P, while later on, even though you correctly believe that at T you believed that P, you don’t at that later time know that at T you believed that P (even though you aren’t demented, haven’t forgotten anything, etc.). That’s a nice result, but not what we need for a one-off case that can act as a precedent for the live sceptic’s arguments. The eliminativist sceptic wants to argue that a live hypothesis can take away your present knowledge that you presently believe that P. So we’re trying to ﬁnd a relatively uncontroversial oneoff case in which you ﬁrst know that you believe that P, but then, due to misleading developments, your present true belief that you presently believe that P no longer counts as knowledge, because you can’t rule out the live hypothesis that you don’t presently believe that P. Consider another example, one that comes closer to doing the trick. Someone convinces our scuba-diver Sam that one doesn’t believe something unless one is positive that it’s true. There is no such thing as believing something a little bit or half way or even 95 per cent; you either believe it 100 per cent, or nothing; there is no

102

The Sceptical Solution

partial believing or degrees of believing. (A sharp friend of mine, who isn’t a philosopher, actually holds this view on reﬂection.) Many experts agree, and Sam is aware of this. Now scuba Sam begins to wonder whether he really believes that those ﬁsh were purple. He knows that he doesn’t feel certain that they were purple, since when he saw them, he was underwater, and he isn’t certain whether that affected the viewing conditions in some way so that the ﬁsh looked purple but really weren’t. His friends include a cognitive scientist and a philosopher of mind, who agree that belief is a very special cognitive relation and doesn’t come in degrees. They listen to Sam describe his attitude regarding the ﬁsh’s colour. They agree that his attitude doesn’t amount to belief. Instead, they say, he has a mere hunch that the ﬁsh were purple, and hunches aren’t of course knowledge. Sam asks himself: do I believe that the ﬁsh were purple, given that according to my smart philosophy and cognitive science friends belief requires subjective certainty and I don’t have that certainty? He begins to have his doubts. But then Sam concludes his ruminations with the following. But I really do believe it. Yeah, I’ve got some reasons for thinking that maybe, just maybe the ﬁsh weren’t purple because I was underwater and all. I’m not 100 per cent sure. But in the end I believe that they were purple. I don’t think the underwater conditions changed anything. I’m putting my money on the claim that they were purple, and that’s enough for belief. His expert friends say not. Believing is being 100 per cent positive or sure or subjectively certain, and Sam admits that he isn’t that positive or sure or subjectively certain. Sam doesn’t know what to say about the reasons his friends have given him; he just refuses to engage them and puts his foot down. In this scenario Sam does believe that the ﬁsh were purple, and he believes that he believes that the ﬁsh were purple. And before being accosted by his friends, he knew that he had the purple ﬁsh belief. But it seems as though after being accosted he doesn’t know that he has that belief. He has to rule out the belief-equals-being-100 per centcertain hypothesis. Do we have our precedent? The only problem here is that the belief-equals-being-100 per cent-certain hypothesis is not entirely innocent. It’s highly theoretical and philosophical. So it isn’t much better than the eliminativist hypothesis.

The Sceptical Solution

103

Here’s another attempt to ﬁnd a precedent. Someone wrongly thinks that you are in a Kripke puzzle (1979). You say that there are two philosophers ‘A. Goldman’ (Alan and Alvin). And you’re right. But Paul insists that it’s the same philosopher who has pursued many different lines of research (and it’s ‘Alvin’, as you must have misread ‘Alvin’ as ‘Alan’ at some point). Not only Paul says this: Pauline, Alexander, Alexandra, Joseph, Josephine, and many others concur. They even have independent evidence: they show you a journal article in which the author wrote something like ‘as Goldman has attempted to do in his publications (1978, 1982, 1983, and 1986)’, where the references indicate articles that you were convinced were written variously by the two Goldmans. (Of course, the journal article author is in a reverse Kripke puzzle situation, falsely thinking that there is just one ‘A. Goldman’ when there are two.) Your friends also convince you, or you convince yourself, that if you’re in a Kripke puzzle situation, thinking there are two when there is just one, then you don’t have any of those ‘Goldman’-ascribed beliefs, as your attitudes are too fouled up to satisfy the conditions for believing. You don’t believe your friends who say you’re in the Kripke puzzle, although you can do nothing to put any hole in their story. You’re just being stubborn. You really do believe that Goldman (Alvin) is an externalist, and you continue to believe that you believe that Goldman is an externalist—despite the fact that your friends are telling you that you don’t believe that Goldman is an externalist. In this story you once knew that you believed that Goldman is an externalist; that was your situation before your confused friends bothered you. But after your friends get through with you, although you continue to believe that you presently believe that Goldman is an externalist, you don’t know that you have that belief. As before, a live hypothesis has ruined your knowledge of your own current beliefs. You need to rule out the live possibility (i) that you are in a Kripke puzzle and (ii) that being in a Kripke puzzle cuts you off from the relevant beliefs. But once again, due to its lack of empirical support and highly philosophical nature, this hypothesis is not as secure as the ones for pain or colour given earlier. This is not to suggest that there is something controversial about the existence of Kripke puzzles. While it is hard to describe the mental states of someone in such a situation, it happens frequently enough. (I was writing a paper on Kripke’s puzzle with a friend, and while writing it he fell into a Kripke puzzle situation.) So (i) is innocent. But (ii) is not. Still not a perfect precedent.

104

The Sceptical Solution

The combined effect of the considerations in this chapter is the draining away of most of the motivations for rejecting the live scepticisms, or so I believe. After encountering the live sceptic, one can easily get embroiled in the fault-ﬁnding process, but the usual motivations for doing so are mostly absent. When confronted with traditional arguments for scepticism, one rejects them before one can put one’s ﬁnger on any really damaging criticisms of the arguments. The conclusions are outrageous, the hypotheses are outrageous, and the idea that the loopy hypotheses could ruin one’s ordinary knowledge is outrageous; and those three outrages are enough to justify our setting out on our investigation of the sceptical arguments with the conﬁdent attitude that the arguments are unsound. But what is it that is so bad about the new scepticisms? Where is the outrage? True, they deny you a huge portion of knowledge. But they ﬂatter you as well: your epistemic status is superior to that of the non-mortals; you fell into the pit of scepticism in virtue of entirely familiar and epistemically virtuous methods (the power of contrary and appreciated expert opinion); there are precedents for most of this lack of knowledge; and you can climb out of the sceptical pit if you become an immortal. The live sceptic says: what are you afraid of in the new scepticisms? What is it that is so unacceptable about them? There is of course the pull of the Moorean response here, the one that is captured by ‘Oh come on! Look at what you’re saying! Do you really think that the premisses of your new sceptical argument are more compelling than the claim that we know that ﬁre-engines are red?!’ The attraction of this response will be ripped to shreds in section B of Chapter 11. The burden of this chapter has been to show why live scepticism is not nearly as outrageous as traditional forms of scepticism. First, the sceptical hypotheses themselves are surprising, but nowhere near as crazy as the BIV hypothesis, for instance. The former are actually believed by many sober experts. Second, the live sceptic’s conclusions are reasonable as well, for four reasons. (a) The argument template that generates the sceptical conclusions is pretty familiar and sound. (b) On the commonsensical assumption that non-mere mortals can know what mere mortals cannot, the live sceptical conclusions are not permanent, blanket, or negative. In fact, mere mortals have all the positive epistemic warrant that non-mere mortals (but not immortals) have. (c) Mere mortals are free to have lots of other bits of knowledge closely akin to the bits denied them by the live sceptic’s

The Sceptical Solution

105

argument. (d) The live sceptic has a good response to the objection that ‘The mere mortal knows that P’ is just as good as ‘The non-mere mortal (but not immortal) knows that P’, even though the former is false and the latter true. Finally, there are precedents for most of the live sceptic’s conclusions—relatively ordinary, non-philosophical, and uncontroversial cases in which we fail to know the very truths that the live sceptic says we do not know, even though we believe them for the right reasons and in the ordinary, reliable way.

7

Universal Scepticism It is a curiosity of the philosophical temperament, this passion for radical solutions. Do you feel a little twinge in your epistemology? Absolute scepticism is the thing to try. . . . Apparently, the rule is: if aspirin doesn’t work, try cutting off your head. Jerry Fodor, Banish DisContent

The Holy Grail for the sceptic is universal scepticism, according to which you know nothing at all, not even that you have hands, that 2 þ 2 ¼ 4, that you exist, or even that it seems to you that you exist! Many people are attracted to philosophy in large part because of the nature of many philosophical puzzles: it seems that every solution must violate common sense in a radical manner. For them, arguments for universal scepticism can be particularly delicious. Amazingly enough, the new sceptical arguments can be modiﬁed in order to produce the conclusion that we mere mortals not only fail to know facts about beliefs and colours and chest pains and character traits, but that certain classes of mortals fail to know anything at all. However, this universal argument will employ a principle less plausible than the principles involved in the sceptical arguments given in previous chapters, at least on my accounting. Even so, many epistemologists will ﬁnd the principle compelling; furthermore, I can’t ﬁnd anything objectionable about it other than the fact that it leads to universal scepticism.35 In order for you to know P, you have to be able to rule out some scenarios that you’re aware are incompatible with your knowing P. The most obvious scenarios are of course ones in which P is false: P’s 35 There is of course a simple argument from scepticism about belief to universal scepticism: accept the KK principle that if you know P, then you know that you know P. That is, if you don’t know that you know P, then you don’t know P. Since we have argued against your knowing that anyone knows anything, the KK principle allows the inference to universal scepticism. But the KK principle is controversial.

Universal Scepticism

107

being false is obviously incompatible with you knowing P. For instance, in order for you to know that the tree is a hemlock, you have to rule out the possibility that it’s a spruce, because you’re aware that the tree’s being a spruce is incompatible with its being a hemlock. In order to know P, you need to rule out various :Ps because you’re aware that they are incompatible with you knowing P.36 But :P scenarios aren’t the only scenarios incompatible with your knowing P. Suppose you are in the process of waking up and seem to hear a bird singing outside your window. In order for you to know P, that a bird is singing outside your window, you need to rule out the possibility, Q, that you’re dreaming. But the dream possibility Q isn’t incompatible with P: the bird could be singing outside your window while you’re dreaming. Many philosophers, such as Moore, have noted this fact. Apparently, you need to rule out Q not because it conﬂicts with P (it doesn’t conﬂict), but because it conﬂicts with you knowing P—and it conﬂicts with you knowing P because it conﬂicts with you being warranted in believing P. If Q is actual, then your belief that P is unwarranted, as it is based on mere highly unreliable dreams, and so you don’t know P. So in order to know P, you need to be able to rule out Q. But, again, Q and P are perfectly compatible. The general principle suggested (but only suggested) seems to be this: in order to know P, you need to be able to rule out scenarios that you are aware are incompatible with your knowing P, although not necessarily incompatible with P itself. That covers the hemlock– spruce case as well as the dreaming–bird case (and some obvious cases involving hallucinations). Of course the principle is much more plausible if we make you aware of the fact that Q is a live possibility, as I did in the examples above. Since you were in the process of waking up, weren’t entirely sure if you were awake, and knew that you can easily dream of a bird singing, it’s a live possibility that you were dreaming the bird-song. Q is quite relevant to P; and you are aware of it. The same holds for the hemlock–spruce case: you’re perfectly aware of the fact that the trees are very similar in appearance and are found in the same forests. So the principle we really appear to be moving towards accepting is this: If you were aware of the fact that Q was a live possibility incompatible with your knowing P, then if you knew P anyway, you must have been able to rule out Q. 36

:Ps are claims inconsistent with P; they aren’t all negations of P.

108

Universal Scepticism

It’s no mystery how the live sceptic can use this principle to her advantage. Eliminativism is a live possibility incompatible with our knowing P, for any P. So according to the principle just formulated, we need to be able to rule out eliminativism in order to know anything. But we’re mere mortals; so we can’t rule it out. So we know absolutely nothing. Philosophers really are a bunch of know-nothings. This argument for live universal scepticism comes from a set of individually highly plausible but mutually inconsistent claims about a possible community very much like ours (one in which eliminativism is live) and a philosopher or cognitive scientist S in that community (who is a mere mortal with respect to eliminativism). 1. S knows some truth P (it doesn’t matter what P is). 2. S is a mere mortal with respect to E, eliminativism (the thesis that no one believes anything—all occurrences of ‘S believes P’ are false—and knowledge requires belief). 3. In S’s intellectual community E is live. 4. If S knows any truths, then S knows (the logical truth) that if S knows P, then :E (as S is a competent intellectual who has reﬂected appropriately on that very simple logical truth).37 In fact, if S knows any truths, then S has consciously reﬂected on this logical truth and is fully aware of it. 5. If S knows any truths, then S knows that E is live (as S is a competent philosopher of mind or cognitive scientist aware of what’s going on in her profession, etc.). In fact, if S knows any truths, then S has consciously reﬂected on this empirical truth and is fully aware of it. 6. A general principle: if (a) f consciously and reﬂectively knows and is fully aware that f’s knowing a entails :b, and (b) f consciously and reﬂectively knows and is fully aware that b is live, then given that f has put together in her mind (a) & (b), in order for f to know a, she must rule out (be able to rule out) b. This is similar to the Live Hypothesis Principle. 7. Another general principle: if b is live, and f is a mere mortal with respect to b, then f can’t rule out b. This is similar to the Modesty Principle. 37 This isn’t to say that it’s impossible to be a knower and not know that if S knows P, then :E. Rather, it’s to say that if forms of scepticism are false, so S knows things, then given S’s background, S knows that if S knows P, then :E. Given her background, if she knows anything at all, then she knows billions of truths, as she is a competent and ordinary adult. This holds for (5) as well.

Universal Scepticism

109

The conjunction of (1)–(7) entails a contradiction. Claim (1) is anti-scepticism about S, saying that S is a knower; (2)–(5) are true by the stipulations regarding S and her intellectual community. So if we want to hold that S knows anything, we have to hold that either (6) or (7) is false. The universal sceptic’s argument is just (2)–(7), with the negation of (1) as the conclusion. Given the deﬁnitions of ‘mere mortal’ and ‘live’, I ﬁnd it difﬁcult to see how either (6) or (7) could be false. There are implausible principles that are similar to (6) or (7), but once one realizes what ‘mere mortal’ and ‘live’ mean, one also realizes that (6) and (7) are pretty attractive. In particular, (7) isn’t saying that just because some famous people have believed that P and S can’t refute them, S thereby doesn’t know P. I don’t know of any counter-examples to either (6) or (7). Much of the importance of the preceding considerations is captured by (1)–(7). They form a new paradox, a set of individually highly plausible, yet mutually inconsistent, claims. There is a convenient way to resist this argument for universal scepticism. First I will spell it out. Then I will evaluate it. What I’ll call Traditional Ruling Out, or TRO, is the following principle (I’ve added some conditions that make the principle stronger than usual): If: A. S consciously and reﬂectively knows and is fully aware that P entails :Q, and B. S consciously and reﬂectively knows and is fully aware that Q is live, then, given that S has put together in her mind (A) & (B): C. In order for S to know P, S must rule out Q. The universal sceptic’s main principle, (6), has a quite different form, the difference showing up in the ﬁrst conjunct of the antecedent: If: D. S consciously and reﬂectively knows and is fully aware that S’s knowing P entails :Q, and E. S consciously and reﬂectively knows and is fully aware that Q is live, then, given that S has put together in her mind (D) & (E): F. In order for S to know P, S must rule out Q.

110

Universal Scepticism

This principle, (6), Universal Ruling Out, or URO, is quite different from TRO. TRO demands that one rule out some :Ps; URO says one has to rule out some :Ps as well as other claims inconsistent with ‘S knows P’—but these other claims need not be :Ps. So perhaps the lesson we learn from the universal sceptic’s argument is not that her conclusion is true but that whereas TRO is true, URO is false: although we do have to rule out at least some of the scenarios in obvious conﬂict with knowing P (e.g., the contextually relevant ones), the only scenarios that need to be ruled out are the ones in obvious conﬂict with P itself, not with ‘S believes P’. It is well known that G. E. Moore endorsed claims that entail URO (1962: 240–6). Barry Stroud has done so as well: As soon as we see that a certain possibility is incompatible with our knowing such-and-such [but not necessarily incompatible with such-and-such itself], it is suggested, we immediately recognize that it is a possibility that must be known not to obtain if we are to know the such-and-such in question. (1984: 21)38

Both Moore and Stroud are well aware that the possibility in question need not be incompatible with the proposition alleged to be known. Please note that URO is much more plausible than the principles endorsed by Moore and Stroud, for neither of them require liveness, awareness of liveness, or simultaneous and fully conscious awareness of the subject’s belief in P, the entailment, and the liveness. Keith DeRose ﬁnds intuitive, but does not go so far as to endorse, an idea quite similar to URO. [T]he BIV hypothesis seems to undermine my putative knowledge of I am in Houston as well as of I have hands, but, of course, that I am a bodiless BIV is compatible with my being in Houston. Perhaps if S is to know that P, then S must know that not-Q for any Q (but here restrictions must be added [e.g. contextual relevance]) such that if Q were true, S would not know that P. Thus, the range of Qs that must be known not to obtain may be broadened so as to include not only propositions that are incompatible with P, but also others such that if they were the case, then S would not know P. (1995: n. 18)

DeRose is registering our attraction to the following principle.39 I’m assuming that in order to know :Q under these conditions, one must rule out Q. Qualiﬁcations would have to be added, as DeRose focuses on knowledge attributions in context (see his 1999). Furthermore, although the DeRose quote above seems to suggest that the condition in (G) is just ‘If Q were true, then S wouldn’t know P’, I suspect that he really wants knowledge of this condition, so I have gone with the latter. 38 39

Universal Scepticism

111

If: G. S knows that if Q were true, then S wouldn’t know P, and H. Q is appropriately contextually relevant, then: I. In order for S to know P, S must rule out Q. Unfortunately, if any of these philosophers are right, then live universal scepticism is true for mere mortals in eliminativism-live intellectual communities, as we can plug eliminativism in for Q, and the universal sceptic’s argument goes through. So if we want to reject that kind of scepticism, then we cannot accept any of those rulingout principles. However, we can’t just pronounce them mistaken and move on. Moore, Stroud, and DeRose are grappling with the reasonable intuition that TRO is too restrictive. In effect, they are arguing that since TRO restricts its focus to Qs inconsistent with P, it is missing out on an important and more encompassing ruling-out principle. Often in order to know P, we have to rule out some Qs that are logically, metaphysically, nomologically, and physically consistent with P. TRO may be true, but it’s only part of the full story regarding the ruling out of possibilities. Thus, we have a clash of intuitions. On the one hand, in order to know P, it seems that :Ps are central to the ruling-out task; that’s why philosophers have focused on closure principles similar to TRO. On the other hand, we have the intuition that we need a ruling-out principle that is more encompassing than TRO, yet unlike the principles of Moore, Stroud, DeRose, and the universal sceptic, can’t be used to derive universal scepticism. An ideal solution would preserve the central importance of the :Ps while supplying a ruling-out principle more expansive than TRO but not so strong as to pave the way to universal scepticism. That’s our Christmas wish list; can Santa deliver? Here is a solution. Moore, Stroud, and DeRose are right that often in order to know P, one needs to rule out some Qs logically, metaphysically, nomologically, and physically consistent with P. However, we need to rule out those Qs only because we need to rule out some :Ps, and the only way to do so, given the speciﬁcs of the knower’s situation, is to rule out those Qs. For instance, suppose you and I want to know when the Apollo 11 lunar module landed on the Moon. We are in my ofﬁce with our colleague Andy; you ask him

112

Universal Scepticism

what the answer is, and he correctly tells us ‘1969’. But I would have hesitated to ask Andy because, unlike you, I’m aware that I have excellent reason to think that Andy is a joker with questions such as these. It may seem that in order for me to know P, that the answer is ‘1969’, I need to rule out the possibility Q1 that Andy is a joker. That’s right, but a bit misleading. What I really need to rule out is the possibility Q2 that Andy was joking this time round and gave us an incorrect answer. Q2 is inconsistent with P; it’s one of the :Ps. It’s just that in order to rule out Q2 , the possibility inconsistent with P, I most likely need to rule out Q1 , the possibility consistent with P. So TRO is on the right track in focusing on :P possibilities alone. Here is the correct ruling-out principle, CRO: If Q is fully consistent with P, but S needs to rule out Q in order to know P, then there is some R inconsistent with P such that (a) S needs to rule out R to know P and (b) S needs to rule out Q to rule out R. The idea now is this: in order to know P (for virtually any P), we do have to rule out some :Ps (as TRO has it), and in order to rule out those :Ps, we will probably have to rule out some Qs consistent with P (as Moore, Stroud, DeRose, and the live universal sceptic claim), but among the Qs we won’t ﬁnd eliminativism. So we don’t have to rule out eliminativism in order to know that 2 þ 2 ¼ 4. Which you probably already knew anyway. That’s what I suspect is the best non-sceptical solution to the puzzle generated from (1)–(7). How good is it? If one likes cost–beneﬁt analyses that hold that rejecting common sense is a large cost, then one may well endorse this solution. But why? Why is it that we should accept CRO and reject the analogous principles advocated by DeRose, Moore, Stroud, and the universal sceptic? The only problem we have found with the latter principles is that they lead to universal scepticism. That may be enough to justify our rejecting them, but we still don’t know why they are mistaken. I ﬁnd that deeply unsatisfactory! It doesn’t help to reply with ‘Yes, but with CRO we can have everything we want: to know P, the :Ps are primary, but sometimes we need to rule out some Qs consistent with P’. Yes, that’s all ﬁne and good, but I’m none the wiser as to why URO, for instance, is mistaken. Perhaps it’s just a conceptual truth that knowing P means ruling out :Ps, and anything else—including the idea that in order to rule out those :Ps we often have to rule out some Qs consistent with

Universal Scepticism

113

P—is a matter of details. Maybe so. But if so, then that’s a conceptual truth that isn’t self-evident, at least not to me. Is the universal scepticism generated by these principles insane? Perhaps not: these arguments for universal scepticism share many of the interesting features of the live arguments regarding colour, pain, ﬂatness, and belief. In particular, none of the three central features of traditional universal scepticism—a sceptical predicament that lasts for ever, applies to everyone, and indicates a cognitive deﬁciency of its victims—applies to live universal scepticism. So almost all of the usual criticisms of universal scepticism simply don’t apply to the versions generated by the arguments given above. This universal scepticism might be palatable after all. So I don’t think we have been given any good reason to reject my new argument for universal scepticism.

8

The Consistency of Scepticism and Knowledge There is nothing so absurd but some philosopher has said it. Cicero, De Divinatione

Obviously, if the live sceptical arguments are sound, then they establish an interesting version of scepticism: for mere mortals with respect to the live hypotheses, they neither know P nor have justiﬁed or warranted beliefs in P. But the converse is not as clearly true. That is, even if our knowledge attributions have the truth-values that common sense says they have, so that, contrary to what I am arguing, millions of them are literally true despite being targeted by the live sceptical hypotheses, this does not mean that the live sceptic is defeated. Even if we mere mortals know that ﬁre-engines are red, the live sceptic wins. As I understand it, scepticism’s fundamental claim is that although millions of our beliefs just may be true, the warrant that these true beliefs have (or even that had by our false ones), if any, is much less, on balance, than any of us have ever supposed in our anti-sceptical moments. The sceptic says that the epistemic quality relevant to knowledge—whatever that amounts to—of our beliefs, even the best ones, is much less than we thought. We can think of scepticism and anti-scepticism as attitudes on a scale from 1 to 10. At the extreme sceptic end, at 1, there is the position that all our beliefs have no positive epistemic status or quality at all. The belief that Jack Kennedy has literally been in my left pocket for 300 years is just as good, epistemically speaking, as my belief that I have ten ﬁngers or that I have seventy-ﬁve ﬁngers, or that I have 34,343,343 ﬁngers. All these beliefs are at the same epistemic level; and that level stinks. At the extreme anti-sceptic end, at 10, the position is that many, many ordinary beliefs amount to infallible certainty. (We could go higher,

Consistency of Scepticism and Knowledge

115

but no one has ever produced even a bad argument for a higher conclusion.) If the reality turns out to hit point 7, the anti-sceptic wins; if it’s 3, then the sceptic wins. If it’s point 5, then everyone declares victory and goes home. This is vague, to be sure, but then so is the gist of scepticism. Various examples can be used to show that scepticism is consistent with knowledge. Pretend that, contrary to almost all recent speculation, all it takes to make ‘S knows P’ true is the truth of ‘S has a true belief P’. In other words, knowledge is just true belief, nothing more. Suppose, further, that as a matter of fact none of our true beliefs has any more than the tiniest bit of justiﬁcation. Lastly, assume that all it takes to make ‘S’s belief P is justiﬁed’ true is something like the truth of ‘S’s belief P has non-zero justiﬁcation’. Then, even though ‘S knows P’ and ‘S’s belief is justiﬁed’ are quite often true, and in the cases we would expect them to be true, surely scepticism is true too. After all, if that’s all our best beliefs amount to, true belief with the slimmest smidgen of justiﬁcation, then the sceptic wins even in the face of widespread knowledge and justiﬁed belief. In this scenario our knowledge and justiﬁed beliefs are so bankrupt, and the truth conditions for the corresponding claims are so pathetically austere, that the upshot is that scepticism rules the day, because the sceptic has been so much closer to the truth than the anti-sceptic. This simple point is easily overlooked, because it is easy to assume, contentiously in my judgement, that the truth conditions for knowledge and warrant claims are not austere compared to what we have always thought. Here’s an analogy intended to make the point ﬁrm. Pam and Pat are debating whether it’s possible to have good evidence for a miracle. They are going through the usual considerations. Then Pita (‘pain in the ass’) joins them and convinces them that the truth conditions for ‘X is a miracle’ are just this: X is something that has never happened before to anyone’s knowledge, X is mysterious to almost everyone, and people consider it to have come from God. Pita convinces Pam and Pat that this is what ‘miracle’ means. Surely the reaction to this announcement, assuming it’s accepted, is that it is largely irrelevant. Although Pam and Pat must now agree that many occurrences of ‘X is a miracle’ are true, none of this helps them much in their debate. For now the truth conditions for ‘X is a miracle’ are so uninteresting, so philosophically destitute, that scepticism about miracles is compatible with the truth of claims made with ordinary

116

Consistency of Scepticism and Knowledge

uses of sentences of the form ‘X is a miracle’. Pam and Pat will now say to Pita something like, ‘Well, that’s all very interesting, but we’re focusing on whether one can have good evidence that something has happened that ﬁts such-and-such conditions. That’s what we all have agreed we will mean by ‘‘miracle’’ in our discussion’. Pita has to admit that even though there have been many miracles, and many people have had excellent evidence for believing in those miracles— even evidence sufﬁcient for knowledge—there is still an interesting philosophical question surrounding the notion of being warranted in believing in a ‘miracle’—as understood in that discussion. In fact, the truth of claims made with ordinary uses of sentences of the form ‘X is a miracle’ and ‘S has excellent reason to believe and even knows that X is a miracle’ settle virtually nothing of philosophical importance. The epistemological sceptic is saying something similar: even if the claims made with ordinary occurrences of ‘S knows P’ and ‘S’s belief that P is justiﬁed’ are true, in line with common sense, we still need to determine whether those facts have any or much relevance to the philosophical question of scepticism. This is not to say that those facts about sentences are irrelevant (as in the miracle case), but only that we need to check whether they are relevant; we had better not assume that they are relevant. This generates a problem for anti-sceptics. They usually take their task to be this: show that most of the knowledge claims we take to be true and unproblematic are really true. If we can do that, so the antisceptic typically thinks, then it’s time to break out the champagne. (This is particularly true of contextualists and relevant alternative theorists.) But this isn’t quite right. The anti-sceptic has to show that our knowledge claims are true and that they are made true by some fairly robust epistemically relevant facts. Once again, if they’re made true by impoverished facts—e.g., mere true belief—then the sceptic wins if those beliefs have only piddling amounts of warrant. Let’s say that the persistent sceptic is the sceptic who admits that ‘S knows P’ is true pretty much when common sense says it’s true, but insists that the knowledge in question is impoverished compared to what any of us expected in our anti-sceptical moods. It should be clear that the persistent sceptic’s position is not this: although ‘S knows P’ is true, S’s knowledge is not ‘high standards’ knowledge. This sceptic isn’t complaining that our knowledge doesn’t satisfy some super-duper high-octane condition that only a philosopher

Consistency of Scepticism and Knowledge

117

could love. Neither is she saying that knowledge is unimportant or not an epistemologically central notion. This point is most easily made with examples, such as the one according to which ‘S knows P’ is true if and only if ‘P and S believes P’ is true, ‘S has a justiﬁed belief P’ is true just in case ‘S believes P as do most other people’ is true, and our true beliefs fail to come close to meeting any of the substantive epistemic conditions that philosophers have ﬁled under ‘warrant’ and ‘justiﬁcation’. The persistent sceptic who objects that if this scenario was real, then knowledge and justiﬁed belief would exist but be pathetic, isn’t whining that knowledge and justiﬁcation aren’t what she always fantasized them to be. Here is another way to put it. Pretend once again that warrant comes in units. All anti-sceptics until now have thought that our beliefs require at least 1,000 warrant units to amount to knowledge. Then someone comes along with some surprising semantic analysis that proves that the truth conditions for ‘S knows P’ are just ‘P, S believes P, and S has 150 warrant units for her belief that P’. Suppose, further, that we almost never have more than 350 units. Is this a victory or a defeat for the persistent sceptic? I say it’s a victory, a victory without whining. If instead of 150 and 350, it had been 750 and 950, then the victory goes to the anti-sceptic, and the sceptic can only whine. And then of course there is the middle ground, where there are no winners or losers but lots of whiners.40 So the anti-sceptic has to show that the truth conditions for ‘S knows P’ are not impoverished. Here are two reasonable replies to this request for additional argument. First, we have no good reason (yet, anyway) to think that the warrant standards governing ordinary knowledge claims are impoverished to any signiﬁcant degree. Sure, if it turned out that knowledge were mere true belief, and justiﬁed belief were merely common belief, then the anti-sceptic would have a lot more work to do, as the 40 DeRose (1995; 5) seems to recognize this problem, declaring as ‘timid’ the sceptic who admits the truth of commonsensical knowledge claims. He says, I think entirely appropriately, that the problem is ‘a topic for another paper’. However, I disagree with the use of ‘timid’. Further, I don’t think anything I’ve written above conﬂicts with contextualist or relevant alternative theories, which are typically used to show nothing but the truth of ordinary knowledge attributions. These theorists quite reasonably assume that the truth conditions for ‘S knows P’ go signiﬁcantly beyond ‘P and S believes P’, for instance, and indicate something epistemically worthwhile; otherwise they wouldn’t be fussing over truth-values in the ﬁrst place. But there is a real temptation to pauperize the warrant condition in order to save the truth of the knowledge attribution.

118

Consistency of Scepticism and Knowledge

truth conditions for knowledge would be so weak. In that case scepticism might be true even in the face of widespread knowledge; the persistent sceptic could be right. But the antecedent is obviously false, so who cares? We have no reason to think that the truth conditions for ‘S knows P’ are impoverished. Second, if the truth conditions of our knowledge claims are generally met when we think they are, it then looks as if the persistent sceptic—the one who insists that this epistemic achievement is still not up to snuff—is putting too much emphasis on a folk theory of knowledge, as follows. This sceptic has some highly theoretical notion of knowledge; a set of ‘philosophical’ conditions for knowledge claims generated from reﬂections on knowledge. Let’s assume that even though occurrences of ‘S knows P’ often satisfy their truth conditions, they fail to satisfy the philosophical conditions. But this just shows that we can very easily be led into a mistaken theory of knowledge. Our conclusion shouldn’t be scepticism, but that our folk theory of knowledge, when developed in a sophisticated yet natural way, is just plain false. This isn’t to say that such a conclusion is uninteresting; the folk theory might be tremendously intuitive and compelling. And we don’t blame the philosopher who thinks that knowledge claims have to meet the higher standard; such a thought might be quite compelling. But it is to say that the result is importantly different from scepticism. I think neither of these responses is good. I shall defuse the ﬁrst argument in Chapters 10–12, by showing how certain anti-sceptical solutions to the live sceptical puzzles threaten to make the knowledge saved impoverished. But I will rebut the second argument here. That argument tries to drive a wedge between scepticism and a signiﬁcantly mistaken folk theory of knowledge. The former says, roughly, that there’s no knowledge; the latter says, again roughly, that there’s plenty of knowledge, but it’s of a much lower quality than what the folk thought. I fail to see much interesting difference between the two positions. With the second option we end up with realism about knowledge at the price of having knowledge be destitute. What scepticism and mistaken folk theory have in common is the interesting possibility that our best true beliefs are epistemically quite inferior to what virtually anyone thought they were. Whether those beliefs amount to knowledge, strictly construed, is a side issue. Naturally, one can’t get very far on these matters unless one determines just how ‘impoverished’ knowledge is if such-and-such

Consistency of Scepticism and Knowledge

119

an epistemological theory is true. An epistemic theory should be responsible not just for positing truth conditions but also for making at least a rough claim about how those truth conditions relate to various theories of knowledge. If a theory makes the truth conditions austere and easily met, then it can claim to rescue knowledge, but it also makes knowledge so meagre in nature that the sceptic is closer to the truth than the anti-sceptic. If a realist about knowledge posits the truth conditions as opulent and met only with difﬁculty, then she can claim to make knowledge strong enough to defeat the sceptic. However, the realist then has the trying task of showing that those conditions are satisﬁed. The middle ground is harder: the posited truth conditions are middle strength, so we don’t know who wins if they are satisﬁed. I suppose it may be just a matter of philosophical taste; a matter of where you think the big issues lie. I’m not going to quibble over the use of the word ‘sceptic’. There’s no point devoting hours to investigating this monster of a problem unless we have excellent reason to think there really is a problematic middle ground. Before you spend hours worrying about the possible monster under your bed, you might as well just look to see if there’s one there to worry about. We will encounter such monsters in Chapters 10–12. The same problem holds for many areas of philosophy. For a relevant instance, if eliminativists lose the battle over the existence of belief, they could still win the battle over belief ’s signiﬁcance. One popular way to argue against eliminativism is to argue that the truth conditions for ‘S believes that P’ are ‘austere’, which means that it takes very little to satisfy them. In particular, the eliminativist argument that goes ‘If ‘‘S believes that P’’ is true, then such-and-such causal-intentional-modular-what-not conditions would have to hold, but they probably don’t’, is thereby disarmed if the austere theory is right, as the initial conditional is false. But if it is so easy to believe that P, then it’s a wonder that this belief could be such a powerful causal force in our lives. On the face of it, you can’t have it both ways: if it’s a piece of cake to believe that P, then it’s hard to see how belief facts could do everything that folk psychologists say they do; surely something else is behind socially relevant human cognition and action. Further, the easier it is to have beliefs, the harder it is to resist the idea that virtually anything can be a believer. One can think of other disputes that have the same problem. For instance, if you make free will compatible with determinism, then you might drain it of the power needed to make more than 15 per cent of our

120

Consistency of Scepticism and Knowledge

folk-psychological theory of free will true. Even more pressing is the case of moral facts: sure, one can dilute the notion of a moral fact so much that one can ﬁnd good reasons to think that there are such things, but if they have just 15 per cent of what anyone would have expected of such a fact, then we have hardly defeated the person who says that there are no moral facts. In what follows I have tried to avoid this thorny issue regarding the consistency of live scepticism and the knowledge it targets as often as possible. I have done so by focusing on the sufﬁcient condition for scepticism: if we can show that all the elements of an interesting set of our knowledge claims are false (even though many of the true-belief claims are true; so S believes the truth that P but doesn’t know that P), as the live sceptical arguments attempt, then an interesting if restricted scepticism follows.

9

Epistemic Threats, Context, and our Anti-Sceptical Strategies The fundamentalists, by ‘knowing’ the answers before they start, and then forcing nature into the straitjacket of their discredited preconceptions, lie outside the domain of science—or of any honest intellectual inquiry. Stephen Jay Gould, Bully for Brontosaurus

The live sceptic’s argument boils down to the following. Because of the liveness of the new hypotheses and your status as a well-informed mere mortal (conditions (i)–(v) of liveness and conditions (a) and (c) of mortality), those hypotheses pose signiﬁcant epistemic threats to many of your beliefs: threats that must be defeated if your beliefs are to count as knowledge. But your status as a mere mortal (specifically, conditions (b) and (d) of mere mortality) shows that you can’t defuse or defeat or neutralize those threats. Thus, your beliefs don’t amount to knowledge. Here we jettison the ‘ruling-out’ language in favour of the more neutral ‘defeat’, ‘defuse’, and ‘neutralize’. The point of these locutions, to be further elaborated below, is to indicate that the putative knower must have some epistemic item or feature—appropriate evidence, reliability, or what not—that cancels out the threat posed by the live hypothesis. In this chapter I will do several things. First, I will examine the notion of the epistemic threat to a belief that P posed by contrary hypotheses, hypotheses entailing :P. In doing this we will see how contextual factors affect the live sceptic’s argument. I will then describe the three strategies available for constructing antisceptical responses to the live sceptic’s argument. The goal here is twofold: to provide a way of classifying possible anti-sceptical solutions and, more hopefully, to supply materials useful for constructing an adequate anti-sceptical solution to live scepticism. Finally, I will

122

Anti-Sceptical Strategies

compare the power of the live and classic hypothesis-based sceptical arguments. I will be arguing for the fairly obvious thesis that the live ones are signiﬁcantly stronger. A Epistemic Threats, Context, and Live Scepticism Consider one of the standard barn cases discussed throughout the literature. I’m driving through the Iowan countryside and see a barn under unusual circumstances: although what I saw was a normal barn, for a hundred miles in every direction there are dozens of fake barns but no real ones. I say, ‘My, that’s a lovely red barn’; so I have the true belief that that’s a barn. Tom is in the back seat with his wife, and he says to her ‘Ah ha! See? He knows that that’s a barn’ (perhaps he said this because I’m from a foreign country that doesn’t have American barns and have just learned what they are). None of us know anything about the fake barns. It is usually thought that in order for Tom’s assertion to his wife, ‘He knows that that’s a barn’, to be true, I have to be able to rule out the possibility that it’s not a barn but a fake. I have to be able to do this because of the existence of fake barns all over the place. The fake-barn possibility—the possibility that the thing I saw was not a barn but a fake—poses a signiﬁcant epistemic threat to the truth of Tom’s assertion (and the knowledge status of my belief).41 Now consider a world nearly identical to the last one except that there are no fake barns anywhere (and no plans to make any, etc.). In this situation it is intuitive that the fake-barn possibility poses less of a threat. Even if the traditional sceptic is right that Tom’s assertion is false in both worlds, the fake-barn possibility constitutes a higher hurdle in the ﬁrst scenario than in the second.42 It’s more ‘relevant’ in 41 This is a Gettier case: I have a justiﬁed true belief that that’s a lovely barn, but I don’t know it, because the fake-barn possibility is not neutralized and needs to be neutralized for my belief to amount to knowledge. At least, my belief has a signiﬁcant amount of justiﬁcation. The live sceptic cannot be thought of as capitalizing on Gettier cases. According to live scepticism, we have true beliefs with loads of positive justiﬁcation that don’t amount to knowledge, due to some negative justiﬁcation (or ‘defeaters’). The negative justiﬁcation (or warrant) is enough to falsify both ‘S knows P’ and ‘S justiﬁably believes P’ and ‘S warrantedly believes P’. Also see my comments on the Harman cases in Chs. 4 and 6. 42 Some sceptics won’t allow epistemic threats of varying strength, but I don’t know of any good arguments for those positions.

Anti-Sceptical Strategies

123

the scenario in which there are loads of fake barns. The only difference between these two situations is in certain features of the alleged knower’s context (the presence/absence of lots of fake barns); call these contextual features of the subject CS factors. Because all that differs in the two situations are CS factors, and in one of the situations the fake-barn possibility is signiﬁcantly more threatening, CS factors can raise the threat posed by a contrary possibility. The upshot: having appropriate CS factors can boost the epistemic threat supplied by a contrary possibility. According to some philosophers, features regarding the knowledge-attributor or her audience have the very same effect. Here is an illustration of how such features are supposed to do this (adapted from DeRose 1992: 913; see Cohen 1999 and DeRose 2005 for further intriguing examples). Case X. It’s Friday and Jon needs to deposit a cheque; so do Jill and I. The three of us are standing outside the bank looking in. The line inside is very long, so Jon mutters to himself, ‘Heck, I’ll do it tomorrow morning when it opens.’ Jill and I are observing this, and I say to her, ‘He knows that the bank will be open tomorrow. So we can wait until tomorrow to deposit the cheque.’ Jill says, ‘Right’, and we go away conﬁdent that we’ve made the right decision. As a matter of fact, for many years all banks have been open on Saturdays. And we have no particular reason to hurry with our cheque deposit. Case Y. Almost everything is as in case X. But now Jill responds to me with, ‘I don’t think so. What if the bank has changed its hours? That guy didn’t check the hours listed on the door. Besides, we’ve got to deposit ours before Monday morning or we’ll be in big trouble when that cheque bounces!’ As a matter of fact, for many years all banks have been open on Saturdays. Jill is right that we absolutely must deposit the cheque before Monday. The knowledge claim K in question is expressed by my utterance (in both cases) of ‘He knows that the bank will be open tomorrow’. The possibility P in question is ‘The bank is closed on Saturdays’. In case X, the intuition is that P doesn’t really pose much of a threat. However, in case Y, P is more formidable due to factors regarding the attributor, A, and A’s audience: the respectful consideration of P on the part of both Jill and myself, coupled with our increased need

124

Anti-Sceptical Strategies

to make our deposit before Monday. Even if the traditional sceptic is right that K is false in both cases, X and Y, it seems to some philosophers that P poses more of a threat in case Y than in X. The bank is the same in X and Y, as is Jon and every other bank. Because all that differs in the two situations are factors having to do with the knowledge attributor A and A’s audience (the people she is talking to), and in one of the situations P is signiﬁcantly more threatening, it follows that attributor and audience factors—CA factors—can raise the threat posed by a contrary possibility. The upshot: having the right CA factors can boost the epistemic threat had by a contrary possibility. At least, we will buy this idea if we have the right kind of intuitions regarding the bank example (more on this below).43 Thus, different kinds of factors can boost the relevance of a possibility. You believe that P, and if the right factors are present, then some of the :P possibilities pose formidable threats to your belief in the sense that you need some powerful epistemic characteristics to defuse those threats. The tools available for defusing or ‘neutralizing’ hypotheses are things like my evidence, reasons, justiﬁcation, proper functioning, and reliability. Call those tools, each of which has to do not so much with the subject’s context as with the subject herself, Subject factors.44 The view we have been inching towards is that there might be two scenarios in which I believe that P, but due to differences in CS or CA factors, a certain :P possibility poses more of a threat in one scenario than in the other. In scenario A but not in B the :P possibility in question is boosted by many CS and CA factors. In order for me to know P in scenario B, my Subject factors can be modest; in scenario A, my Subject factors need to be stronger. More generally, certain contextual factors can raise a possibility to ‘highlighted’ degree N, which amounts to an epistemic threat of a certain strength proportional to N. The stronger and more numerous 43 One could attribute the threat difference to CS factors, as the facts about the attributor and his audience are features of the subject’s context. However, this makes CS factors cover too much, as we could alter the example so that Jon is very removed from Jill and I in space, time, and purpose. 44 Evidence can of course be knowledge. Therefore, when a Subject factor is my knowledge that P, this Subject factor says something about me (I know and believe P) and about my environment (P is true). We could classify evidence-as-knowledge as a combined CS –Subject factor; it doesn’t matter.

Anti-Sceptical Strategies

125

the highlighting factors are, the greater the strength of the epistemic threat posed by the possibility so raised. The higher the degree of threat, the stronger the Subject factors have to be to neutralize the highlighted possibility. I trust this is plausible; I’m going to assume it is on the right track from now on. Return to the original dinosaur story from Chapter 1. Subject S is an undergraduate studying palaeontology at a well-respected university. She is aware of the live status of the solar ﬂare and supervolcano theories. She believes the true meteor one. Now one of her classmates, A (for ‘attributor’), asserts that S knows that a meteor wiped out the dinosaurs. She says this to some other palaeontology student (A’s audience). If what we have said previously about the relevance-boosting power of CS and CA factors is correct, then there are twenty-one factors that boost the relevance of the solar ﬂare and supervolcano theories. (The numbers come out differently if CA factors don’t have threat-boosting power; more on that below.) In thinking about the twenty-one factors, consider H to be either the solar ﬂare or the supervolcano hypothesis. 1. In S’s intellectual community H has been through a signiﬁcant (not to say exhaustive) evaluation by experts over many years. 2. This is true for A (as A is in S’s community). 3. It’s also true for A’s audience (as she is in S’s community as well). 4. In S’s intellectual community H is judged actually true or about as likely as any relevant possibility by a signiﬁcant number of well-informed, well-respected, highly intelligent, non-crackpot experts in the ﬁeld(s) to which H belongs. Also assume that there are lots of these experts. 5. This is true for A as well. 6. This is true for A’s audience too. 7. Those experts in S’s community reached that favourable opinion based on H’s merits in a familiar, epistemically responsible way (i.e., the way they reached their opinion seems as epistemically responsible as any). 8. Same for A. 9. Same for A’s audience. 10. Those experts in S’s community consider there to be several decent and independent arguments that support H. So it’s not the case that the only reason people pay serious attention to H is the presence of one weird experiment or line of reasoning.

126

Anti-Sceptical Strategies

11. Same for A. 12. Same for A’s audience. 13. Many of those experts in S’s community consider H to be a ‘real, live possibility’ (i.e., that’s what they’d say if you asked them). 14. Same for A. 15. Same for A’s audience. 16. S is at least somewhat familiar with H and the issues surrounding H, including the fact that H is live in the above sense. That is, S is about as aware as anyone of the fact that H is live. 17. Same for A (as she is a palaeontology student as well). 18. Same for A’s audience (same here). 19. S is about as aware as anyone of the fact that the hypothesis H is inconsistent with the proposition P alleged to be known. 20. Same for A. 21. Same for A’s audience. These twenty-one conditions come from the liveness and mortality characterizations. When H is the solar ﬂare hypothesis, (1)–(21) are true. So there are twenty-one factors boosting the epistemic threat posed by the solar ﬂare hypothesis to A’s utterance of ‘S knows that the meteor story is true’. Now consider the BIV hypothesis: (1)–(18) are false and (19–21) are true. That is, there are just three factors boosting the epistemic threat posed by the BIV hypothesis to A’s utterance. This just goes to prove what we knew all along: the BIV possibility is simply not as epistemically threatening as the solar ﬂare possibility. The ﬂare hypothesis has twenty-one offensive weapons trying to defeat the truth of A’s utterance, whereas the BIV hypothesis has just three of those twenty-one. Furthermore, there are excellent reasons for thinking that S’s Subject factors are not up to the job of neutralizing the threat made by the solar ﬂare hypothesis: 22. S’s intelligence, understanding, and knowledge with regard to H and the issues surrounding H are not extraordinary for people in her intellectual community. 23. Same for A. 24. Same for A’s audience. 25. If S has any reasons or evidence she can marshal for casting doubt on the hypothesis, and if they were carefully considered by the members of that community—in particular, by the wellinformed, well-respected, and highly intelligent experts who

Anti-Sceptical Strategies

127

were thoroughly familiar with the hypothesis—they would be nearly universally and conﬁdently rejected as clearly insufﬁcient to rule out the hypothesis. 26. Same for A. 27. Same for A’s audience. All six are true for the solar ﬂare hypothesis, whereas only three (22– 24) are true for the BIV hypothesis. We can think of these as more weapons against S, diminishing the chances that A’s utterance of ‘S knows P’ is true. Thus, the solar ﬂare story generates a total of twentyseven weapons with which to ﬁght the truth of ‘S knows P’, whereas the BIV story generates a mere six of those twenty-seven weapons. At no point in this line of argument have we assumed that the traditional sceptic is wrong. But it doesn’t take a genius to spot the beginnings of a powerful anti-sceptical lesson here. The nature of our epistemic practices shows that the threat posed by the BIV and other traditional sceptical hypotheses is tiny compared to that posed by the live hypotheses. So the mere fact that the traditional sceptic uses hypotheses inconsistent with our beliefs does little to show that they pose epistemic threats of any signiﬁcance. The dinosaur case does not apply to people outside the community in which the sceptical hypotheses are live, people who aren’t mere mortals within the community, or times during which sceptical hypotheses aren’t live. This is deﬁnitely not to say that other contexts are without any interest, or that the live scepticisms don’t threaten them. It is only to make our task manageable while conserving philosophical import. The numbers come out differently if you don’t accept the ‘attributor contextualist’ idea that CA factors affect truth conditions. And plenty of people don’t like that idea. It seems odd that factors having to do with the attributor A and her audience could be epistemically relevant to whether A’s use of ‘S knows P’ is true. Surely, one might think, whether or not S knows P is determined by P and S, not by A or A’s audience! By making CA factors have ‘threat-boosting’ powers, attributor contextualism has the odd consequence that ‘S knows P’ can have different truth-values on different occasions of use, even though in each use it’s the same S and P. This will happen provided the CA factors change appropriately from context to context. That’s what happened with my ‘He knows that the bank will be open tomorrow’ in cases X and Y. (The two occurrences of the same

128

Anti-Sceptical Strategies

sentence type get different truth-values because the occurrences of ‘knows’ have different ‘semantic values’—values that are truth-conditionally relevant.) Attributor contextualists have good reasons for accepting these initially counter-intuitive consequences, but it’s clear that the jury is still out on this. If we subtract the CA factors from my lists, then the ﬁnal tally is nine factors for the solar ﬂare hypothesis versus two for the BIV hypothesis (as we lose all the facts about the attributor and her audience). Thus, the ratio stays the same (4.5 to 1). Somewhat arbitrarily, in what follows I will stick with the larger numbers, twenty-seven and six; feel free to substitute the smaller numbers if you wish. B Our Three Anti-Sceptical Strategies In Chapter 6 we looked at one proposed solution to the live sceptical puzzles, the sceptical one. What plausible strategies are there for mounting a successful anti-sceptical solution to these puzzles? Suppose you are faced with a sword-wielding opponent. You have your own sword with which to protect yourself. Suppose, further, that she does not defeat you. How might you have won the battle? You might have won because her sword was nothing but a pathetic toothpick. So even if your sword was no great shakes, it didn’t have to be. Another way you might have won is by doing something very tricky, thereby taking away her sword. So even if her sword was terribly threatening, you never had to battle it, because you managed to make it unusable to her. Whether or not your own sword is just a toothpick, you will win. Of course a third way to win would be to use your impressive sword to destroy her sword. So even if her sword was very threatening and she got to use it on you, you were able to battle it back with your sword in order to win the day. Your sword-wielding opponent is the live sceptic; her sword is generated from the ﬁrst twenty-one factors noted in the previous section; you are the target of the live sceptic’s argument whose sword is fashioned from your Subject factors. Your sword looks weak, as factors (22)–(27) of the previous section count against its strength. Her sword looks strong, because it has factors (1)–(21) going for it. Because of those facts, it looks as though the live sceptic will win; the fans in the stands will be putting their money on her, for sure. You have three strategies to use in defeating the live sceptic. You can argue

Anti-Sceptical Strategies

129

that the sceptic’s sword, the twenty-one factors, is actually a meagre toothpick. Contrary to what I’ve been arguing, the twenty-one factors are pretty pathetic, not really mounting a signiﬁcant epistemic threat to your beliefs. So even if your sword is no better than a toothpick, perhaps because of factors (22)–(27), it doesn’t matter. This is the No-Threat Strategy, which claims that the twenty-one factors are so feeble that they fail to mount a signiﬁcant threat that can’t be neutralized by even very weak Subject factors (so we can be knowers even if factors (22)–(27) count severely against the power of our Subject factors in neutralizing the sceptical hypotheses). For instance, one might use the No-Threat Strategy in arguing that the kind of liveness enjoyed by the new sceptical hypotheses is crucially different from, and signiﬁcantly weaker than, that enjoyed by ordinary scientiﬁc hypotheses, so the former pose no real threat, just a mere toothpick. Alternatively, the anti-sceptic can pull a fast one: be prepared to admit the strength of the twenty-one factors constituting the live sceptic’s impressive sword, but insist that at least in many interesting knowledge-attribution contexts the live sceptic’s sword cannot be used in battle, as it has been magically set aside. This is the Disarmed-Threat Strategy. For instance, perhaps in many contexts we implicitly just set aside colour error theory, and that contextual setting-aside factor cancels out the threat posed by the liveness and mortality facts. Or perhaps the practical purposes for the sake of which we make certain knowledge claims render ‘purely theoretical’ issues, such as colour error theory, moot. Either way, the opponent’s sword, whether or not it’s impressive, poses no threat because it has been side-stepped. So you don’t even have to pick up your own sword. Finally, the anti-sceptic can argue that, regardless of the strength and usefulness of the live sceptic’s sword (the twenty-one factors), our Subject factors are sufﬁciently strong to defeat it head-on, so to speak. This is the Defeated-Threat Strategy. For instance, perhaps facts about the reliability of our belief-forming process types constitute an impressive defensive sword sufﬁcient to beat back the live sceptic. Or maybe facts about how ‘safe’ or ‘sensitive’ (in senses to be described below) our belief contents or states are can save the day. Actually, there are many battles to ﬁght. Each time someone says ‘S knows P’, where the live sceptic challenges the attribution,

130

Anti-Sceptical Strategies

the anti-sceptic needs to employ some strategy or admit defeat. The live sceptic’s argument, simply put, is that for a vast, diverse, and interesting set of knowledge-attribution contexts, none of the anti-sceptical strategies shows that the live sceptical hypotheses are appropriately neutralized; so the attributions are false. Think of the anti-sceptic as having three methods to use in defeating the sceptic, and using different ones depending on the attribution in question. What won’t do, I think, is to reject all the strategies and deny the Mere Mortal Premiss that many of us are mere mortals. That’s just a dead end in my opinion, going against brute facts. In order to make that anti-sceptical approach work, one would have to argue that virtually none of us, even in close possible worlds, is a mere mortal. Perhaps in some worlds every one of us can rule out eliminativism, for instance, merely by saying to ourselves, ‘It sure seems to me that I believe that 2 þ 2 ¼ 4’. I’ll be considering such ideas later in the book. But even in those worlds philosophers sufﬁciently sceptical about the truth of common sense will not consider such a line of reasoning as sufﬁcient to rule out eliminativism. So condition (d) in the characterization of ‘mere mortal’ will still hold. In the remainder of this essay I will assume that denying the Mere Mortal Premiss won’t work and that any plausible anti-sceptical solution will have to utilize some combination of the three options sketched above and developed below. Let’s see how the three strategies match up with the live sceptic’s ofﬁcial argument, which you’ll recall has the following uncontextualized premisses. The Modesty Principle If S is mortal, then live H isn’t ruled out. The Live Hypothesis Principle If S is mortal and live H isn’t ruled out, then S doesn’t know P. The Mere Mortal Premiss We are mere mortals regarding H (in this world or a very close one). We are assuming the truth of the last premiss. So the Live Hypothesis Principle amounts to this: if live H isn’t ruled out, then we don’t know P. The No-Threat and Disarmed-Threat strategies reject this principle. The Disarmed-Threat Strategy says that since H has been somehow disarmed, we still know P even if we can’t rule out H. The

Anti-Sceptical Strategies Are factors (1)−(21) armed?

131

Is the Live Hypothesis Principle true?

No

Yes

DisarmedThreat Strategy

Are the live hypotheses a significant threat?

No Yes

No

No-Threat Strategy

Yes

Are our subject factors weak?

Is the Modesty Principle true?

No-Threat Strategy

DisarmedThreat Strategy

No Defeated-Threat Strategy

No

live scepticism

No

Yes

Yes

Yes

Are factors (1)−(21) armed?

Defeated-Threat Strategy

live scepticism

No-Threat Strategy says that H never posed any threat at all, so even if we can’t rule it out and even if we don’t disarm it in some way, we still know P. The Modesty Principle amounts to this: live H isn’t ruled out. The Defeated-Threat Strategy targets that principle, claiming that we all have impressive Subject factors that rule out the formidable H. Finally, here are two decision trees that can be used to classify responses to the live sceptic’s paradox. The anti-sceptic may, if she is a certain kind of attributor contextualist, admit that perhaps in some rareﬁed contexts the live sceptical hypotheses are imposing and unmet threats; so in those contexts ‘S knows P’ is false. This kind of concession to the sceptic is not uncommon now. Such ‘sceptic-friendly’ contextualists often allow that in weird enough knowledge-attributor contexts—say when we are discussing BIV scenarios in a certain way in a philosophy class— when the sceptic says ‘We know next to nothing’ or ‘You don’t know

132

Anti-Sceptical Strategies

that you have hands’, she speaks the truth. But, they insist, in ordinary attributor contexts it would be true to say, e.g., ‘We know millions of things about the human eye’. The problem with the weird attributor contexts is that the :P possibilities that get boosted into epistemic relevance go beyond the resources of our Subject factors, and nothing else is available for neutralizing duty. But in all the other contexts, which is almost all contexts, those hypotheses are irrelevant in the sense that they pose no threat (No Threat), because they have been set aside or disarmed (Disarmed Threat), or because their threat is adequately defeated by some impressive Subject factors (Defeated Threat). In (virtually) all contexts we have the following anti-sceptical solution: the live sceptical hypotheses aren’t highlighted so as to produce an epistemic threat that cannot be defused in one way or another. C Scepticism: Live versus Classic On the face of it, it seems that none of the three strategies will work against live hypotheses. In the case of the BIV hypothesis, on the other hand, it’s reasonable to pursue the No-Threat Strategy: since there are just the measly six boosting factors, the BIV hypothesis isn’t much of a real threat. It is a threat, but a threat so incredibly meagre that it can be adequately defused via reliability and other ubiquitous Subject factors. If you’re breathing and sane, then you have ‘done’ enough to defuse the threat generated by the classic sceptical hypotheses. The traditional sceptic gets overly excited by the fact that sceptical hypotheses are logically inconsistent with the propositions we ordinarily take ourselves to know, but this logical fact ‘highlights’ or brings into prominence those hypotheses only a little bit—a bit that our automatically generated Subject factors are strong enough to defuse (or so the anti-sceptic will plausibly argue). There are a mere six factors constituting the threat posed by the classic sceptical hypotheses; so it makes sense that we can defeat that threat with some epistemically ubiquitous and primitive Subject factors. But the live sceptical hypotheses have all twenty-one threat-boosting factors; so the No-Threat Strategy doesn’t look anywhere near as attractive for live scepticism as it is for classic scepticism. Or perhaps we can go with the Defeated-Threat Strategy for the BIV hypothesis: admit the presence of an impressive BIV threat

Anti-Sceptical Strategies

133

but claim that we have even more impressive and ubiquitous Subject factors. However, this doesn’t look promising for the live hypotheses, as the twenty-one factors make the threat signiﬁcant, and the six anti-neutralization factors make our Subject factors look woefully inadequate to defeat the threat head-on. So Defeated Threat won’t work. (It doesn’t seem very plausible for the classic hypotheses either, since they don’t seem to amount to a serious epistemic threat.) The Disarmed-Threat Strategy might work for many contexts for the BIV hypothesis, because that hypothesis is so absurd that we are almost universally disposed to dismiss it as completely irrelevant— i.e., implicitly setting it aside as truth-conditionally irrelevant to uses of ‘S knows P’. But once again, as reasonable as this strategy might be for traditional hypotheses, because the live hypotheses are live and we are mere mortals aware of their live, respected status, it can hardly be said that we are disposed to completely ignore them!45 We can take a brief look at some particular knowledge-attribution contexts just to see the initial difﬁculties in using the anti-sceptical strategies against the live sceptic. Since the reader is undoubtedly a philosopher, imagine that you have just a BA or MA in philosophy and you are speaking with a friend who has the same in cognitive science. You’re both mortals. The two of you are discussing theories of belief. Perhaps something mostly philosophical, say Fodor’s Language of Thought thesis, or perhaps something more scientiﬁc, say connectionism. You might say something theoretically loaded such as, ‘Yes, but we know that people often have contradictory beliefs’, or ‘We know that people have beliefs in non-existent objects’, both utterances having the aim of making some theoretically relevant point. Assume that people often do have contradictory beliefs and people do have beliefs in non-existent objects. Assume, further, that these facts are ‘common knowledge’ in the sense that you and just about everyone else in the ﬁeld have ordinary, but considered excellent, reasons for believing them. So you have two true second-order beliefs, the contradiction one and the non-existence one, with the ordinary amounts of warrant coming from ordinary Subject factors. But, given the liveness of the eliminativist sceptical hypothesis and the mortality of you and your friend, are your two knowledge 45 ‘Yes, but we ignore them in lots of everyday contexts.’ Right, and the consequences of that fact will be investigated in sect. B of Ch. 12, on the Disarmed-Threat Solutions.

134

Anti-Sceptical Strategies

attributions true only if the eliminativist hypothesis is neutralized by some strong Subject factors? It certainly seems so. Consider again the analogous water-as-element contexts. (Recall that in this ﬁctional scenario we’re in a time when the scientiﬁc community is still debating whether water is an atomic element.) You and another mere mortal are discussing how the boiling point of water is affected by the force due to gravity, and in the conversation you say, in an appropriate way, ‘Yes, but we know that water is made of oxygen as well as hydrogen. So the effect must be larger than you said it would be.’ Assume, further, that neither of you have set the live water-as-element hypothesis aside with something like ‘Setting aside the whole element-or-compound debate for the moment’. This is just like the case concerning eliminativism. It’s clear that your knowledge assertion ‘We know that water is made of oxygen as well as hydrogen’ is false. If someone had said, ‘What about Smith and Jones? What about the element idea?,’ you would have two options. First, you could retract or qualify your assertion: ‘Oh, that’s right. Well, assuming that it’s a compound, then the gravity effect has to be larger.’ Of course you might not retract: ‘Smith and Jones?! Come on; you don’t take that seriously’; or ‘No, I don’t think so. Their theory just doesn’t stand up.’ In those cases we have the same result: your boiling-point knowledge assertion is false; you didn’t know that water is made of oxygen and hydrogen. Surely the same holds for the contradiction and non-existence assertions about belief. The important thing to remember is that the sceptical hypothesis doesn’t have to be ﬂashing on a screen in front of the conversationalists in order to be highlighted, something we need strong Subject factors to neutralize. When attributor contextualists treat the conversational relevance of the traditional sceptical hypotheses, such as the BIV one, what they say is reasonable: unless the hypothesis is ﬂashing on the screen (so to speak), we need not have any signiﬁcant Subject factor that puts it away. But that’s reasonable because it is hard to see how any other mechanism could make the classic sceptical hypotheses threats, things that only reasonably strong Subject factors can disable. With the new sceptical hypotheses it is clear that other mechanisms can do the trick: the hypothesis is live, all the conversationalists are aware of it, the subject is aware of it, all of them are aware of its liveness, it is obviously inconsistent with what is being asserted in the conversation, etc. That’s plenty to make it highlighted!

Anti-Sceptical Strategies

135

What appears above is a good warm-up, showing that some pretty fancy footwork is needed to defeat the live sceptic’s arguments. All I’ve attempted to show here is that we can’t use any of the three strategies in any easy way to defeat the live sceptic—as we can (perhaps) for the traditional sceptic. The rest of the work appears below, where I try to ﬁnd clever ways of using the three strategies to defeat the live sceptic. In the next three chapters I articulate and evaluate thirteen anti-sceptical solutions to the live sceptical puzzles. Two use the Disarmed-Threat Strategy; seven use the No-Threat Strategy; three employ the Defeated-Threat Strategy; and one other (the Moorean one) doesn’t ﬁt neatly into any category. I wish I could say which proposal I think is best, but I can’t. However, my ﬁrst main point is that the sceptical solution given in Chapter 6 is good enough to take seriously, as a real contender along with the other proposals. So we don’t begin by restricting our space of seriously considered solutions to the anti-sceptical ideas, as is almost always done with traditional sceptical puzzles. My defence of that claim consists in my arguments for the sceptical solution coupled with my criticisms of the anti-sceptical solutions, which make up the rest of the essay. My second main point is that even if some anti-sceptical solution holds, it tells us fascinating things about knowledge and the nature of philosophical reasoning. No matter what the solution to the live sceptical puzzles turns out to be, it will be philosophically important. In this chapter I ﬁrst argued that even if the radical sceptic is right, there are scenarios in which I believe that P but, solely due to differences in CS or CA factors, a certain :P possibility poses more of a threat in one scenario than in the other. I then argued that in the dinosaur story the epistemic threat posed by the solar ﬂare and supervolcano hypotheses was much more signiﬁcant than that posed by the BIV hypothesis: the former have twenty-seven factors trying to defeat the knowledge in question, whereas the latter has just six of those factors. The implication of course is that the live sceptical hypotheses are just like the solar ﬂare hypothesis in generating those twenty-seven threat-boosting factors. I then formulated three strategies that we could use to block the live sceptic’s argument. These strategies look powerful enough to defeat the threat posed by traditional sceptical hypotheses. At the very least, they are powerful enough to prove that the traditional sceptic’s argument is weak (but they don’t show that it is wrong). I suggested that even if the

136

Anti-Sceptical Strategies

strategies work against the traditional sceptic, there is little reason to think that they can work against the live sceptic as well. Instead, we must look fairly closely at well-developed anti-sceptical moves that employ those strategies to see if any of them have a chance at success. In the next three chapters I will be arguing that no such move succeeds.

10

Disarming the Live Sceptical Threat By a ‘silly’ theory I mean one which may be held at the time when one is talking or writing professionally, but which only an inmate of a lunatic asylum would think of carrying into daily life. . . . It must not be supposed that the men who maintain these theories and beliefs are ‘silly’ people. Only very acute and learned men could have thought of anything so odd or defended anything so preposterous against the continual protests of common sense. Charles Dunbar Broad, Mind and its Place in Nature

We disarm the live sceptic’s threat by taking away her sword: making the twenty-one threatening factors lose their punch without meeting them head-on. In this way, the mere mortal need not have any impressive Subject factors, such as evidence, that neutralize the sceptical hypotheses, as the latter never posed any threat that had not somehow been rendered truth-conditionally irrelevant to knowledge assertions. In this chapter I present two such strategies. The ﬁrst, the Set-Aside Solution, claims that we explicitly or implicitly set the live sceptical hypotheses aside as truth-conditionally irrelevant when we assert ‘S knows P’ and P conﬂicts with the hypotheses. The second, the Practicality Solution, claims that in ordinary contexts practicality and other contextualist factors bracket the sceptical hypotheses so that they aren’t threatening.

A When Hypotheses are Set Aside Again: there are twenty-one reasons for thinking that the live sceptical hypotheses are ‘relevant’ in the sense that we require some heavy-duty Subject factors to neutralize them. The anti-sceptic who wants to employ the Disarmed-Threat Strategy to any signiﬁcant

138

Disarming the Live Sceptical Threat

effect will have to reveal some other factor that cancels the effect of these twenty-one highlighting factors—some feature that picks up and sets aside these factors and thereby renders the live sceptical hypotheses unthreatening. Can this be done? Yes it can, at least for many interesting contexts. Fortunately for the anti-sceptic, there are many contexts in which all twenty-one factors are clearly present, the knowledge attribution is true, and yet the subject need not have any Subject factor that neutralizes the sceptical hypotheses. For such contexts the Disarmed-Threat Strategy is successful in blocking the live sceptic’s argument, at least in a sense. Here are the contexts. Often in scientiﬁc conversations and research contexts certain live possibilities are temporarily put aside for the purposes of investigating some theory or research project incompatible with that hypothesis. To see this, consider the ﬁctional water-as-element example given earlier, in which the scientiﬁc community had not yet determined that water is a compound. Suppose two scientists (both mere mortals) have the true belief P that water has to be made of something like nitrogen or oxygen in addition to hydrogen—or it seems that they have this belief, anyway, from a superﬁcial glance at their assertions. Both scholars have had a decent education in science, and each bases his opinion on the assertions of experts and his own scientiﬁc knowledge. One scientist says to the other: ‘Okay, so we know that water has to be made of something like nitrogen or oxygen’. If, previously, the scientists in the conversation had explicitly noted the water-as-element hypothesis and had agreed to set it aside for the sake of argument, all the while respecting its live status, then their use of ‘Okay, so we know that water has to be made of something like nitrogen or oxygen’ would very likely be true. Their use of that sentence would be true because there is an agreed-upon presupposition governing the truth conditions of sentences in the conversation, something that renders the water-as-element hypothesis impotent. The presupposition could be expressed with ‘Setting aside the water-as-element hypothesis for the time being’.46 However, in spite of the truth of their use of ‘We know that water has to be 46

I don’t think that ‘setting aside possibility P (that water’s an element), Q (water is made of nitrogen or oxygen)’ can be replaced by anything like ‘If :P, then Q’, although the speaker might endorse both sentences. The point of the setting-aside conversation is that P is, well, set aside as truth-conditionally irrelevant. Starting out with ‘If:P’ makes P relevant when in fact the whole point was to ignore it. I won’t take any stand on how

Disarming the Live Sceptical Threat

139

made of something like nitrogen or oxygen’, they needn’t even believe that water has to be made of something like nitrogen or oxygen! This is obvious: after all, the participants in the conversation might be advocates of the water-as-element hypothesis who are simply investigating a rival theory. Well, then, what is it exactly that they know in such a situation when they say ‘We know that water has to be made of something like nitrogen or oxygen’? Perhaps what they believe (and know) is better expressed by putting the presupposition this way: they believe and know that, setting aside the element hypothesis, water has to be made of something like nitrogen or oxygen. So the presupposition is part of their belief content. Or perhaps the presupposition is some kind of non-truthfunctional operator in front of the knowledge claim: setting aside the element hypothesis, they know that water has to be made of something like nitrogen or oxygen. Or it is some other kind of relation. And it isn’t clear what the presupposition amounts to: an ignoring of the water-element hypothesis, an assumption of its falsehood, or something else. In any case, this true use of ‘Okay, so we know that water has to be made of something like nitrogen or oxygen’ is elliptical for something else whose truth condition differs from the truth conditions had by ‘literal’ uses of that sentence. But of course this is not the usual way we talk, with explicit settings aside of live hypotheses. Still, even without such an explicit manœuvre, we may implicitly have set the hypothesis aside. Once again, this is clearest if we are advocates of the water-as-element hypothesis who are considering a rival hypothesis. Eventually I may say in the conversation, ‘Well, then. We know that water must be composed of something like nitrogen or oxygen.’ Clearly, as an advocate of the water-as-element hypothesis I am not expressing or reporting or implying the existence of my belief that water must be composed of something like nitrogen or oxygen. I don’t believe that for a minute, and I presumably haven’t asserted it in any way, as I typically don’t assert that I know P when I take P to be false. What I have asserted has a more complex content, one consistent with my advocacy of the water-element hypothesis and perhaps expressed with ‘We know that, setting aside the whole element–compound debate, water must be composed of something like nitrogen or oxygen’, or ‘We know that, assuming water isn’t an element, it must be made of exactly to interpret ‘set aside’, as it won’t matter for my purposes. In all probability, it has no univocal meaning.

140

Disarming the Live Sceptical Threat

something like nitrogen or oxygen’. This holds even without the earlier explicit setting-aside of the water-element hypothesis. What if I’m not an advocate of the water-element theory, and I say, ‘So we know that water must be composed of something like nitrogen or oxygen’? If someone had responded with, ‘Wait a minute. What about the element hypothesis of Smith and Jones?’ and I reacted with, ‘Well of course, but I wasn’t implying they are wrong about anything. I was saying . . .’, then I may not have said anything inconsistent with the water-as-element hypothesis. An advocate of the element hypothesis could perhaps have knowingly agreed with my knowledge claim. I needn’t even believe that the element hypothesis is false. I set it aside beforehand, even if not explicitly. So in this case I have not asserted anything inconsistent with the wateras-element hypothesis, so that hypothesis isn’t incompatible with the truth of the knowledge claim. However, extreme care must be taken here. Just because someone says ‘I know P’, someone else counters with ‘What about suchand-such :P possibility Q?’, and the ﬁrst person responds with ‘Well, of course, but I wasn’t implying that :P’, does not mean that there was a presupposition or other factor that set aside Q as posing no threat. You can’t get off the hook after the fact, not that easily. Just because you say that Q isn’t relevant doesn’t mean it really is irrelevant. It would be nice if it were that easy to respond to a challenge, but it’s not. Think of the theist here, who is a mere mortal regarding atheism. She says that she knows that P, where P obviously entails theism: e.g., ‘I know that God is a loving person’. Just because she might claim that atheism isn’t relevant would not mean that she is right (in fact, she would probably be incorrect). But what if everyone said that Q was irrelevant? Perhaps Q would then be irrelevant, as we control our language to some extent. But when Q is a real, live possibility in the sense uncovered in this essay, then it seems relevant unless one has set it aside by some wellunderstood means. Q is not set aside simply because you say that it is irrelevant. All sorts of nonsense would follow if it were that easy to uphold the status quo. Here is an argument: clearly, my father has not set aside colour error theory in making his colour knowledge claims, and just as clearly we are communicating smoothly in our colour language. Since he isn’t setting the theory aside, and I’m using the same language to express the same thoughts and meanings, I’m probably not setting it aside either.

Disarming the Live Sceptical Threat

141

If instead I had reacted to the challenge that runs ‘Wait a minute. What about the water-as-element hypothesis of Smith and Jones?’ with ‘Oh, that’s right. Fair enough. I forgot all about the element idea. Well, if water is a compound, then my proposal is right’, then it looks as though I initially did say something inconsistent with the water-as-element hypothesis but then retracted it and asserted something else consistent with it. And surely my original use of ‘We know water must be made of something like nitrogen or oxygen’, which we just said was inconsistent with the water-as-element hypothesis, can’t be true unless I have some combination of Subject factors that neutralize that hypothesis. This is why I’m retracting it in the ﬁrst place. I know that I’m in no position to know that water is made of nitrogen or oxygen. I can’t say, straight off, literally, without any assisting presuppositions, that I know that water must be composed of something like nitrogen or oxygen unless something about me neutralizes the water-as-element hypothesis. If I reply with ‘Oh come on. I don’t take that theory seriously’, then we have an assertion and a belief inconsistent with the water-as-element hypothesis. And once again, my knowledge attribution is probably out of bounds: I know no such thing unless the water-as-element hypothesis is neutralized. In addition, uses of ‘S knows that water has to be made of something like nitrogen or oxygen’ in some conversational contexts might mean something better expressed with ‘S knows that water has to be made of something like nitrogen or oxygen—not sodium’. Alternatively, S knows that water has to be made of something like nitrogen or oxygen rather than sodium. Another shot: S knows that of the three options—sodium, nitrogen, and oxygen—only the last two are real possibilities. In a context with the twenty-one contextual factors listed earlier, such a knowledge ascription could express something true even if S didn’t have Subject factors that neutralized the water-as-element hypothesis. But once again, these uses strike me as non-literal uses of ‘S knows that water has to be made of something like nitrogen or oxygen’. For one thing, literal and true uses of ‘S knows that water has to be made of something like nitrogen or oxygen’ entail that water has to be made of something like nitrogen or oxygen, whereas the ‘rather than’ sentence, ‘S knows that water has to be made of something like nitrogen or oxygen rather than sodium’, does not. It isn’t clear what the latter sentence entails. For another thing, if you asked the conversational participants if what they really meant was the ‘rather than’ sentence, and they didn’t concur, then it

142

Disarming the Live Sceptical Threat

seems to me that they didn’t mean the ‘rather than’ sentence. Not because we’re infallible regarding semantics, but because we are usually sharp enough to know when we’re employing a ‘rather than’ qualiﬁcation. In the sodium case, the conversationalists were probably weighing the merits of water’s being composed of sodium, nitrogen, and oxygen. After ruling out to their satisfaction the sodium possibility, they came up with ‘Okay, so we now know that water has to be made of something like nitrogen or oxygen’. In this case the ‘rather than’ reading is plausible; otherwise not. (Plausible, but certainly not required. They may have thought that the sodium, nitrogen, and oxygen possibilities were the only ones with any plausibility at all; so after ruling out the sodium one, they concluded that it’s one of the other two—full stop. They don’t conclude that it’s one of the other two rather than sodium.) It would be silly to think that all belief has the ‘rather than’ form. For instance, I believe that ﬁre-engines are red, full stop; and not just ‘red rather than orange’ or ‘more red than orange’. I have those ‘rather than’ beliefs as well, but the belief I’ve had since childhood is the one expressed without a ‘rather than’ clause. And the same holds for you! So different uses of ‘We know that water has to be made of something like nitrogen or oxygen’ have very different consequences. Sometimes what is expressed is true but consistent with the water-aselement hypothesis. There is nothing surprising about this; the live sceptic will surely not be bothered with such contexts of use in which the Disarmed-Threat Strategy is successful. However, it seems that what we won’t ﬁnd is a use of that sentence that expresses just its literal, non-elliptic, straightforward meaning and is such that it can be true even if the subject has no Subject factors that neutralize the water-as-element hypothesis. A use of ‘S knows that water has to be made of something like nitrogen or oxygen’ won’t be true if it is literal, the twenty-one contextual factors are present, and there is no appropriate combination of Subject factors that neutralize the wateras-element hypothesis. As we have seen, a use of that sentence can express a truth in no need of such Subject factors, but only when construed as elliptical or otherwise non-literal. It seems to me that this is pretty solid. So this use of the Disarmed-Threat Strategy fails right where we want it to succeed: with literal, non-elliptical knowledge assertions. It also seems to me that what holds in the water-as-element case holds for the live sceptical hypotheses: given that S, A, and A’s

Disarming the Live Sceptical Threat

143

audience are mere mortals and the sceptical hypotheses are live in their shared community, it follows that A’s literal use of ‘S knows P’, where P is obviously inconsistent or incompatible with the hypotheses, can be true only if S has some potent Subject factor that neutralizes the hypotheses. That certainly holds for the wateras-element or whales-aren’t-ﬁsh cases and thousands of other ordinary cases. In effect, this looks simple: Alf needs to rule out the water-aselement hypothesis because otherwise he would be believing it in the face of contrary evidence that hasn’t been ruled out and is hanging over us all, assuming he can’t rule it out. You just can’t get away with that: if you look contrary evidence in the eye, understand it and its signiﬁcance as contrary and well-respected evidence, and then go on to believe P anyway, well then your belief won’t amount to knowledge unless you can rule out, in some sense, the contrary hypothesis. That is, you need a combination of Subject factors that neutralizes the live contrary hypothesis; the Set-Aside Solution fails for virtually all contexts, the ones in which the knowledge ascription is literal. Okay, pretend that I’m completely wrong: in virtually all contexts there are presuppositional factors or ‘rather than’ clauses or initial linguistic operators that change the most obvious truth conditions for the naked ‘S knows that water must be composed of something like nitrogen or oxygen’. Well, then: now it looks as though our knowledge assertions ‘S knows P’ are quite consistent with our saying at the very same time ‘But P is false’. That’s quite unlikely. Or is this the truth: that not only do our uses of ‘S knows P’ not entail P, but they don’t entail that S believes P either? This is implausible as well. Switching to another case: do any philosophers or colour scientists believe that ﬁre-engines are red or not? If not, then we have the amazing result that none of us mere mortals believes anything inconsistent with the sceptical hypotheses. That’s incredibly unlikely. No one believes that ﬁre-engines are red, full stop? None of us believes that Moore believed scepticism to be false? And even if we really don’t have any of those anti-live hypothesis beliefs, the sceptical result would follow anyway : none of us knows that ﬁre-engines are red or that Moore believed scepticism false because we don’t even believe those things. If that is the solution to the sceptical puzzles, so be it; it’s still shocking, and it’s even a sceptical solution, not an anti-sceptical one. Or is there no such thing as a naked proposition in these cases, but rather all such belief

144

Disarming the Live Sceptical Threat

contents have hidden presuppositional or ‘rather than’ or whatever parts? But wouldn’t that affect the sceptical hypotheses as well—so the inconsistency is retained? Now it looks as though we may, in order to save the Set-Aside Solution, have to erect an entirely new theory of belief content, completely different from anything we have worked on. Concerning these matters I’m conﬁdent that any hopeful proposed solution will be philosophically quite interesting. B Knowledge Attributions in Ordinary Contexts Recall that contextualists hold that different uses of ‘S knows P’ can have different truth-values even when concerning the same person, truth, and time. The idea is that depending on the context of utterance, different levels of epistemic standing are necessary for the truth of the use of ‘S knows P’. In some contexts—for instance, when a lot is riding on whether P is true or S’s belief can be relied on—then S’s warrant has to meet a high standard to make the knowledge attribution true. In other circumstances—when, for instance, little rides on the truth of P—S’s warrant need meet only a low standard. This theory schema can be ﬁlled out in many ways, depending on how one ﬁlls out ‘standards’, what one says about which contextual factors can raise or lower standards, and what one says about how much those factors can raise or lower standards. It is thought to offer a plausible response to scepticism by making room for the claim that in discussions of scepticism the standards get raised absurdly high, so that when the sceptic concludes with ‘So, we don’t know anything’, she speaks the truth—even though in other contexts of utterance, governed by low standards, we speak the truth when we say ‘I know that she hates my motorcycle’. But can the theory provide a decent response to live scepticism? Take as our example eliminativism. As I mentioned at the end of Chapter 9, it’s natural to say that in theoretical contexts (e.g., discussions of connectionism or Fodor’s theory of content) eliminativism deﬁnitely poses a threat that can be neutralized only with some strong epistemic factors. The live sceptic is right about knowledge attributions in those contexts. But what about the completely ordinary conversational contexts in which mere mortals discuss their

Disarming the Live Sceptical Threat

145

alleged knowledge of a friend’s beliefs? Or the colour of their socks? Or whether the doctor is aware of the new pain in their father’s hip? Here the error theories are far from anyone’s mind. Sure, the twentyone liveness factors are still present, but they aren’t strong enough to mount an epistemic threat that needs to be defused. Or so the objection to live scepticism goes. For the sake of argument in the remainder of this section, I will assume that contextualism is true. The issue here is the truth-value of a speciﬁc kind of contextualism, one that implies that, for instance, ‘I know you think motorcycles are dangerous’ is true when the mere mortal conversational participants are not thinking or talking about live eliminativism (that’s the objection to live scepticism), even though when they discuss eliminativism or other relevant theoretical matters appropriately, the very same knowledge sentence is false (that’s the concession to live scepticism). I have six things to say about this kind of contextualism/practicality move. First point. I personally dislike burden-of-proof arguments more than most other philosophers do, as I think that on virtually any important issue everyone has ‘the’ burden of proof, even though that makes our lives so much more difﬁcult. But it seems to me morally certain that the onus is on the anti-sceptic to provide a damn good argument that, despite the twenty-one highlighting factors, despite the fact that the sceptical hypotheses have not been set aside, and despite the fact that we have no potent Subject factors to neutralize the sceptical hypotheses—all of which this practicality anti-sceptic is prepared to admit—our ‘S knows P’ assertions can be true anyway, due to facts about practical purposes. It is plausible to pursue the Practicality Solution when it comes to traditional sceptical hypotheses. It is plausible to think that the BIV hypothesis doesn’t threaten the truth of ordinary knowledge attributions unless it is currently being discussed or thought about appropriately. It is hard to see what else could make the BIV hypothesis a threat to one’s belief that one has hands. However, this doesn’t mean that occurrent discussion and thought are the only ways to make a hypothesis threatening. On the face of it, liveness conditions make a hypothesis even more of a threat than mere occurrent attention does. Given the twenty-one highlighting factors, there is no way that we can get away with saying ‘Well, the new sceptical hypotheses ain’t highlighted either’; one needs a very fancy argument to make this at all plausible. Only three of those twenty-one factors are present for the traditional sceptical hypotheses,

146

Disarming the Live Sceptical Threat

so there is no reason to think that the practical, contextual irrelevance of the traditional hypotheses carries over to the live hypotheses. We can’t just say, ‘Well, we’ve already seen that sceptical hypotheses can be truth-conditionally irrelevant in some contexts; so that’s what must be happening here too’. Neither can we get away with ‘Well, we know that if one is just minimally rational, then one has neutralized the BIV hypothesis; so the same must be true for the colour error theory hypothesis’. Contrary to what you might have thought initially, in the ﬁrst part of this section I did not deliver a decent objection to the live sceptic limited to ordinary contexts of knowledge attributions. All I did was motivate one. Unless we have the magical reason why the twenty-one highlighting factors collectively fail, the objection goes nowhere. I don’t deny that one can formulate a contextualism according to which, if we’re not talking or occurrently thinking about colour error theory, then it isn’t highlighted and so presents no threat even if it is live and we are mere mortals. But this strikes me as an inferior kind of contextualism; this leads to my second criticism of the Practicality Solution, as follows. Second point. Consider how such a theory would work with a different example. Suppose that every baseball book says, correctly, that Hank Aaron hit 755 home runs in his career. Sarah knows that that’s what everyone thinks; and she agrees with them because she has read those books. Then Smith and Jones get archival evidence that suggests that in a 1958 game in which Aaron hit two home runs in the ﬁrst four innings, the game was halted after four innings on account of rain. And the game was never continued. And the newspapers made a mistake the next day and reported the game as rained out after six innings. And a game is ofﬁcial if and only if it goes at least ﬁve innings. And because the newspapers said that the game went six innings, the statisticians treated the game as ofﬁcial. And unless a game is ofﬁcial, what happens in it doesn’t count towards one’s statistics. (Most of this is almost true.) So if Smith and Jones are right, Aaron hit only 753 home runs. Suppose, further, that many people are very impressed by the evidence uncovered by Smith and Jones; the Smith–Jones hypothesis is live (but still the underdog). Sarah hears about the new hypothesis and becomes aware of its respected standing. Still, she thinks it’s probably false, although she is a mere mortal regarding it. So she still believes that Aaron hit 755 home runs. Can you think of a context in this story in which ‘Sarah

Disarming the Live Sceptical Threat

147

knows that Aaron hit 755 home runs’ is true even after Sarah becomes a mere mortal with respect to the live 753-homer hypothesis? I can; but only if the sentence has some non-literal reading. If it meant something better expressed by ‘Setting aside the whole Smith– Jones thing, Sarah knows that Aaron hit 755 home runs’, then maybe it would express a truth. The ‘setting aside’ sentence may be true, but I don’t think it conﬂicts with (the literal reading of) ‘Sarah does not know that Aaron hit 755 home runs’, as it doesn’t even entail that Aaron hit 755 home runs. That is, ‘Setting aside the whole Smith– Jones thing, Sarah knows that Aaron hit 755 home runs’ doesn’t entail the truth of ‘Aaron hit 755 home runs’. The ‘setting aside’ sentence is more like ‘Sarah knows that: assuming Smith and Jones are wrong, then Aaron hit 755 homers’, which obviously doesn’t entail that Aaron hit 755 homers. Even Smith or Jones could comfortably say ‘Setting aside our hypothesis, Sarah knows that Aaron hit 755 home runs’. They would say it, believe it, and in saying it they would not be contradicting their belief that Aaron didn’t hit 755 home runs. We saw these points in the previous section of this chapter. But other than taking ‘Sarah knows that Aaron hit 755 home runs’ to have some non-standard interpretation, I don’t see how it could be true unless she can rule out the 753-homer hypothesis, and the reason is that the latter hypothesis is live and Sarah is aware of that fact, and she is aware that it conﬂicts with her belief. If Fred isn’t thinking of the 753-homers hypothesis, he says ‘Sarah knows that Aaron hit 755 home runs’, and he means it literally, then what he says is false even if it’s conversationally appropriate. It simply doesn’t matter that the 753-homer hypothesis isn’t ﬂashing on a screen over Fred’s head; whether Fred is thinking of, or has recently mentioned, the 753-homer hypothesis just doesn’t seem relevant. I’m not beating up on contextualists here; this isn’t to say that contextualism is false. I can go along with the contextualists when they claim that in some contexts ‘Sarah knows that Aaron hit 755 homers, assuming Smith and Jones are wrong’ and ‘Sarah knows that Aaron hit over 700 homers’ are true while in others they are false due to various traditional sceptical hypotheses becoming highlighted for standard contextualist reasons. I’m addressing only what I consider to be a relatively extreme kind of contextualism. I’ve overstated the case slightly. I can see the attraction of claiming that a god or a suitably wise twenty-ﬁfth-century person would speak

148

Disarming the Live Sceptical Threat

truly using ‘In the twenty-ﬁrst century Sarah, like most people, knew that Aaron hit 755 homers’. Perhaps they would speak truly because they know that the 753-homer idea was a false and short-lived aberration, and from their god’s-eye perspective can be ignored. I doubt it, but in any case, for anyone in Sarah’s community— especially the mere mortals—a use of ‘Sarah knows that Aaron hit 755 homers’ is false in every way. Third point against the Practicality Solution. If this solution assumes that the practical irrelevance of the error theories is so extreme that they cannot really pose an epistemic threat large enough to challenge signiﬁcantly one’s everyday, non-theoretical knowledge attributions, then the objection rests on a falsehood. We must remember the distinction between the practicality of the attributionaction (the action of saying ‘S knows P’) and the truth of the attribution-proposition (the proposition expressed by one’s utterance of ‘S knows P’). The objection can confuse the two. I fail to see why we should think that practicality concerns introduced by context affect truth conditions to such an extreme extent when it’s obvious that they affect in the ﬁrst instance the conversational appropriateness of actions. In Chapter 6 the live sceptic admitted, and even accounted for, the conversational and functional worth of the attribution ‘I know that the table is brown and the cause of the pain in Fred’s knee after he ran into it; and I know that Fred believes all this too’.47 Fourth point. Many of us philosophers have long been acutely aware that in ordinary life we say that we know truths that are obviously inconsistent with eliminativism, colour error theory, and many other live philosophical theories. The funny thing is that most of us don’t feel at all embarrassed by this curious situation. If challenged, we are inclined to say to ourselves things like this. Well, yes, of course colour error theory is a real, live possibility. It’s been endorsed by lots of people and there are good reasons to accept it. Same for eliminativism. I’m not ignorant of the state of play. And I’m quite prepared to admit that I wouldn’t have a snowball’s chance in hell of defeating either of those theories—or even countering the 47 I don’t mean to say that contextualists make this mistake; remember that in this section I’m assuming that contextualism is true. Much more could be said about practicality and truth conditions, and much that is worthwhile has been said by DeRose (1999, 2002, 2005). Jessica Brown 2005 is a good rebuttal to DeRose 2002.

Disarming the Live Sceptical Threat

149

arguments for those theories. But still, there is nothing wrong with me saying in ordinary life that I know that my socks are blue or that I know that Moore believed that scepticism is false. On the face of it, this is awfully strange (compare this with the scientist in the analogous purely scientiﬁc situation). Yes, we do this all the time, making knowledge assertions that we know quite well contradict what our esteemed colleagues say—colleagues we often look up to and consider brighter and better informed than we are. I don’t mean that when we’re working on theories of content and colour we conﬁdently assume that we know people have beliefs and that ordinary physical objects are coloured. I mean that in ordinary contexts we conﬁdently say that we know that she believes that your driving is awful and that his tie is multicoloured even though we are perfectly aware that some of our best and brightest say that we can’t know those things because they aren’t true. And I’m not saying that it’s strange that we say ‘His socks are blue’ even though we’re aware that so-and-so brilliant philosopher-scientist says that no socks are blue. That is a bit odd as well, but quite different. In that case we’re just registering our disagreement. The oddity I’m focusing on is that we say that we know that his socks are blue, even though we’re aware that the esteemed colleagues say that the socks aren’t blue. This is a more extreme case, one that makes us look arrogant, at least on the surface. I suspect that there are two good reasons why this situation is not epistemically vicious in spite of its oddity. First, the truth of the literal construal of our words often doesn’t matter in ordinary life, and we act as though we know this. I suspect that most philosophers underestimate the importance of this point. Sometimes in making ordinary knowledge claims we are tagging on operators like ‘For all practical purposes’ or ‘Setting aside philosophy for now’—at least, we would give explanations of our linguistic behaviour that suggest such interpretations. Other times we care only about the practicality of knowledge attribution-actions, and not so much the truth of attribution-propositions. In my own case, when doing my day-today chores or engaging with friends, the truth of what is said is often not to the point (but not most of the time!). At least that’s my observation, and my conversations are pretty ordinary. Second, we don’t know what else to do. We simply cannot stop making these knowledge assertions regarding claims inconsistent with the new sceptical hypotheses. We can easily talk in such a way

150

Disarming the Live Sceptical Threat

that we remain assertively agnostic regarding whether anyone knows that black holes really exist or whether there is intelligent life on other planets. But we simply cannot do so regarding whether we know that Moore believed that scepticism is false or whether we know that ﬁreengines are red. At least, we cannot do so while remaining communicating members of our communities (especially if we are teachers or parents). But surely this doesn’t mean that we can know that Moore believed that scepticism is false without having some epistemically relevant ‘warrant factor’ available to neutralize the live sceptical hypotheses. Just because you are compelled to speak as though S knows P even though P is contradicted by real, live possibility Q, doesn’t mean that the truth-value of ‘S knows P’ isn’t threatened by the liveness of Q. In some communities you can barely communicate unless you talk as though you are a theist who knows all sorts of things about God, but when you say ‘S knows P’ where P is something that obviously entails the falsehood of atheism, you aren’t off the hook just because it’s so hard to avoid making such knowledge claims in those contexts. Maybe our own talk regarding how we know things about how the sun moves behind the trees commits us to clearly false astronomical theories. We engage in the talk anyway, because the truth-value of the literal construal of our words doesn’t really matter (that’s the point of the previous paragraph) and because it would be a real pain in the ass to stop talking that way (that’s the point of this paragraph). Perhaps our knowledge assertions regarding numbers commit us to heavy ontological claims regarding mathematical objects. But in either case the truth of our knowledge assertions is often of little practical importance, and the fact that we can’t stop making the claims proves nothing. Philosophers often don’t want to admit the inconclusive nature of the most crucial points of their arguments. Let’s not have that happen here: the above arguments against this Practicality application of the Disarmed-Threat Strategy are not conclusive, at least by my lights. So suppose that despite everything I just argued, the antisceptical solution for ordinary, everyday contexts is right, so that in ordinary-life contexts ‘I know my socks are blue’ is true—even though all twenty-one highlighting factors are present in both everyday and theoretical contexts. I don’t care why this would be true. Perhaps in some manner we implicitly set aside as truth-functionally irrelevant the live sceptical hypotheses. Or perhaps practicality facts set them aside. Or some other fact turns the trick; it doesn’t matter.

Disarming the Live Sceptical Threat

151

My last two points address the signiﬁcance of the success of this antisceptical idea. Fifth point. I have given some pretty good arguments that in theoretical contexts—ones that analytic philosophers and cognitive scientists ﬁnd themselves in virtually every day (depending on how hard you work!)—the new sceptical arguments go through. I ﬁnd that sceptical result still pretty amazing. If it’s right, then you don’t know much of anything right now about the colour of the objects around you, anyone’s beliefs, or the pains in your knees. And you don’t know any of that in many, many philosophical and even scientiﬁc contexts. In spite of the recent endorsement of scepticfriendly versions of contextualism, according to which we know very little if we’re discussing a classic sceptical hypothesis and argument in the right way (thereby making those hypotheses mount real epistemic threats, despite being nowhere near live), I still ﬁnd it almost unbelievable that we don’t know, right now, simple facts about the colours of our socks, our aches and pains, or what we believe. The sceptic-friendly response to live scepticism is much more radical than the sceptic-friendly response to traditional scepticism. Last point. Suppose, as in the previous point, that the antisceptical solution for ordinary, everyday contexts is right. Then in my judgement the anti-sceptic has won a battle but lost the war. Knowledge attributions are often true (that’s the battle won), but the truth conditions of those attributions are so meagre that we should feel a little ashamed (that’s the war lost). Is that all it takes to make ‘I know my socks are blue’ true—and my belief only barely meets that pathetic standard? There are twenty-one reasons for thinking that I need some powerful Subject factors to defuse the expertly endorsed, highly respected contrary scientiﬁc hypotheses—and yet you’re telling me that I can know lots of things inconsistent with those hypotheses just because I don’t currently care about them and they don’t matter when I’m getting dressed? Or because I have somehow set them aside even though the thought of setting them aside has never occurred to me (and virtually no one else)? Tell me why this is supposed to be a victory that the anti-sceptic should celebrate. On the contrary, we see that our epistemic position in such a case is so impoverished that even if it warrants the title ‘knowledge’,48 this is to concede to the 48 Earlier, in sect. A, we saw that this concession cannot be made easily. If my claim ‘I know that P’ really means something along the lines of ‘Setting aside Q, I know P’, then it

152

Disarming the Live Sceptical Threat

sceptic her fundamental point that our epistemic position is impoverished compared to what anyone thought it was. We can stick to our anti-sceptical urges if we want; there is nothing really stopping us; but we will ﬁnd that we are winning a prize not worth having. We have defeated the letter of scepticism and saved knowledge, but the champagne is pretty ﬂat. isn’t at all clear that the truth of my knowledge assertion even requires my belief in P. So it might not require or even reliably indicate my knowledge of P.

11

Do Live Sceptical Hypotheses Pose Real Threats? Who can refute a sneer? William Paley, Moral Philosophy Russell’s view that I do not know for certain that this is a pencil or that you are conscious rests, if I’m right, on no less than four distinct assumptions. . . . And what I can’t help asking myself is this: Is it, in fact, as certain that all these four assumptions are true, as that I do know that this is a pencil and that you are conscious? I cannot help answering: It seems to me more certain that I do know that this is a pencil and that you are conscious, than that any single one of these four assumptions is true, let alone all four. G. E. Moore, Four Forms of Scepticism

The No-Threat Strategy claims that the twenty-one factors are too feeble to mount a threat, so it doesn’t matter whether factors (22)–(27) count heavily against the power of our Subject factors in neutralizing the sceptical hypotheses. On the face of it, the twentyone factors make it look as if the sceptical hypotheses are about to ruin our chances at knowledge, but this is an illusion. In this chapter I present and evaluate eight solutions that ﬁt the No-Threat Strategy. The Metaphysical Distance Solution complains that the sceptical hypotheses aren’t just false, they are nowhere in the metaphysical vicinity of true, and this ruins their ability to mount anything but the most meagre threat. The succeeding four solutions develop in different ways the idea that there is something peculiar about philosophy or common sense that taints the sceptical hypotheses and renders them non-threatening. The Double Standard Solution says that the quality of being philosophically

154

Are the Threats Real?

live (as opposed to scientiﬁcally live) does not carry much epistemic weight. The Robustly Live Solution is similar: the notion of liveness involved in the sceptical hypotheses, although at least partially scientiﬁc, is importantly different from the liveness of epistemically threatening scientiﬁc hypotheses. The Bad Reasoning Solution claims that there is some bad reasoning involved in the evidence for the sceptical hypotheses that drains them of epistemic import. The Hume–Moore–Reid Comparison Solution claims that the positive warrant for the ordinary claims contradicted by the sceptical hypotheses so outweighs the negative warrant produced by the liveness of the sceptical hypotheses, that the epistemic damage produced by those hypotheses is easily absorbed without loss of knowledge. This latter proposal isn’t an instance of any of the three anti-sceptical strategies, as it fails to say why the live sceptic’s argument fails, but I include it here due to its afﬁnity with the other three ‘Philosophy versus Common Sense’ proposals. Sixth, the Limited Scope Solution insists that since almost all of us can see through the meagre support had by the new sceptical hypotheses, for only very few people are they live in any threatening way. Seventh, the Ambiguity Solution claims that live scepticism is true under one disambiguation and false under the other. Finally, the Testability–Indeterminacy Solution complains that the sceptical hypotheses cannot be fruitfully established one way or the other, and somehow this negates the threat posed by those hypotheses. As in the previous chapter, I will attempt to show that each of these strategies is either impotent or renders knowledge fundamentally impoverished. A Metaphysical Distance from Actuality The worlds in which the new sceptical hypotheses are true are so metaphysically distant from our world that even if some experts have confused them with the actual world, they still don’t pose much of an epistemic threat to our beliefs. Sure, our beliefs might lose a bit of their warrant status when the scientists and philosophers take seriously a theory that conﬂicts with those beliefs, but when the theories are so radically wrong—eliminativism!—they are too distant to have much epistemic import. In such cases, because the conversational participants are in a possible world metaphysically distant from worlds in which the sceptical hypotheses are true, we don’t need

Are the Threats Real?

155

the powerful Subject factors because the hypotheses aren’t really highlighted. This is the Metaphysical Distance proposal. This proposal is intuitive, as it’s sometimes the metaphysical remoteness of the traditional sceptical hypotheses that lies at the base of our hunch that they don’t pose a real epistemic threat to our beliefs. Eliminativism and colour error theory are different from the evil demon and BIV hypotheses in being live, but it is easy, perhaps comforting, to think that all these hypotheses, new and traditional, are in the same metaphysically distant boat. Thus, although my live sceptical hypotheses are novel in being epistemically different from the traditional ones, they fail to sabotage our knowledge for the same reason the classic ones do. Unfortunately, the counter-examples to this solution are so numerous and ordinary that it is drained of virtually all of its force— even when we waive the problem that the objection is trying to get a particularly heavy epistemological consequence out of metaphysical fact.49 The scientiﬁc hypothesis I started with—water is an atomic element—is true only in metaphysically distant possible worlds (if any!). And of course this is typical: there have been loads of scientifically live but metaphysically exceedingly distant hypotheses. Yet the metaphysical distance does not rob the hypotheses of the ability to ruin our knowledge. So the mere fact that eliminativism is true only in metaphysically distant worlds does nothing to diminish its epistemic threat. The only real difference between the water-as-element hypothesis and the eliminativist or colour error hypotheses (or the no-character-traits or pain-only-in-the-head hypotheses) is that only the latter are anti-commonsensical. That’s precisely why the latter produce philosophical interest while the former does not. My guess is that it’s also the only reason that we (or, to be more precise, I) bother to search for anti-sceptical solutions to the live sceptical arguments, whereas we (I) don’t do so for the philosophically uninteresting ones. Since we are all familiar with science poking holes in common sense, we shouldn’t get too worked up when colour scientists and cognitive scientists try to do the same, even when philosophers accompany them. 49 Obviously, the notion of metaphysical distance would need to be carefully and appropriately elaborated in order for the proposed solution to stand any chance at being adequate. I’m ignoring that problem here, as I think the other problems are sufﬁcient to doom the proposed solution.

156

Are the Threats Real?

B Philosophy versus Common Sense Philosophy never seems to resolve anything. At the very least, it has a particularly awful record when it comes to overthrowing common sense. However, it’s clear that the live sceptic is attempting to do just that. So we have an inductive argument against the live sceptic. Needless to say, this isn’t a very strong objection. Can it be made into something formidable? At the end of Chapter 5 I stated that the sceptical argument template probably didn’t generate a convincing argument when non-scientiﬁc hypotheses (e.g., ‘God doesn’t exist/does exist’, ‘Lying isn’t/is morally wrong’) are plugged into it. This brings up the ﬁrst promising ‘philosophy versus common sense’ solution to the sceptical puzzle. When I warmed us up to the live sceptic’s reasoning in Chapter 1 I used a scientiﬁc hypothesis for illustration: water is an atomic element. That was adequate to motivate the operative principles behind the new sceptical arguments. But when I turned to squeezing some philosophical juice out of those principles, I plugged in philosophical hypotheses, not scientiﬁc ones. And, one might think, there’s the rub. We defer to scientiﬁc experts and liveness. But we don’t for philosophical ones. With good reason! There seems to be a pretty tight connection between being scientiﬁcally live and being probably-roughly-true: if a hypothesis has the former quality, then there is good reason to think that it has the latter quality. That is why a scientiﬁcally live hypothesis that conﬂicts with your belief poses a formidable epistemic threat to your belief, a threat that must be defused in order for the belief to be knowledge (at least provided you’re aware of the threat). But no such connection holds between the characteristic of being philosophically live and being probably-roughly-true. So just because eliminativism is philosophically live doesn’t make it threaten our beliefs, contra the live sceptic. One can add a Moorean point to this pessimism about the power of philosophy. If I know that (a) hypothesis H contradicts proposition P, (b) P but not H is part of robust common sense (something along the lines of being part of ﬁrmly ingrained, cross-cultural, and virtually timeless common sense; the vagueness here won’t matter), and (c) the only support for H is philosophical, then H doesn’t pose any signiﬁcant epistemic threat to my belief in P. (More on Moore below.)

Are the Threats Real?

157

The upshot of this pessimism is that we should hold a double standard: the sceptical argument schema is correct for scientiﬁcally live hypotheses, but not for merely philosophically live ones. The quality of being philosophically live does not carry epistemic weight. More precisely: although it carries some weight, it’s not much compared to being scientiﬁcally live. We are not saying that a purely philosophical hypothesis that goes against a belief of yours doesn’t pose any threat at all to that belief; we’re only saying that it doesn’t pose a formidable one. Thus, this Double Standard Solution is a version of the No-Threat Strategy. It is saying that the sceptical hypotheses would be highlighted only if they were sufﬁciently scientiﬁc or otherwise respectable, but they aren’t. I have several things to say about this Double Standard Solution. First, it isn’t clear that its main source, a fairly deep pessimism regarding the power of philosophy, is warranted. One often hears, and voices, complaints about the failure of philosophical argument to settle any interesting dispute. This makes it look as though historical facts suggest that philosophical reasoning is epistemically quite weak. But, on the other hand, the complete and utter failure to defuse certain philosophical arguments suggests, to me anyway, that the philosophical reasoning in these arguments is not weak—on the contrary, it’s extraordinarily strong. I made a similar point earlier, when articulating the sceptical solution. Many of the greatest minds in history have tried but failed to defeat traditional philosophical arguments for scepticism. Centuries of exceedingly careful and ingenious investigation of these arguments have yet to put them to rest. Sure, the anti-commonsensical conclusions are counter-intuitive, but after so much brilliant evaluation, how rational is it for a philosopher to announce that those arguments are so weak? Perhaps the right conclusion is that historical facts decide nothing regarding the strength or quality of philosophical reasoning.50 This issue raises difﬁcult questions about the nature of philosophy and philosophical progress that I would like to avoid at all costs. So, for the sake of argument, I’ll grant the Double Standard Solution in its entirety (not because I agree with it but because I have no good argument against it). Thus, I will assume that if we plug a hypothesis 50 In addition, depending on how one understands ‘probably-roughly-true’, there just might be a tight connection between being philosophically live and being probablyroughly-true.

158

Are the Threats Real?

motivated by nothing but pure philosophy into the argument schema, then the resulting sceptical argument is not strong enough to be taken seriously. So it won’t do to focus on ‘No one is free’, ‘Murder isn’t morally prohibited’, ‘God doesn’t exist’, ‘Double negation isn’t valid’, ‘Chairs don’t exist’, ‘Numerical statements aren’t truth-apt’, etc. Even if this is right, the solution needs a lot of supplement to block other interesting applications of the argument template, the ones I’ve been emphasizing in this essay. First, in close possible worlds cognitive scientists consider eliminativism a live hypothesis. That hypothesis is live in the scientiﬁc community. The same holds for colour error theory; in actual fact, it is clearly one of the dominant scientiﬁc views, although, as you would expect, most scientists don’t consider how their theories affect the truth conditions or truth-values of sentences such as ‘Fire-engines are red’ and ‘I know that ﬁre-engines are red’. The same holds for pain location. So perhaps the Double Standard Solution should conclude like this: the argument template is correct for hypotheses that are live for scientiﬁc reasons, not those whose liveness is traced to philosophical reasons. But that won’t do either, because there are non-philosophical reasons for eliminativism and, more convincingly, colour error theory is ﬁrmly rooted in purely scientiﬁc investigation, especially in many of our neighbouring possible worlds. Moreover, colour error theory is an obvious option to anyone who thinks about colour in an intelligent way. As a matter of actual brute fact, there often is no philosophical reasoning behind the personal discovery of colour error theory.51 Similar points hold for character trait error theory. It is true that philosophers dreamt up eliminativism, but obviously this is a purely accidental matter. It is easy to imagine the clever scientist who studies cognition via studying the brain beginning to wonder how plausible it is that she will ever ﬁnd beliefs in the head. She could quite easily discover and investigate several forms of eliminativism all the while thinking that philosophy is nothing more than navel gazing. Even if all philosophy departments were banned, no one read Descartes, and all philosophical literature destroyed, these sceptical hypotheses 51 I say this from experience. I studied enough physics to earn a Master’s degree, and many good physics students discovered for themselves and took seriously colour error theory. Of course further education often doesn’t take away the attraction of error theory, as many scientists advocate it as well.

Are the Threats Real?

159

could very easily reach live status—in a society without any traditional philosophy. The point is that sceptical hypotheses can have scientiﬁc credentials and originate from scientiﬁc investigations. Say scientists study cognition for 200 years and develop a great set of uniﬁed theories— just as in chemistry. ‘Belief ’ never shows up anywhere. And there’s nothing even remotely corresponding to it. Now the scientists start saying, ‘Look, if you think it’s really strictly true that we have loads of beliefs, then show me where they are. If they are really cognitive states of particular individuals, then why is it that we see none of them when we study the cognitive states of the individuals? The science simply goes against folk psychology. And don’t give me any watered-down truth-maker crap that doesn’t involve individual cognitive states. If there are beliefs, they’ve got to be in the head; everything else is just loose talk.’ Disagree all you want; they have their reasoning. The lesson of these considerations is that we could reasonably add a sixth condition to the notion of liveness: the liveness of the hypothesis is scientiﬁc and is, or very easily could be, originally due almost entirely to scientiﬁc, non-philosophical considerations. One might object that when scientists endorse eliminativism or colour error theory they are doing so not as scientists but as philosophers, and pretty lousy ones at that. In some cases this is surely true. But eliminativism is often, even usually, taken seriously because of developments in cognitive science, connectionism research in particular. If one thinks that connectionism provides a good understanding of the mind, then one need not be a genius or a philosopher to see that it is going to be difﬁcult and forced to ﬁnd beliefs in the head. The only way to save the Double Standard Solution is to make it extreme: if there is even the slightest hint of philosophizing behind some hypothesis that contradicts some of our most cherished ordinary beliefs, then the hypothesis does not amount to a threat; it is not highlighted. As a matter of unpleasant fact, most philosophers are often in the position of the ﬁrst-year graduate student: they admit they hold beliefs inconsistent with live theories, even though they are no more expert on the relevant topics than the graduate student is. Is this some kind of institutionalized arrogance (or at least wild selfconﬁdence)? The radical Double Standard advocate says not; she thinks there is something very special about philosophical reasoning

160

Are the Threats Real?

that drains it of any signiﬁcant epistemic force. If the aetiology of the liveness of a sceptical hypothesis has even a hint of philosophy in it or is otherwise less than robust as described earlier, then it’s automatically divested of epistemic power. If this is the right solution, then surely it says something interesting and highly negative about philosophy. If you advocate it, then why are you a philosopher anyway? Why on earth would you bother reading a philosophical work such as this essay? It also suggests that you know that ﬁre-engines are red and that Moore believed scepticism false partly due to the fact that philosophical reasoning is fatal to epistemic liveness. Let’s move on to consider a slightly different and more interesting anti-sceptical idea. We should ﬁrst admit that colour error theory and eliminativism are endorsed by scientists and this for scientiﬁc reasons. Even so, the kind of liveness that works in the sceptical argument template is truth endorsement on behalf of experts, but all we have in the case of eliminativism and colour error theory is some importantly weaker kind of endorsement that runs, ‘Well, given the state of play in this area, this hypothesis is deﬁnitely one that we need to examine in order to further the ﬁeld’, or, ‘Given the assumptions currently in play, this theory is just about as probably true as any other’. Let’s call a hypothesis robustly live if plenty of experts rate its probable truth more than 50 per cent, say. Then revise the Live Hypothesis and Modesty principles so that they turn on robust liveness only. Now we see where the live sceptic has gone wrong: the eliminativist and colour error hypotheses are not robustly live. Forced to produce a ‘percentage chance at truth’ endorsed by a sizeable pool of experts, we could give those hypotheses no more than a 10 per cent chance at truth. Ten per cent is enough for being live in a certain pragmatic way regarding a certain kind of problem: if no theory looks acceptable, if they all have serious problems, then 10 per cent is enough for liveness. But it isn’t enough for robust liveness. Call this the Robustly Live Solution. Once again, this is a version of the No-Threat Strategy. The proposal does put its ﬁnger on an interesting phenomenon. I wouldn’t be surprised to discover that most philosophers have very few beliefs in substantive philosophical hypotheses. The question is: in the general case when we advocate a philosophical theory, do we believe that it’s true, or do we just hold a more cautious, weaker, attitude to it? Did David Lewis really believe that there is a plurality of

Are the Threats Real?

161

genuinely real worlds? Does Nathan Salmon really believe that anyone who recognizes herself as believing that Twain was famous also recognizes herself as believing that Clemens was famous—even though she would emphatically deny ‘Clemens was famous’? Do Peter Unger, Hartry Field, Larry Hardin, Peter van Inwagen, Tim Williamson, Roy Sorenson, Stephen Stich, and the Churchlands— just to stick with eminent, well-known contemporaries—really believe the anti-commonsensical things they advocate so strongly? Or isn’t it rather that they hold that if such-and-such philosophical claims are true, then those theories are correct? Isn’t it the case that we endorse philosophical theories not as true but only as trueon-certain-background-assumptions? Or perhaps we believe things like ‘This is the best theory we’ve got to work on’. When it comes to philosophical hypotheses of interest, we are agnostics who disagree only on the relative promise of competing hypotheses. So when the ﬁrst-year graduate student knows that my theory is false, she is not really disagreeing with me—because I don’t believe the theory either. Advocating a theory doesn’t mean believing it. Therefore, even its strongest advocates need not believe a philosophically live theory. The Robustly Live Solution looks pretty good to me, but it has some serious problems. First problem: the possible worlds in which eliminativism and colour error theory (as well as some of the other sceptical hypotheses) are judged to have a 50 per cent chance at truth by a large and diverse group of experts (based on several strands of evidence, etc.; see the liveness conditions in Chapter 2) are very close to our actual present intellectual community. They are so close that they may become actual soon. We saw this earlier when we discussed how the philosophy of mind could be permeated (if you like, infected) with the sceptical attitude towards common sense that one ﬁnds in many other central areas of philosophy. Do we want to say that the mere mortals in those communities don’t know that ﬁreengines are red? If so, if we are willing to admit the truth of that sceptical thesis, then what if our society becomes just like theirs? Your grandchildren end up not knowing the colour of your car; neither do they know that you believe that scepticism is false—even though they’ve seen your car a million times and you’ve told them your opinion regarding scepticism on many occasions. To admit that the mere mortals in the possible or future 50 per cent communities don’t know that ﬁre-engines are red or that Moore believed scepticism false is to admit something pretty shocking in my estimation. In fact, colour

162

Are the Threats Real?

error theory is more respected in scientiﬁc circles than it is in philosophy; in my judgement it is robustly live now, in the actual world. I do have a hard time seeing how some particular philosophical hypotheses could become robustly live in an advanced and realistic scientiﬁc community. Regarding those societies in which the nonexistence of dreams, say (from Norman Malcolm), is a robustly live hypothesis among expert scientists who have investigated the matter in an intelligent and epistemically responsible way (to preclude the scientists just waxing poetic about issues they know nothing about, as happens from time to time), my intuitions run out on whether they know, for instance, that people often dream of ﬂoating several feet above the ground. That society is too strange for me to have conﬁdent intuitions about. But the society of cognitive or colour scientists who make eliminativism or colour error theory robustly live is highly realistic. So it seems to me that even if we amend the argument in the way suggested, by requiring the new sceptical hypotheses to be robustly live, this proposal is really a sceptical proposal, not an anti-sceptical one. Second problem: the mere mortals can’t tell the difference between robustly and non-robustly live without a lot of sociological investigation. Everything seems the same to them in each world. This is ﬁne if a certain kind of externalism is true; deadly if it’s false. Third problem: why exactly should the robust/non-robust distinction mark such a radical epistemic difference? Unless we have a detailed answer (I don’t), the proposal is too meagre to endorse. In sum, I don’t see any compelling reason to believe any version of either the Double Standard or the Robustly Live Solution; but neither do I have any conclusive counter-arguments. What may be really bothering the advocates of the Robustly Live and Double Standard Solutions, at least when they admit the scientiﬁc credentials and origin of the liveness of the new sceptical hypotheses, is that there was some really bad reasoning involved in the birth of those hypotheses. It doesn’t really matter whether scientists or philosophers are the guilty parties; it doesn’t matter whether the reasoning is scientiﬁc or infected with philosophy in some sense; and it doesn’t matter what the precise nature of the liveness consists in. Rather, it’s the poor quality of the reasoning for those hypotheses that robs them of their power to constitute signiﬁcant epistemic threats to our second-order and colour beliefs. The people who made the errors might be very bright, but everyone has their fumble. We are not

Are the Threats Real?

163

epistemically responsible to the bad mistakes made by our experts, even if they become fashionable and temporarily institutionalized among experts. So the new sceptical hypotheses aren’t threats. This is the Bad Reasoning Solution. But how exactly is the reasoning bad? It isn’t merely that it led to a false conclusion, even one metaphysically or practically distant (as we saw earlier). It isn’t bad in the sense that only a really thick person would engage in it. It isn’t bad in that it was easily avoidable. In the case of colour error theory, it is remarkably easy for clever folk to embark on a line of reasoning that makes the hypothesis seem probable. It isn’t bad in the sense of being clearly quite weak; once again the expert advocates prove otherwise. It isn’t because the objective support for the live hypotheses is incredibly weak; again, the brute fact is that the evidence is enough for many of our best and brightest to judge them likely or even true. The truth is that the reasoning that generates the hypotheses isn’t bad but difﬁcult. It’s just really hard to get one’s head around the issue of the truth-value of claims like ‘My trousers are blue’, ‘I have a pain in my chest’, and ‘She believes your driving is awful’ when we are exposed to scientiﬁc theories about colour, pain, and belief. When reasoning about something easy, virtually any mistake makes for bad reasoning, but when reasoning about something difﬁcult, sophisticated, and hard to evaluate, it is to be expected that mistakes will ﬂourish. This is what has presumably happened to generate the live but false sceptical hypotheses. We can of course virtually effortlessly reach true beliefs about trouser colour, pain locations, and Mary’s beliefs (remember that we are assuming that the sceptical hypotheses are false). But the live sceptical arguments are targeted at people in a peculiar if widespread epistemic position. The position of the target group comes from a quite general puzzle about scientiﬁc investigation into matters we already have ordinary, pre-scientiﬁc, and highly useful beliefs about. Here is the situation that generates all the live sceptical puzzles. The (nearly actual) scientiﬁc community says that each of the following is a real, live possibility: Beliefs don’t exist. Pain exists only in the head. The table isn’t brown. No one has any character traits.

164

Are the Threats Real?

Sometimes the scientists are right: the earth isn’t ﬂat and motionless with the sun going around it once a day. Other times they aren’t: we have beliefs, pains often exist in ﬁngers, and the table is brown—or so we are assuming in this essay. If you’re no expert or genius but appreciate the liveness of these hypotheses, and you notice the conﬂict between your belief and those hypotheses, but you go on with your ordinary beliefs anyway, then it seems that you have not given the contrary evidence its due; your beliefs are not sufﬁciently warranted to amount to knowledge. That’s the new scepticism generated by scientiﬁc advances. Hume, Reid, and Moore motivate another way to pit common sense against the live sceptic. When discussing miracles, Hume asked us to compare or weigh reasons when faced with a challenge to common sense. Moore is famous for doing the same. I will focus on Moore.52 Moore’s proof of the external world is beautifully simple: (a) here is a hand; (b) here is another hand; thus, (c) there are external objects. Obviously, if he were interested in validity, he could have thrown in the premiss ‘Hands are external objects’. Roughly put, Moore’s position is that he knows the truth of the premisses as well as the argument’s validity and non-question-begging nature; so it’s a proof of the conclusion. He also thinks that he cannot prove (a) or (b). Presumably what he means here is that even though he has epistemic warrant sufﬁcient to make his true beliefs in (a) and (b) knowledge, such warrant does not and could not come from anything like a proof. One might think that one could mount an analogous argument against the live sceptic: here is a belief; here is another belief; thus, there are beliefs. In addition: here is a blue sock; here is another blue sock; thus, there are ordinary coloured physical objects. So we have proofs that there are beliefs and coloured socks. Or even this: I know I believe in ghosts; I know I believe that 2 þ 2 ¼ 4; so there is knowledge of belief. Lastly: I know that this sock is blue; I know that that sock is blue; so there is knowledge of the colour of ordinary physical objects. Now we have ‘proofs’ that even for mere mortals there is knowledge of belief and coloured socks. Can this apparently question-begging argumentative strategy work in any of these cases? 52 I utilize his ‘Proof of an External World’ and ‘Four Forms of Scepticism’, both in his 1962. Greco (2002) has convincingly shown Moore’s debt to Reid.

Are the Threats Real?

165

I will do the following. I will attempt to state brieﬂy and in terms I prefer part of what Moore was up to in his response to the external world sceptic. Then I will indicate why I think Moore’s arguments won’t work against the live sceptic even if they work against the traditional sceptic. Moore (my Moore) ﬁrst tried to show why the external world sceptic is wrong to think that there are no facts that make it true that the average Joe knows that there is an external world. Joe knows this, according to Moore, because of the existence of worldly facts that serve as truth-makers for (a)–(c). a. Joe knows that he has hands. b. Joe knows that hands are external objects. c. Joe can and has expertly deduced that there are external objects from the claims that there are hands and hands are external objects. d. Thus, Joe knows that there is an external world. The Moorean claims that the facts that make (a)–(c) true are sufﬁcient to make (d) true; so (d) is true. Further, the facts that make (a)–(c) true constitute a proof of the proposition in (d) that there is an external world. Now of course the external world sceptic objects that no fact exists to support (a). Indeed, that is her real position: Joe doesn’t know anything substantive about the existing external world—such as the fact that he has hands. This is one reason why Moore’s argument is often seen as question-begging. The sceptic’s main claim isn’t that ‘we can’t prove the existence of the external world’; it’s that ordinary knowledge about the external world—such as the knowledge that one has hands—doesn’t exist. So using ‘Here is a hand’ to prove ‘There is an external world’ does nothing to put a dent in the sceptic’s arguments or position—any more than using ‘Fred is a ghost’ to prove ‘There are ghosts’ puts a dent in the enlightened position that there is no proof that there are ghosts. But Moore was fully aware of all this, and so he gave arguments to back up his claim that we know that we have hands. That is, he backed up his claim that (a) is true. What we have seen thus far is just a fragment of Moore’s reasoning. If he had not given further arguments, then his argument would be question-begging, but even a casual look at his lectures shows that the supporting arguments were there. There may be questionbegging going on in Moore’s line of reasoning, but it isn’t here.

166

Are the Threats Real?

Back to Moore’s sceptical opponent. She agrees that if such a fact underlying (a) existed, then perhaps the combination of the truthmakers for (a)–(c) would sufﬁce to make (d) true. Moore asks the sceptic why she thinks there is no fact that makes (a) true. The sceptic replies by claiming that there are no facts analogous to the ones that would make (d) true that collectively serve as a truth-maker for (a). That is, the sceptic wants (a) to be made true in just the way (d) is made true—with some kind of proof-like structure of facts. Just as (d) is to be made true by the truth-makers for (a)–(c), so (a) is supposed to be made true by some truth-makers for another trio of claims, (a1 )–(a3 ). Moore seems to want to agree with the sceptic to this extent: there is no such set of facts (a1 )–(a3 ) making (a) true (this is Moore’s admission that he has no ‘proof ’ that he has hands even though he has a ‘proof ’ that there is an external world). At the least: Moore hasn’t been able to satisfy himself that there is such a collection of facts (a1 )–(a3 ) for (a). But he then questions why such a set of facts is necessary for Joe’s knowledge that he has hands, as the sceptic is assuming. Why can’t Joe’s knowledge of his hands be constituted by some other combination of facts, a combination that isn’t analogous to the ones making (d) true? The sceptic replies that Moore himself has not come up with any proposals on what the alternative truth-makers could be for knowledge claims. Moore agrees: he has done no such thing. That makes it look as though Moore has no real response to the sceptic. That may just seem obvious: how is one to defend the view that knowledge can be had in some unfamiliar way unless one has a pretty good story regarding what that way is? But no. He then introduces the key Moorean insight: he doesn’t need any such proposal in order to defeat the sceptic! He offers the sceptic (and agnostic regarding the actual existence of knowledge) a test. Ask yourself what is more probable: (3) or the conjunction of (1) and (2)? 1. If Joe knows that he has hands, then there exist some facts of such-and-such a kind (e.g., an (a1 )–(a3 ) proof) backing up that knowledge. 2. There are no such facts. 3. Joe knows that he has hands (this is just claim (a) again). Moore insists that (3) is much more certain than the conjunction of (1) and (2). Sure, the sceptic has presented some interesting reasoning for the conjunction of (1) and (2), but let’s not get too worked up

Are the Threats Real?

167

over the quality of those reasons. Either the sceptic has gone wrong in thinking that there’s no set of ‘proof ’ facts backing up Joe’s knowledge of his hands (i.e., (2) is false), or the sceptic has gone wrong in thinking that such a complex of facts is needed to back up Joe’s knowledge of his hands (i.e., (1) is false). Of course, the Moorean response—call it the Comparison Solution—generalizes to any sceptical argument, as follows. The sceptic endorses (A) and (B) and thereby rejects (C) since (C) is inconsistent with the conjunction of (A) and (B). A. If Joe knows that he has hands, then E is true. B. E is false. C. Joe knows that he has hands. There are many candidates for claim E; here are several classic ones: . . . . .

There are some facts of such-and-such a kind (e.g., an (a1 )–(a3 ) proof) backing up that knowledge. Joe is absolutely certain he has hands. Joe is able to rule out the hypothesis that he’s dreaming. Joe knows he isn’t dreaming. Joe knows he isn’t a BIV.

For any of these candidates Moore thinks that the amount of certainty or plausibility attaching to the conjunction of (A) and (B) is minuscule compared to the certainty or plausibility (C) has. Since (C) is obviously inconsistent with the conjunction of (A) and (B), we must reject that conjunction. For some sceptical arguments Moore thought that (B) was the culprit (e.g., ‘Joe doesn’t know he isn’t dreaming’), but clearly the basic Moorean move is the comparison of (C) with the conjunction of (A) and (B). And then this fact about (C)’s dominance over (A) and (B) is supposed to mean that the sceptic’s argument—which relies on the strength of (A) and (B)—is utterly lacking in strength. Thus, we have the surprising result that we can defeat the sceptic without doing the real work: we can defeat her without showing just what it is that makes it true that Joe knows he has hands or even showing where the sceptic’s argument has gone wrong. We get the anti-sceptical result without doing any of the work! This is Moore’s achievement. But what exactly is Moore’s conclusion? Is it that we know that Joe knows he has hands? Or is it that we might not quite know that about Joe but it’s true anyway? Or is it that epistemologists have ﬁnally

168

Are the Threats Real?

refuted the sceptical arguments—at least the arguments given thus far? Or is it just that Joe has a justiﬁed true belief that he has hands? Or is it that we know that the conjunction of (A) and (B) is false—or just very improbable? Or what? Fortunately, I don’t have to sort that out for our purposes. Return to Moore’s argument: why should we agree with Moore that (C) is much more certain than the conjunction of (A) and (B)? Suppose scientists were able to discover rock-solid evidence for (A) and (B). In that case, Moore would retreat. As I understand it, Moore’s argument for his crucial premiss that (C) is so much better than (A) and (B) is that the support for the conjunction of (A) and (B) is, in fact, tiny compared with that for (C), for any (A)–(C) ever actually offered. He isn’t claiming that we can’t imagine scientists eventually coming up with excellent evidence for some substitution instances of (A) and (B). That may well happen in our future. If it did, that would not mean that (C) is false of course; all it would mean is that we can’t ‘defeat’ the sceptic via the Comparison Solution. But in any case, for any substitution instances of (A) and (B) ever actually offered, (C) has won hands down, so to speak. So we have yet to be given any powerful reason for scepticism. Of course this is not to say that we should ignore the sceptic’s arguments for (A) and (B). Indeed, investigating them might be the most productive thing for epistemologists to do in order to understand knowledge and warrant. But we can start off our investigation with the attitude that there is no good argument for scepticism. Not because its premisses (A) and (B) are so bad! We earn the anti-sceptical attitude based on the fact that the commonsensical (C) is so good compared to (A) and (B). That is, the idea behind this Hume–Reid–Moore Comparison Solution is not that there is something terribly wrong with philosophical reasoning or the sceptical hypotheses—as posited in the previous three antisceptical solutions employing the No-Threat Strategy. At least, it need not have that motivation. Rather, it’s that the beliefs being targeted have an enormous amount of positive warrant, so the legitimate negative warrant revealed by the sceptic’s arguments is overwhelmed. The Moorean need not be dismissive about the power of philosophy; rather, he is impressed with the certainty of parts of common sense. We can think of (C)’s epistemic superiority to the conjunction of (A) and (B) as a superiority of epistemic warrant (Moore said ‘certainty’, not ‘justiﬁcation’ or ‘warrant’). And we can think of the balance of warrants in either of two ways. Suppose that Joe has 8,000

Are the Threats Real?

169

positive units for his belief in his hands, acquired in the usual way. He has 500 negative units due to the sceptic’s arguments. So he has 7,500 units net. But he needs just 1,000 units for knowledge. So he knows he has hands. Of course, no one thinks that warrant comes in units. Still, the notion of some warrant factors outweighing others isn’t silly at all. It is more plausible to hold that some kinds of warrant don’t match up in this simple quantitative manner. Instead—and this is the second way to conceive of the balance of warrants—some negative warrants amount to vetoes of certain kinds of positive warrant. In that case, the sceptic has yet to reveal any negative warrant attaching to Joe’s belief about his hands that vetoes its high-quality positive warrant. So he knows he has hands. It’s interesting to note that the basic Moorean anti-sceptical proposal—the certainty comparison—has nothing essentially to do with scepticism or epistemology or even philosophy. The Moorean comparison argument can be mounted against any argument that concludes with the negation of some claim C of robust common sense and employs some condition E that some anti-commonsensical theorist has argued goes with C. i. If C holds, then E holds. ii. :E. iii. C. The anti-commonsensical theorist argues mightily for (i) and (ii), concluding that (iii) is false. And of course the Moorean points out that despite the perhaps brilliant efforts of the anti-commonsensical theorist, (iii) is much more certain than the conjunction of (i) and (ii); so the anti-commonsensical argument fails. In fact, if we let C be ‘No one knows how many grains of sand there are in the world’, then the Moorean response to someone who thinks she knows how many grains of sand there are, is a sceptical response, not an anti-sceptical response. The Moorean argument can be applied, with various degrees of success, to any of the following anti-commonsensical philosophical theses (each of which negates some commonsensical Cs ). 1. 2. 3. 4.

No one believes anything. (Eliminative materialists.) There are no after-images. (Some physicalists.) 2 þ 2 6¼ 4. (No positive mathematical truth.) There are no people. (Peter Unger.)

170

Are the Threats Real?

5. Thermometers have beliefs. (Certain information-fanatic philosophers.) 6. There are no chairs. (No non-living composite physical objects exist.53) 7. Nothing is coloured, not even tomatoes. (Lots of people, even scientists.) 8. No tomatoes are red; only mental things are red. 9. Stones are not solid objects. (Inspired by Sir Arthur Eddington.) 10. No one has ever had a dream. (Norman Malcolm.) 11. Cats don’t feel any pain when their paws are cut off. (Descartes.) 12. A severed ‘hand’ is not a hand. (Aristotelians.) 13. The world could not have turned out even a bit better than it actually is. (Leibnizians.) 14. It isn’t wrong to torture young children purely for fun. (No moral truths.) 15. Kant didn’t live after Descartes died. Alternatively: Nothing ever happened in the past. (Time doesn’t exist; isn’t ‘real’.) 16. No one has ever done anything because they wanted to do it. (Various reasons.) 17. Rocks have mental characteristics. (Idealists.) 18. There could be two wholly physical objects that during their entire existence occupied the very same space and were composed of the very same particles in the very same manner. (Some contemporary metaphysicians.) 19. Taking one cent from a rich person can make them no longer rich. (Epistemic theory of vagueness.) 20. No one is free to do anything. (No one is free, full stop.) 21. There are no necessary truths. (Some self-styled Quineans.) 22. No one knows anything, or much of anything. (Radical sceptics.) 23. Pain only occurs in the head, so no one has pains in their bodies. (Some scientists and philosophers.) Extending the list is a matter of patience and familiarity with the history of philosophy and science. It is easy to see the power of the Moorean response to various anticommonsensical arguments from philosophy. Indeed, the Comparison Solution is suggested in Hume (and in Reid too). This is not to 53 This area of metaphysics is particularly rife with outrageous theses, ones that in my opinion make radical scepticism downright tame. In fact, almost no one defends common sense in this area, exactly the opposite of the situation in epistemology!

Are the Threats Real?

171

say that there aren’t serious problems with the application of the strategy to even traditional scepticism. The most obvious, perhaps, is the question of what is being compared under the rubrics ‘certainty’ or ‘plausibility’ when we compare (C) with the conjunction of (A) and (B). The Moorean has to cash this out in a way that isn’t question-begging (e.g., ‘The warrant had by the sceptic’s premisses is far less than the warrant held by the negation of the conclusion’) or irrelevant to the debate (e.g., subjective certainty). But in any case, on my accounting, there is no obvious question-begging going on here. For the sake of argument I’ll admit the solution’s truth in its entirety as a response to the traditional sceptic. But it seems to me that even if it works against the traditional sceptic, the solution does not work against the live sceptic.54 The live sceptic ﬁlls in the Moorean template (i)–(iii) with the following (ignoring the complexities of context for the sake of simplicity): .

. . . . . .

. .

If Joe knows that Moore believed that scepticism is false, then E: either the eliminativist hypothesis isn’t live, or Joe isn’t a mere mortal regarding it. :E: but it is live, and Joe is a mere mortal regarding it. Joe knows that Moore believed that scepticism is false. If Joe knows that ﬁre-engines are red, then E: either colour error theory isn’t live, or Joe isn’t a mere mortal regarding it. :E: but it is live, and Joe is a mere mortal regarding it. Joe knows that ﬁre-engines are red. If Joe knows that Mary has a pain in her toe, then E: either the pain-only-in-the-brain theory isn’t live, or Joe isn’t a mere mortal regarding it. :E: but it is live, and Joe is a mere mortal regarding it. Joe knows that Mary has a pain in her toe.

54 For some of the reasons it doesn’t work against the eliminativist sceptic, which I will detail below, in my judgement it doesn’t work against the eliminativist herself. I concede that I might be wrong about that, as I have no convincing arguments going for me. (I won’t bother you with my unconvincing arguments.) But even if I am wrong, the Moorean– Humean refutation of eliminativism or the other hypotheses via the Comparison Solution is simply not common knowledge, in either this world or many others, even among epistemologists, and thus those sceptical hypotheses are free to be live. Thus, even if there is a ‘refutation’ of the eliminativist hypothesis around, it can still be live. This point was made earlier, in Ch. 2. However, for a reason to be given below, the Moorean proposal hardly merits the title ‘solution’ even if it fully succeeds against the new sceptic, the eliminativist, and the colour error theorist.

172

Are the Threats Real?

I will argue for two claims. First, for each of these cases we can’t be conﬁdent that the amount or quality of warrant had by the third member is signiﬁcantly higher than that had by the conjunction of the ﬁrst two. I’m not arguing that the balance of warrant favours the live sceptic; I’m not even saying that the balance doesn’t favour the anti-sceptic. I’m just making the milder point that the comparison reveals no winner, and so the Comparison Solution is useless as an argument meant to defeat the live sceptic’s argument or conclusion. Second, even if I’m wrong, and the Moorean view is entirely correct even as applied to the live sceptic, in an important sense this doesn’t amount to anything. The ﬁrst claim, that the Moorean is wrong, will be defended with six arguments; the second claim, that the Moorean solution is grossly inadequate even if true, will need just one supporting line of argument. I take the ﬁrst claim ﬁrst. First argument for the ﬁrst claim (that Moore is wrong about the comparison). It may seem crashingly obvious that we know we have beliefs, we know that some tables are brown, and we know that there are honest-to-goodness chest pains. That obviousness simply overwhelms the warrant had by the two premisses of the live sceptic. But how much weight does being ‘crashingly obvious’ come to? It once was crashingly obvious that we knew that: 1. 2. 3. 4. 5. 6. 7. 8. 9.

God exists. The sun goes around the still earth once a day. The earth is ﬂat. Many common parts of nature have been designed by a higher power. There can be no precise science of chance or probability. Solid objects are solid through and through. Colours are just run-of-the mill physical properties of objects, with no dependence on us or any other minds. A purely physical object could not think of, say, Alpha Centauri, or God, or beauty, or inﬁnity. There’s no good reason to think that Z is any better than X: there are three identical containers X, Y, and Z; one contains a fortune; the other two are empty; you get to choose one; you choose X; the presenter reveals that Y is empty; you are asked if you’d like to switch your choice from X to Z.

Most if not all of (1)–(9) are false; none of them is crashingly obvious to us any more; so clearly the idea that we know them is even less

Are the Threats Real?

173

plausible; all are currently, in philosophy, science, or mathematics, highly contentious or the object of ridicule; all were, for long periods of time for most people, as crashingly obvious as just about anything. So it seems to me that the property of being crashingly obvious does not correlate as well with high-quality warrant as the Moorean claims. Now one can admit all this but then object that the beliefs targeted by the sceptical hypotheses are crucially different from those in the list above. I don’t think this hair-splitting tactic will succeed. So I would not rate common-sense beliefs nearly as high in positive warrant as the Moorean does. Second argument. Everyone admits that there is something just right about our common-sense knowledge attributions. We rightly have enormous conﬁdence in those claims. (Naturally, much of our conﬁdence in the rightness of common-sense knowledge attributions comes from our conﬁdence in the rightness of the claims one gets when one cuts off the ‘S knows that’ part of the common-sense knowledge attributions.) So we come to think that those attributions are so great because they are attributing beliefs with high-quality warrant. But I suspect that we mistakenly infer from the extreme acceptability of the knowledge attribution-action that the attributed belief must have a very high-quality positive warrant. The action– proposition distinction is not some controversial bit of philosophy; it’s obvious. It’s also obvious that the commonsensical knowledge attributions in question have an enormous degree of acceptability; it’s obvious that common sense is largely a practical matter; it’s obvious that the acceptability of a knowledge attribution is largely determined by the practicality of the attribution-action; and it’s obvious that the attribution-actions in question have an enormous degree of practicality. Thus, it is natural to attribute the extreme acceptability of common sense about knowledge to that of the practicality of common-sense knowledge attribution-actions. If the practicality of attribution-actions were mostly determined by the warrant of the attributed belief, then we would have a decent argument for the high quality of the attributed belief. But I don’t see much reason to accept the antecedent. Even if some of (1)–(9) are known, it seems to me that very few, if any, of us have very good reason to think so. The third argument acts as a buttress to the second. We tend to be impressed with commonsensical propositions because they seem so widely conﬁrmed. That ﬁre-engines are red seems to be conﬁrmed

174

Are the Threats Real?

a million times a day in many different circumstances. But not really. The ﬁre-engine proposition is indeed tested a million times a day, but only in a quite restricted range of circumstances—the ordinary ones in which we see ﬁre-engines on the street. When we observe ordinary physical objects in highly diverse scientiﬁc settings, then we grow puzzled and become attracted to various kinds of error theories with regard to colour. Ordinary experience is like the experiences one has in reading newspapers. One newspaper says that Smith was born in California. So do three other newspapers. Scientiﬁc experience is like the experience of a private investigator hired to ﬁnd out if Smith really was born in California, as he claims he was. Putting lots of conﬁdence in common sense is little more than buying many newspapers that all say the same thing. It’s when we set the newspapers aside and do some in-depth investigation that cracks in common sense appear. That’s when we see the merits of anti-commonsensical theories such as colour error theory; that’s when we uncover evidence that Smith has lied about his childhood. This holds for the mind as well, as Churchland, Churchland, Stich, and Dennett have pointed out. Theorists like Fodor stick with examining newspapers (e.g., folkpsychological explanations); when one does so, eliminativism looks mad. The Churchlands, Dennett, Stich, and others look at the results of the private investigators (e.g., abnormal psychology); here eliminativism no longer looks mad because we are testing in a much more rigorous fashion the claim that we have beliefs. This is not to say that eliminativism, colour error theory, or other anti-commonsensical theories are correct or that we shouldn’t rely on newspapers. But it is to say, with regard to the Moorean Comparison Solution applied to the live sceptic, that the view that the warrant had by the third proposition (in each of the three bullet lists above) greatly outweighs the warrant had by the conjunction of the ﬁrst two propositions is often based on a diet of too few types of tests.55 Fourth argument. The Moorean implicitly assumes, I think, that if we don’t know that ﬁre-engines are red, then our epistemic situation is appallingly bad. This is quite right when it comes to traditional forms of scepticism. The Moorean is right to think that if he gives up 55 These ﬁrst three arguments against the Moorean Comparison Solution to the problem posed by the live sceptic apply to the general form of the Comparison Solution, thereby putting pressure on the Moorean response to traditional sceptical arguments as well as the new ones.

Are the Threats Real?

175

on Joe’s knowledge of the existence of his hands, then he is going to have to give up much more knowledge than that—and there will be no way to obtain that knowledge, as the resulting scepticism is permanent, blanket, and a mark of epistemic failure. The Moorean infers the high quality of his ﬁre-engine belief partly from the extreme consequences of that belief not amounting to knowledge. But none of those consequences hold for the new scepticisms. As I argued in Chapter 6, the mere mortal who loses her knowledge that ﬁre-engines are red is in at least as good an epistemic position as she was previously, when she had that knowledge. Neither is the scepticism permanent or blanket. The live sceptical arguments are a whole new kettle of ﬁsh. Fifth argument. The Moorean critic of the live sceptical arguments also assumes that the only backing for the conjunction of the ﬁrst two claims will come from abstract philosophical reasoning. Due to the weak credentials of such reasoning when compared to that of common sense, the third proposition wins the warrant comparison. But this isn’t the case for the live sceptical arguments: they are based on commonsensical epistemic principles utilizing scientiﬁcally live hypotheses. We have already seen this in my criticisms of the previous three proposals for solving the live sceptical puzzles. Sixth argument. Pretend that your belief that ﬁre-engines are red (or your belief that you believe that ﬁre-engines are red) really has each of those 8,000 warrant units, and you need just 1,000 units for knowledge. As noted earlier, some sources of negative warrant produce a veto of any positive warrant for the belief in question. It doesn’t matter whether warrant is quantitative or not: if H is scientiﬁcally live, S is a mere mortal, and H obviously entails :P, then H vetoes the warrant S has for her belief in P. Now suppose all my arguments fail and the Moorean proposal is correct and shows that live sceptical arguments fail. Let’s even assume that the Moorean is right that our ﬁre-engine beliefs don’t just barely qualify as warranted but have oodles of warrant (or high-quality warrant), so we can safely dodge the persistent sceptic who complains that although ‘Joe knows that ﬁre-engines are red’ is true, the knowledge in question is so impoverished to be somewhat embarrassing.56 Oddly enough, I think there is an important sense in which 56 This is not a small concession. Certainly the Moorean response does nothing to encourage the idea that knowledge isn’t impoverished.

176

Are the Threats Real?

the success of the Moorean Solution makes no difference to philosophical practice. The success of the Moorean is without consequence. For even if the Moorean proposal is correct, it isn’t even remotely close to being a solution to the puzzles created by the live sceptical arguments, where the puzzles are in the form of mutually inconsistent sets of individually plausible claims. For the Moorean response does not meet the four constraints on an acceptable solution to a sceptical puzzle: identify the false claims in the set, explain why they are false, articulate the truths they are mistaken for (if any), and explain how it is that we made those mistakes. In fact, the Moorean response fails to meet any of those conditions. I suspect that this is one of the main reasons almost everyone hates the Moorean response, even when applied to traditional sceptical puzzles and stripped of any question-begging. It’s not that Moore is wrong about the certainty comparison, although he might be. Rather, even if he is right about that comparison, the problem is that there is so much work left to do in solving the puzzle that we can hardly be said to have made but the most preliminary start. Even today Moorean rhetoric sometimes suggests that once we face the comparison challenge and give the Moorean response, signiﬁcant progress has been made (e.g., Lycan 2001; Greco 2002). But the sceptical problem is nowhere near solved with Moore; at best (that is, if Moore is right about the comparison challenge) just one move among many has been ruled out. The question raised by scepticism isn’t, for me anyway, ‘Is scepticism true?’ or ‘Does this particular argument establish scepticism?’ Instead, the question is, ‘How do we solve the sceptical puzzles these arguments generate?’ The other questions are good ones, to be sure, but one hasn’t got very far on the problem unless one answers the last question. Let’s keep our eye on the ball. To make the matter vivid, imagine that we had an intriguing and apparently valid argument for radical scepticism that had 835 premisses. The corresponding sceptical puzzle can be represented as the 837-member inconsistent set containing those 835 claims, the negation of the conclusion, and the claim that the argument is valid. An adequate solution would have to identify the falsehoods in the set, tell us what is wrong with them, articulate the truths that we are mistaking them for (or explain why there isn’t such a truth, as the case may be), and explain why we made those mistakes. All the Moorean gets us is this: this claim over here, this one out of the 837, the

Are the Threats Real?

177

negation of the sceptic’s conclusion, is true. Gee, thanks, but this hardly helps. Even if God announced that the Moorean is right, this would hardly help us solve the puzzle. The Moorean might grant this objection but protest that if his claims about our responses to his challenge are correct, then at least radical scepticism has been defeated, even though the sceptical puzzles remain. But this is clearly false: this will hold only if the Moorean comparison works for every sceptical argument, and surely modesty prevents us from announcing ahead of time that the Moorean certainty comparison will work for ever against the sceptic. (I don’t think Moore would disagree.) I personally thought that the classic sceptical arguments were the best we would ever see, but now I think the new ones are better; and who’s to say that superior ones won’t be discovered? So even this quite weak Moorean conclusion is unjustiﬁed. The Moorean might retreat further and insist that at least all the extant sceptical arguments have been disarmed, provided we accept his certainty judgement for every sceptical argument ever given. It is true that if we accepted his comparison judgement, we would no longer be advocates of those sceptical arguments, but we still would have no idea why any of them failed. We certainly wouldn’t be aware of the mistake that any instance of sceptical reasoning is based on, for we haven’t a clue what to make of the premisses. Surely this wouldn’t count as anything resembling a solution. So it seems to me that the Moorean certainty comparison, no matter how construed, does not buy us much. We still have to continue exactly as if no one had offered the Moorean proposal: we still need to examine the live sceptical arguments in detail to ﬁnd any faults. In a real sense, we’re none the wiser. Scepticism has predominantly been understood as presenting us with a problem or puzzle, and the Moorean response just doesn’t give us anything but the slimmest help in solving it. We all know where the real work lies, and the Moorean Comparison Solution does at best only the tiniest bit of work.57

57 This is not to suggest that Moore himself did nothing but offer the Comparison Solution to sceptical arguments. He did probe the arguments themselves as well, intelligently searching for problems. Moore is one of my philosophical heroes, and I don’t want to leave the impression that I think ill of his philosophical talents.

178

Are the Threats Real?

C The Population of the Sceptical Snare Perhaps you’re inclined to admit that the poor fool on the street with just enough education to know about the liveness of the eliminativist and other hypotheses but not enough training or smarts or good sense to see through them, falls under their epistemic threat and loses her knowledge. The person on the knife-edge between ignorance and knowledge falls into the sceptical trap. But that’s the special case: someone with just barely enough knowledge to be confused. And of course there are the (how should we put it?) professionally loud philosophers who vigorously champion the live sceptical hypotheses in their mini-research programmes generating gallons of journal ink. When it comes to their attitudes regarding the colour of their diningroom tables, they are such an odd bunch, compared to anyone else, that it doesn’t really matter whether they are victims of some kind of restricted scepticism as a result of their peculiar views. The vast silent majority of mature and sensible philosophers are not subject to the live sceptical arguments. We may not be geniuses, but we all recognize how meagre the support is for any of the sceptical hypotheses. Yes, we do know that ﬁre-engines are red. So even if the live sceptical arguments are sound, their scope is pretty limited. People on the knife-edge can be affected, and perhaps certain excited philosophers are affected as well; we will grant that the arguments have those consequences. But, contrary to what has been suggested in the chapters above, very few philosophers or scientists fall into the live sceptical traps. So this objection to live scepticism is a use of the No-Threat Strategy: the new sceptical hypotheses are not really live in any epistemically threatening way for most of us. I have three things to say in response to the Limited Scope Solution. First point. Even if the solution is perfectly correct, we have the bizarre result that many people, mainly students, are robbed of enormous portions of common knowledge. If you’re a teacher, then you’re writing letters of recommendation for people who don’t know that ﬁre-engines are red or that G. E. Moore believed that scepticism is false. This is just plain shocking. So this solution strikes me as a sceptical proposal, not an anti-sceptical one. If the proposal is altered to contain the claim that the people on the knife-edge do not fall into the sceptical snare, then this is a mere undefended rejection of the sceptical arguments.

Are the Threats Real?

179

Second point. The NEH Summer Institute on Consciousness and Intentionality was held at the University of California, Santa Cruz, from 24 June to 2 August 2002. There was a questionnaire distributed among the participants regarding views on consciousness and intentionality.58 On virtually every point there was nothing even approaching a silent majority of sensible mature philosophers—and these were experts on the conference topics. One can make jokes here (sensible? silent ?), but I won’t. My only point is that the ‘silent majority’ of agreeable philosophers who (a) hold that each of the sceptical hypotheses is false and (b) can see through the meagre support for those hypotheses is actually quite tiny—because no majority ﬁts the ﬁrst condition, (a). So it seems to me that the proposal makes an empirical blunder. Third point. Suppose that you, dear reader, can explain why all the new sceptical hypotheses are false. You have the ability to rule out all these claims: that the table isn’t brown; that pain occurs only above the shoulders; that no one has any beliefs. You have that ability now; you don’t have to struggle for months or years to hit on the reasons those claims are false. You can rule them out. You are an immortal when it comes to those issues. The reasoning in favour of those claims might not be bad, exactly, but it’s certainly not strong enough to sabotage your knowledge. I personally don’t believe any of that. Or: if that’s true for you, then you are an exceptional philosopher. That is, I reject the second condition (b) attributed to the majority: that they are able to see through the meagre support for those hypotheses. I’m not a bad philosopher, yet I can’t rule out any of those claims; more to the point, I’m not sure what to think of the reasons in their favour. I have things to say about them, in particular about the alleged pathetic support they enjoy, but any realistic assessment shows I’m no immortal with regard to any of them. So even if there is a silent majority of philosophers who hold that all the live sceptical hypotheses are just plain false, they do not, with few exceptions, have the ability to see through the allegedly meagre support for those hypotheses. This is just my opinion of course, but one I hope you share. Why is it that philosophers are so reluctant to take those sceptical hypotheses seriously, as I confess a good many are? I suspect the reasons are purely accidental and have more to do with fashion than 58

See .

180

Are the Threats Real?

insight. Many contemporary philosophers are incredibly suspicious of the ability of philosophical argument to overthrow common sense—or, rather, we are exceedingly suspicious of our colleagues’ ability to do so. We don’t want to look foolish in falling for soand-so’s bad reasons for her anti-commonsensical theory. The worst nightmare of a contemporary philosopher is to discover that her silly theory goes ﬂat against established science and even the scientists could pick apart the theory. So, when the scientist says, ‘Why do we ﬁnancially support you navel-gazing philosophers anyway?’, she isn’t being a philistine; instead, you’re the fool. It also gives us a bad taste in our mouths when we contemplate how foolish we will look in the face of history. When we are presented with the oddities of the scientiﬁc facts behind colour, pain, and, perhaps, belief, many of us are loath to take seriously the idea that our common judgements such as ‘Fire-engines are red’ are false. ‘No,’ we say, ‘People aren’t speaking falsehoods all the time when they talk about red ﬁreengines! Get a grip on yourself! Instead, the right view is that the microphysical facts that make those colour judgements true are just odd or unpredictable from the pre-scientiﬁc stance.’ That may be right, it’s the explanation many would bet on, but I for one can’t adequately defend it. Part of the way forward is to give a revealing picture of the truth conditions for the common-sense claims, but this is hard work. As I mentioned earlier, to adopt as a constraint the truth of common sense, which many linguists and philosophers of language do when construing truth conditions, is a contentious approach in my judgement. The proper constraint is practicality; truth is another matter. D The Ambiguity of ‘Scepticism’ Perhaps our ‘autonomic sensory systems’ can rule out some simple propositions regarding physical objects, e.g., the proposition that the table isn’t brown. Those cognitive systems carry certain pieces of information that in some sense count as primitive and warranted beliefs—even knowledge. Perhaps one part of your mind has ruled out the crazy ‘The table is not brown’ hypothesis, even though another part has not and has some (testimonial and expert) warrant for even believing the crazy hypothesis. Your automatic, primitive, biological, epistemic system has ruled out the sceptical hypothesis;

Are the Threats Real?

181

your highly theoretical, evolutionarily recent, language-infected ‘opinion generator’ has not ruled it out. So the questions ‘Has she ruled it out? Is that possibility eliminated for her?’ are rejected as ambiguous and countered with a request for disambiguation: ‘Which part of her do you want me to consider?’59 Your biological system knows that the table is brown even though your ‘higher’ faculty does not. The new scepticisms are true for our opinion generators but not for our automatic belief generators. So are the new scepticisms true? That question is ambiguous: are you talking about something quasibiological (then, no; the new scepticisms are false) or something more advanced (then, yes; the new scepticisms are true)?60 This sounds like a convenient, wishy-washy response. ‘Live scepticism is kinda true and kinda false!’ It seems to be saying that when it comes to ‘opinions’, the live sceptical arguments go through, but when it comes to mere ‘beliefs’, the arguments fail for some reason or other. I assume the idea is that the twenty-one factors don’t threaten the ‘beliefs’ because the latter aren’t responsible to epistemic community matters. It is hard to evaluate this idea in the absence of details regarding these cognitive systems. Suppose you decide, after mulling it over, that the live sceptical hypotheses are all wrong, and you renew your convictions that ﬁre-engines are red, that you believe that baseball is superior to cricket, etc. According to the proposed solution, do you believe twice over that ﬁre-engines are red? That is, do you have two token beliefs—one biological and the other highercognitive—where just one of them amounts to knowledge? If so, then live scepticism is true for higher-cognitive beliefs: they don’t amount to knowledge. Furthermore, does this other cognitive state token, the primitive one, really count as a belief ? Or isn’t it just some rather primitive, automatically generated cognitive state that feeds information or what not into our opinion-producing apparatus? Does the notion of epistemic warrant even apply to it? Should we get worked up over it? Do any of these primitive cognitive states have what it takes to be second-order beliefs—which are of course the target of the eliminativist sceptical argument? On the other hand, suppose you 59 For some discussion of this belief–opinion distinction see Daniel Dennett, ‘How to Change Your Mind’, in his 1978. I don’t want to suggest that Dennett would endorse any of what is written here. 60 This is not a linguistic ambiguity. It is saying that what we’ve learned about cognition demands that we withdraw our old question ‘Does she know?’ and replace it with two questions.

182

Are the Threats Real?

have just one token belief. So the token has both the primitive belief type and the fancy opinion type. Then the Ambiguity Solution must be that that belief token amounts to knowledge when considered one way but not when considered another way. I don’t know what that could mean, or why it isn’t a sceptical solution. E Testability and Indeterminacy The live sceptical hypotheses could easily arise in a wholly scientiﬁc context, but one might suspect that the theories associated with those hypotheses are immune to veriﬁcation or falsiﬁcation. There is simply no way to establish their truth-value. We can establish the sex of the chicken, but we cannot do something similar with the truth-value of eliminativism. Arguments for and against eliminativism will be endless, like those for dualism versus physicalism, compatibilism versus incompatibilism, theism versus atheism, nominalism versus universals realism, moral realism versus moral error theory, and other everlasting debates. There is simply no way to establish the truth-value of the theory; no matter what happens, nothing conclusive will ever show up. That’s the nature of philosophy. I am assuming that it turns out that ﬁre-engines are indeed red, we believe it, and we have tons of positive warrant for believing it. Sure, there’s the liveness of colour error theory, but given that this theory will never go away or be established, how much of a threat can it really pose? Isn’t it idling in some relevant way? This is another instance of the No-Threat Strategy: the idea is that because the hypotheses can never be settled, they can’t present a real epistemic threat to our beliefs. But this is an odd position: it’s supposed to be as easy as pie to establish that the tomato is red, but we can’t establish the falsehood of a theory which says that the tomato isn’t red? Now we have to give up closure: you know that the tomato is red, you know that if the tomato is red, then colour error theory is false (one straightforwardly entails the other), so you should know that colour error theory is false—especially since you’re reading this and following the reasoning ﬁne. Doesn’t that mean that you have settled the truth-value of colour error theory? After all, you know it’s false— we just went through the reasoning together! The attributor contextualist has a way out of this rejection of closure. She could hold that in an ordinary, real-life context,

Are the Threats Real?

183

‘S knows that the tomato is red’, ‘S knows that if the tomato is red, then colour error theory is false’, and ‘S knows that colour error theory is false’ are all correct. But in certain strange contexts in which we are taking external world scepticism or colour error theory seriously, ‘S knows that the tomato is red’ is false, ‘S knows that if the tomato is red, then colour error theory is false’ is true, and ‘S knows that colour error theory is false’ is false. In even stranger contexts, in which universal scepticism is being taken seriously, all three knowledge attributions are false. The point is that in no context do we reject closure by accepting the ﬁrst two attributions while rejecting the third. Her concessive response to the live sceptical arguments will be this: in certain strange contexts in which we are taking the live sceptical hypotheses seriously, ‘S knows the tomato is red’ is false, ‘S knows that if the tomato is red, then colour error theory is false’ is true, and ‘S knows that colour error theory is false’ is false. If you don’t buy attributor contextualism and don’t want to reject closure, then this anti-sceptical solution is probably not for you. But I want to question the idea that we cannot rule out eliminativism in a conventional, scientiﬁc-philosophical way. I don’t see why we can’t do so. We could either show that the only theories of meaning and reference that have any chance of success entail that belief sentences do not need belief particulars as truth-makers, or show that the only theories of belief that have a chance of success are realist about belief without being realist about belief particulars (so either way ‘S believes P’ is true without belief particulars). More generally, we could show that the only viable theories in this area entail that belief sentences do not have any of the truth conditions that eliminativists claim rule out the existence of beliefs. Or, alternatively, we could show that the robust belief particulars really are in the brain (so ‘S believes P’ is true because its truth conditions call out for belief particulars, and the latter exist). None of this is easy, which is why mere mortals do not seem to have ruled out eliminativism yet, but I don’t see why it could not be done. Further, this is an empirical matter, testing various theories of belief, belief ascription, and cognitive processes in the brain. (Just because philosophers who know no linguistics or psychology work productively on the project doesn’t mean that it’s not empirical.) One might object that even if all this happens as I just described, eliminativism will not be epistemically buried; it will just be temporarily, if justiﬁably, abandoned. It has not been ruled out really; it has

184

Are the Threats Real?

merely been beaten up enough that people no longer pay any attention to it. Two replies are in order. First, suppose this is right: eliminativism will never be ruled out. Then I’m convinced that virtually no interesting scientiﬁc hypotheses gets ruled out either, for all such hypotheses are in the same epistemic boat. Yet I think that the Testability Solution is not very plausible for non-philosophical and upstanding scientiﬁc hypotheses. (Think of the dinosaur example.) Second, if it is granted that many scientiﬁc hypotheses do get ruled out, even when it was very hard to do so, then there is no good reason to think that eliminativism can’t be ruled out as well. (Even if in some interesting sense eliminativism, unlike more straightforward scientiﬁc hypotheses, cannot be ruled out, this does not mean that it can’t cease being live. If our best and brightest minds destroy all the arguments for eliminativism, and several new considerations point very favourably in the direction of realist theories, then the eliminativist view loses its status as a signiﬁcant threat. Knowledge would then be free to ﬂow. The lesson: if you want to regain your knowledge, kill the eliminativists, and burn their writings.) There might be another thought lurking around the Testability Solution. It is the idea that sentences like ‘The table is red’ are indeterminate, so obviously their truth-value cannot be tested. Nothing in our linguistic practices or minds or relationship to the environment sufﬁces to give those sentences truth-values. They have degrees of acceptability, but they aren’t determinate enough, in some way, to rise to the level of being either true or false. Given the bizarre facts we’ve learned from physics about the atomic and subatomic worlds, and the facts we’ve learned from colour science, it just isn’t ﬁxed whether humdrum claims like ‘The table is red’ are true. The facts are so unlike anything we could have imagined that it isn’t clear whether they make our sentences true or false. This is a difﬁcult topic, about which I will say little. But consider the outcome: if ‘The table is red’ comes out neither true nor false, then presumably ‘I know that the table is red’ isn’t true, as knowledge requires truth. This is enough for scepticism in my judgement. The same result holds if ‘The table is red’ is false: clearly, if it’s false, then we don’t know it’s true. And in the essay I have been assuming that ‘The table is red’ is true and arguing that ‘I know the table is red’ is false. Thus, the live sceptic has ‘I know that the table is red’ false no matter what the status of ‘The table is red’ turns out to be: true, false, or indeterminate.

12

Defeating the Live Sceptic Head-On And thus a ready mode is provided, by which whoever is on the strongest side may dogmatize at his ease, and instead of proving his propositions, may rail at all who deny them, as bereft of ‘the vision and the faculty divine’, or blinded to its plainest revelations by a corrupt heart. John Stuart Mill, Coleridge One of the most constant characteristics of beliefs is their intolerance. The stronger the belief, the greater its intolerance. Men dominated by a certitude cannot tolerate those who do not accept it. Gustave Le Bon, Opinions And Beliefs

The anti-sceptic can try to argue that regardless of the strength and usefulness of the live sceptic’s sword (the twenty-one factors), our Subject factors such as evidence and reliability are sufﬁciently strong to defeat it head-on, so to speak. So factors (22)–(27) don’t weaken our Subject factors enough to render them inadequate to defeat the live sceptical hypotheses. This is the Defeated-Threat Strategy, which comes in three varieties. The Safety–Sensitivity Solution is an improvement of the Metaphysical Distance Solution of the previous chapter in being a more sophisticated attempt to solve the puzzle by claiming that worlds in which we falsely believe, e.g., that ﬁre-engines are red, are too metaphysically distant (in various ways) to sabotage our knowledge. The Extreme Externalism Solution says, roughly, that the reliability of our beliefs is enough of a Subject factor to neutralize the sceptical hypotheses. The Tenacity Solution insists that knowledge is not the kind of state that can easily be unseated; thus, since we once knew that ﬁre-engines are red, that knowledge stays with us even after we become mere mortals with regard to colour error theory.

186

Defeating the Live Sceptic Head-On

A Safety and Sensitivity Related to the notion of metaphysical distance from actuality are the notions of safety and sensitivity. The former notion is associated with Nozick (1981), Dretske (1970, 1981), and DeRose (1995); the latter with Sosa (1999, 2000) and others. Given the importance of these notions, it is perhaps obligatory to answer the question whether our mere mortal beliefs threatened by the live error theories are either safe or sensitive. First I explain the terminology; then I answer the question and assess its relevance to the live sceptical puzzles. These two notions are difﬁcult to understand, so my ﬁrst approach will be rough-and-ready, and my second will be done in my own way.61 First safety. You believe that black holes exist. Now imagine us taking your belief as ﬁxed and changing the world in various ways. To make you look bad, we want to change the world just enough that you continue to believe in black holes even though we’ve counterfactually eliminated them from the universe. Clearly, your belief is pretty good if we can’t do this easily. That’s roughly what a safe belief is. Your belief is safe if you hang on to it tenaciously but give it up once we make it false. In a good sense, your belief is discriminating: you can discriminate worldly changes from actuality that don’t falsify your belief from worldly changes that do; you hang on to your belief through the ﬁrst set of worldly changes, but not the second. You don’t pay attention to changes in the world until we change things so that there are no black holes: when that happens, you prick up your ears and give up your belief. Here is another way to introduce safety. It’s easy to imagine lots of worlds very similar to ours in which you have the black-holes belief and it’s true. But it’s harder to imagine worlds very similar to ours in which you have that belief while it’s false; those worlds are less realistic. That is, it’s harder to design a world, starting with the actual one as a template, in which you have a false black-holes belief than it is to design one in which you have a true black-holes belief. The truebelief worlds are much closer than the false-belief worlds. In fact, in each of the 100 most realistic worlds in which you believe in black 61 I think this difﬁculty holds mostly for Sosa’s notion of safety (Sosa 1999, 2000). His presentation in terms of subjunctive conditionals invites criticisms (e.g., DeRose, forthcoming) that I suspect do not diminish the notion that Sosa has in mind, although they are fair criticisms of some statements made by Sosa. I will elaborate the notion of safety in my own way, so as to neutralize DeRose’s objections.

Defeating the Live Sceptic Head-On

187

holes, there are black holes; the false belief worlds are farther out from actuality than those 100. Now for sensitivity. Once again, you believe in black holes. It’s not too difﬁcult to think of worlds in which there are no black holes. But how easy is it to design a world, starting with the actual one as a template, in which there are no black holes but you believe there are? If it’s hard to do this, then your belief is doing well. More exactly: if it’s signiﬁcantly harder to design a black-holeless world, again starting with actuality, in which you believe in black holes, than a blackholeness one in which you don’t believe in black holes, then your belief is sensitive. The main point is that it’s harder to make one in which the black-holes belief is false but you believe it anyway, than it is to make one in which the black-holes belief is false and you don’t believe it; that’s sensitivity. The :P-plus-not-believed-that-P worlds are much closer to actuality than the :P-but-believed-that-P worlds. In fact, in each of the 100 most realistic worlds in which there are no black holes, you don’t believe there are black holes; you’re no fool. In order to make these notions a bit more concrete, stick with the fantasy that we can number these possible worlds in manageable terms. Consider the set of possible worlds in which P (‘There are black holes’) is false. Pretend that there are just 1,000 of these. Now narrow the set so that it contains the 100, say, :P worlds that are closest to the actual world. We are conﬁning our attention to the :P worlds that are as much like the actual world as possible. So we have the set of what we will call the relevantly close :P worlds. S’s belief that P is sensitive if and only if in each of those 100 :P worlds S refrains from believing P. Thus, in all the relevantly close worlds in which P is false, S no longer believes P. Suppose worlds WSE1 ---WSE100 (‘SE’ for ‘sensitivity’) are the ones that are relevantly close and in which P is false. Then S’s belief that P is sensitive just in case S doesn’t believe P in any of WSE1 ---WSE100 . So in an intuitive sense it isn’t easy to get S to falsely believe P. This is not to say that WSE1 ---WSE100 are worlds close to actuality. All we are saying is that among the :P worlds these are the most realistic, the closest to actuality in the sense described above. Now let’s go back to safety. Consider the set of possible worlds in which S believes that P. Pretend that there are just 1,000 of these. Now narrow the set so that it contains the 100 S-believes-P worlds that are closest to actuality. We are conﬁning our attention to the S-believes-P worlds that are as much like the actual world as possible. So we have

188

Defeating the Live Sceptic Head-On

the set of relevantly close S-believes-P worlds. S’s belief that P is safe if and only if in each of those 100 S-believes-P worlds P is true. In all the relevantly close worlds in which S believes P, P is true. Suppose worlds WSA1 ---WSA100 (‘SA’ for ‘safety’) are the ones that are relevantly close and in which S believes P. Then S’s belief that P is safe just in case P is true in WSA1 ---WSA100 . So, as with sensitivity, it isn’t easy to get S to falsely believe P—but in a sense different from sensitivity. Thus, we have the following two rough deﬁnitions. Safety: In the 100 most realistic worlds in which S believes P, P. Sensitivity: In the 100 most realistic worlds in which :P, S doesn’t believe P. Let’s abbreviate these deﬁnitions with something a little more manageable (even if they’re a bit untoward). Safety: If S were to believe P, P would be true. Sensitivity: If P were false, then S wouldn’t believe it. To put this yet another way, S’s belief P is unsafe if and only if P is false in at least one of WSA1 ---WSA100 , the closest worlds in which S believes P. S’s belief P is insensitive if and only if S believes P in at least one of WSE1 ---WSE100 , the closest worlds in which P is false. Thus, in looking for worlds that make S’s belief P unsafe or insensitive, we’re looking for the same thing in the two cases: a world in which S believes P and P is false. But in the case of sensitivity we are looking among the closest worlds in which P is false; in the case of safety we are looking among the closest worlds in which S believes P. One interesting difference between safety and sensitivity comes from focusing on your belief Q that you are not a BIV. Then the relevant S-believes-Q worlds, i.e., the WSA1 ---WSA100 worlds we examine when testing safety, are very close to actuality: you believe you’re not a BIV in those. But the relevant :P worlds, i.e., the WSE1 ---WSE100 worlds we examine when testing sensitivity, are very distant: you are a BIV in those. Thus, to see if your belief that you’re not a BIV is safe, we must look at just the very close worlds in which you believe you’re not a BIV. On the other hand, to see if your belief that you’re not a BIV is sensitive, we must look at just the extremely distant worlds in which you are a BIV. That’s a big difference: when testing the beliefs that the classic sceptical hypotheses are false, safety demands that we look at very close worlds, and sensitivity demands that we look at very distant worlds. Q ends up safe but insensitive. This difference

Defeating the Live Sceptic Head-On

189

doesn’t show up for all Qs, however. If Q is the proposition that I ate breakfast, then {WSA1 ---WSA100 } and {WSE1 ---WSE100 } may be virtually identical in distance from actuality, depending on S, my relationship with S, and my breakfast-eating habits. The above characterizations of sensitivity and safety are ﬁrst approximations. The most obvious question regarding these notions is how we are to judge distance between possible worlds, the notion that some worlds are more realistic—or closer—than others to the actual world. That is, what are the 100 ‘most realistic’ or ‘closest’ worlds? But I want to pursue a different point, one related to the scope of beliefs to which the notions of safety and sensitivity are intended to apply. Return to the (ﬁctional) time when it was up for grabs whether water contained hydrogen. Many scientists thought it did, many thought it didn’t, and most believed it was too soon to make up one’s mind, as the evidence was inconclusive. In such a situation pretend that I believe that water contains hydrogen because my father told me and I’m aware that he’s a high school chemistry teacher. So I have a true belief about the composition of water. But I don’t know it, as my father is certainly no expert and didn’t get his opinion from some immortal. I have come to believe, but not know, that water contains hydrogen, even though my belief is both metaphysically safe and sensitive. If in some possible world I believe that water contains hydrogen, then that belief must be true in that world because I couldn’t (metaphysically) even exist without it being true: in a world in which water doesn’t contain hydrogen I certainly don’t exist! So it’s safe. For the same reason, if the belief were false, I wouldn’t believe it—as I would not exist. So it’s sensitive. Clearly, these facts about the metaphysical safety and sensitivity of my watercontains-hydrogen belief tell us nothing regarding the epistemic status of my belief. The intuitive notions of sensitivity and safety, when given a metaphysical construal, aren’t supposed to apply to beliefs like this one, beliefs in necessary or otherwise fundamental truths. When testing to see whether S’s belief that P is sensitive, for instance, what we’re really supposed to imagine is a world in which P is false but S exists, is a believer, and comes to consider whether P is true in something like the way she actually came to consider P (roughly put). Obviously this isn’t going to apply to Ps such that S could not exist if P were false. And it isn’t only necessarily true Ps that engender this problem. A person in the fourteenth century could believe but not know the contingent truth that the earth was at that

190

Defeating the Live Sceptic Head-On

point about three billion years old. Even so, this belief content is metaphysically sensitive and safe, at least according to my metaphysical intuitions. (I omit the arguments.) Other examples present the same problem. Sosa’s use of safety will introduce my reason for thinking that appeals to safety and sensitivity notions won’t generate a refutation of live scepticism. Sosa gives two more conditions for being safe (1999: 149, 153). Your belief in black holes was acquired at some point through some process; call that process-type the ‘delivery process’. The delivery process delivers to you a proposition, the one you believe. What is required for knowledge isn’t just that the belief be safe, as described previously, but that the delivery process be safe as well. Here’s one way to spell this out (I can’t say this is Sosa’s way). Consider all the worlds in which delivery process C delivers some belief to someone. Pretend that there are just 1,000 of these C worlds. Now narrow the set so that it contains the 100 C worlds that are closest to actuality. We are conﬁning our attention to the C worlds that are as much like the actual world as possible. So we have the set of relevantly close C worlds. A delivery process C is safe if and only if in each of those 100 worlds the delivered proposition is true. In all the relevantly close worlds in which C delivers something, the thing delivered, the proposition, is true (or 92 per cent, say, of them are true). Suppose worlds WC1 ---WC100 are the ones that are relevantly close and in which C has delivered some proposition. Then C is safe just in case the delivered propositions are true in WC1 ---WC100 . So it isn’t easy to get C to deliver a false proposition; that’s why C is safe. We could construct a sensitivity condition for delivery processes as well: very roughly, in the 100 closest worlds in which P is false, C doesn’t deliver it. Sosa’s third condition on safety is that the ultimate source of the delivery process must be intellectually virtuous, i.e., ‘in a reliable or trustworthy way a source of truth’ (1999: 153). I’m not going to attempt to tease apart the meaning here, as it doesn’t matter for my purposes in the succeeding paragraphs. Perhaps Sosa is right to demand the safety of not just the belief content but the belief-producing process; and perhaps we should demand the sensitivity of the belief-producing process as well. So we could have the belief content and belief-forming process safe and sensitive. But even that isn’t sufﬁcient: we must go even further to the belief-sustaining process. The main problem here is simple. The

Defeating the Live Sceptic Head-On

191

acquisition of a true belief may be as upstanding as you please, in any of several ways of spelling out sensitivity or safety, but if the belief is sustained in the face of new, contrary considerations that are recognized as contrary and well respected (but are ultimately misleading), then the person no longer knows P unless someone neutralizes that contrary hypothesis (and the neutralization status somehow applies to that person). We need to have safety or sensitivity apply not only to the belief and the process by which it was formed, but also to the process or conditions under which it is sustained. Consider once again the one-off cases in Chapter 6. You see your colleague’s blue socks in the usual excellent viewing conditions. He insists that he is wearing some of his wife’s trick socks, which are really green. You stubbornly persist in your belief that the socks are blue. It’s a true belief reliably produced in the entirely ordinary way, the belief is as sensitive and safe as you please, the delivery process by which you acquired it is also as safe and sensitive as you please, but the belief doesn’t amount to knowledge after your encounter with the weirdsocks story—especially if we have his wife, her co-workers, and lots of other experts come in with other allegedly trick socks, other fancy objects that are proved to be illusory in ways similar to the way the socks are (mistakenly) said to be illusory, etc. Many other examples, even more plausible ones, could be described to make this point (e.g., we can alter the Chapter 6 case about George knowing that Senator Smith is in Bigtown). Thus, safety and sensitivity advocates need a safety or sensitivity condition that applies to the belief-sustaining process, or its fundamental source. I don’t know how exactly the conditions should go. Perhaps a belief-sustaining process D is safe if and only if in each of the closest 100 worlds in which D is sustaining propositions, 93.6 per cent of the sustained propositions are true. If S’s belief that P is being sustained via D, then D is sensitive just in case in the closest 100 worlds in which P is false, D doesn’t sustain it. In any case, surely your sticking with your blue-socks belief in the described circumstances is neither safe nor sensitive in any intuitive sense. Sticking with your belief P in the face of recognized contrary and well-respected evidence is a lousy belief-sustaining procedure. And since I am a mere mortal, I am aware of the conﬂict between my belief in P and the contrary theory, aware that the contrary theory is live, etc. So I’m quite aware that my beliefs are being held in the face of well-respected expert contrary evidence.

192

Defeating the Live Sceptic Head-On

B Extreme Externalism We see a similar point when discussing the Extreme Externalist solution. Maybe the processes by which our second-order, ﬁrstperson beliefs (your beliefs about your own beliefs) are formed are incredibly reliable: just try to describe a possible world in which virtually all of your beliefs about your beliefs are false but you aren’t totally demented. The processes by which we judge colours are pretty reliable too; the same for ﬂatness, solidity, and pain location. And this high degree of reliability holds no matter what anyone thinks about any of these sceptical hypotheses. One might think that these reliability facts provide enough warrant for our true colour, pain, ﬂatness, solidity, and second-order beliefs to amount to knowledge. The live sceptic has no allegiance to internalism in any of its usual formulations (see discussion in section B of Chapter 4). I have allowed the anti-sceptic whatever tools she wants to defeat the live sceptic, including externalistic ones, but I have not anywhere assumed that those tools are legitimate or illegitimate. As with every topic he touches, Dretske (1991) has something illuminating to say about the motivations behind the externalism/internalism debate. One starts doing epistemology by looking at certain paradigm cases and going from there. You could start with scientiﬁc knowledge; that’s the top-down strategy. Here language, conceptual sophistication, articulation, critical reasoning, and justiﬁcation are centrestage. Or you could start with Fido the dog; that’s the bottom-up strategy. When we start there, we become engrossed in considering reliability, proper function, safety, sensitivity, and the co-ordination of behaviour with the immediate physical environment. Pursuing the former often lands one in internalism; the latter leads one to externalism. It might be thought that the live sceptic is a top-downer. After all, she points to the highly theoretical developments among scientists and philosophers in order to generate her sceptical results. Reliability isn’t enough; you need fancy reasons to neutralize the live hypotheses. But this is misleading. The live sceptic is neutral on the topdown/bottom-up debate. Suppose the bottom-uppers are right. Even so, the live sceptic is looking at members of highly sophisticated epistemic communities, not young children or non-human animals. Not only that: she is considering members of those communities who

Defeating the Live Sceptic Head-On

193

have certain kinds of special awareness of, and participation in, epistemically advanced circles. The live sceptic has allowed that the 3-year-old effortlessly came to know that the ﬁre-engine is red. Problems didn’t arise until adulthood—and even then only because of certain highly advanced circumstances in her intellectual community.62 The externalist can, against the live sceptic, insist that those high-level developments cannot take away the knowledge gained as a 3-year-old, but only at a price, as we are about to see. The live sceptic does have to balk at extreme kinds of externalism. In what follows I will ﬁrst argue that reliability factors (or other external factors such as proper functioning) do not provide enough warrant to make our threatened true beliefs knowledge even if those reliability facts are sufﬁcient to guarantee signiﬁcant amounts of high-quality warrant, even warrant sufﬁcient to defeat the traditional sceptic. Keep in mind, however, two important qualiﬁcations. First, I’m not arguing against the reliabilist solution to traditional scepticism or against the epistemic importance of reliability. For all I argue below, some reliabilist responses to traditional scepticism are entirely successful. Second, the traditional reliabilist is by no means forced by her theory to deny the live scepticisms. There is nothing in reliabilism or many kinds of externalism that forces one to deny all sceptical conclusions. The only reliabilists and other externalists I consider here are ones who wish to use reliabilist or other externalist moves to block the live sceptical arguments. The reliabilism solution is an instance of the Defeated-Threat Strategy: we have Subject factors—reliability features—that are strong enough for the big neutralization task. Suppose that these reliability facts sufﬁce to place a signiﬁcant amount of positive epistemic warrant on our second-order, ﬁrst-person beliefs. Perhaps they supply enough warrant that in ordinary circumstances, when we aren’t in a society fallen to serious consideration of eliminativism, true second-order, ﬁrst-person beliefs are often good enough for knowledge. Perhaps they supply so much warrant that even in an eliminativist-live society people can have second-order knowledge if they are quite divorced from and ignorant of both philosophy and cognitive science. Even so, eliminativism is actually live, and I’m aware of and it and its status. Suppose I happen to reject the 62 Dretske’s judgements about the examples he describes in that 1991 article support, to some extent, the new sceptic’s position. Other remarks, in his 1981, suggest the opposite.

194

Defeating the Live Sceptic Head-On

eliminativist hypothesis, and so believe that I have beliefs. Still, I’m no immortal regarding eliminativism. It seems that there is now a signiﬁcant amount of negative epistemic warrant placed on my second-order, ﬁrst-person beliefs (this is the negative warrant revealed by the eliminativist sceptical argument). And there is enough of it to render the positive reliability-based warrant insufﬁcient to render my second-order beliefs warranted enough for knowledge. Only an exceedingly strong form of reliabilism can back up a neutralization of the eliminativist hypothesis. No reliabilist I know of has endorsed such a theory. Everyone has admitted that factors like the recognition of expert counter-opinion, liveness, independent contrary evidence, and what not can defeat a reliably formed belief—even highly reliable ones. The one-off sceptical arguments I gave in Chapter 6 conclusively demonstrate the point. So no real reliabilist has said anything to suggest that he or she would endorse the reliabilist solution given here. The crucial point here is the same as the one we saw with safety and sensitivity: if the degree of reliability of the belief-producing process has signiﬁcant epistemic weight (as the reliabilist reasonably holds), then the degree of reliability of the belief-sustaining process also has signiﬁcant epistemic weight. This is no mystery. The beliefsustaining process has to be reliable to retain knowledge, and in the cases described in this essay—liveness, mere mortality—to continue to have beliefs that are known to be in contradiction with the live hypotheses is clearly highly unreliable. That is, the following beliefsustaining procedure—which applies to the beliefs in question—is unreliable: continue to believe that P even though a contrary hypothesis Q has become live and one is a mere mortal with respect to Q. The reliabilist might be right that the reliability of the belief-forming process is often enough to make a true belief knowledge; so when mere mortals were children, their beliefs amounted to knowledge. Our question has to do with the retention of that knowledge upon coming aware of contrary expert evidence. The reliabilist who wants to use reliability factors to defeat the live sceptical arguments has to claim not only that the belief-producing reliability confers a signiﬁcant amount of warrant, but also, mysteriously, that belief-sustaining reliability is irrelevant. Reliability of belief formation gives me 1,218 positive warrant units; standing fast with the belief that P in face of appreciated, well-respected, expert evidence produces just 200 negative warrant units; I needed just 1,000 units to have my true belief

Defeating the Live Sceptic Head-On

195

turned into knowledge. Alternatively: belief-forming reliability gives me positive warrant of sufﬁcient quality for knowledge, and beliefsustaining unreliability amounts to no veto. Should we swallow this extremism? The generality problem for reliabilism is relevant, but I won’t address the issue here. I will make just one comment. One cannot object to my argument of the previous paragraph by claiming that the mere mortal isn’t aware that she is doing something that falls under the belief-sustaining-process description ‘continue to believe that P even though a contrary hypothesis Q has become live and one is aware of the inconsistency’. Since she is a mere mortal, she is aware of the live error theory, its liveness, and its inconsistency with her threatened belief. To that extent, the belief-sustaining method is particularly relevant to the epistemic status of the belief. For those reasons it seems to me that reliability factors are probably not sufﬁcient to neutralize the live sceptical hypotheses. The basic point again: once we grant the epistemic power of the reliability of the belief-producing process, we must grant the epistemic power of the belief-sustaining process. C The Tenacity of Knowledge I would like now to consider a more radical approach, one that insists on the tenacity of knowledge.63 Roughly put, once you come to know that P, it is very hard to lose that knowledge without losing the belief. The reasons and other epistemic factors underlying one’s knowledge do not have to hang around lifting up the knowledge for ever. Instead, the reasons and other Subject factors hoist the belief on to the table of knowledge and are then no longer needed to keep supporting it. Now that it’s on the table, it is very hard to knock it off. This is a view regarding the survival skills of knowledge states. It says that once you’ve got knowledge, it can survive many potent assaults. But the new sceptic’s views apply to knowledge gained as a non-mortal—she says you lose your knowledge that ﬁre-engines are red—as well as beliefs gained while a mortal. In the latter case, if you receive a pair of socks for your birthday next week, you, being a mere 63 Special thanks go to Andrew McGonigal for making me appreciate the force of this objection.

196

Defeating the Live Sceptic Head-On

mortal with respect to colour error theory, do not come to know that they are blue. Here the question isn’t whether knowledge can survive the transition from non-mortality to mortality but whether belief acquired during mortality amounts to knowledge. The anti-sceptic impressed with the tenacity of knowledge can take either of two views. The moderate Tenacity view holds that knowledge gained as a non-mortal survives mortality. Thus, since you knew long ago that ﬁre-engines are red, you still know it today. The strong Tenacity view agrees with the moderate view concerning survival but in addition holds that in virtue of the old knowledge remaining, the related true beliefs acquired during mortality also amount to knowledge. For instance, in virtue of the tenacity of knowledge, I keep my childhood knowledge that ﬁre-engines are red (that’s the moderate Tenacity view); and because my new car is obviously the same colour as that ﬁre-engine across the street, I can easily come to know—today as a mere mortal—that my new car is red (that’s the strong Tenacity view). Of course, the anti-sceptic needs to show how this happens. First problem. The moderate view collapses into either the strong view or live scepticism. The strong Tenacity advocate reasons this way: because the mere mortal is in the same epistemic position with regard to the new car she just saw across the street (today, as a mere mortal) as she is with the ﬁre-engines she saw in her youth, and she also knows that the ﬁre-engines are red (as she retains that piece of knowledge), she now knows that the new car is red. One can also reason in this way: Moore believed that scepticism is false (this is our old belief); and Goldman is no different from Moore on that matter (a new belief); so Goldman has that belief as well (new knowledge inconsistent with eliminativism). If Moore has beliefs, then surely we do as well, even right now! But of course the live sceptic just turns this around: since the mere mortal doesn’t know that the new car is red, and she is in the same epistemic position with regard to the new car as she is with regard to ﬁre-engines, the mere mortal doesn’t know that ﬁre-engines are red either, even though she once knew. Tit for tat. Thus, it’s either the strong Tenacity view or live scepticism. Second problem. Return to the whales-as-ﬁsh case, in which common sense is right that whales are ﬁsh and the scientists who justiﬁably think otherwise are mistaken. One method for ﬁnding out that something is a ﬁsh—such as ﬁnding out it’s a whale—is a relatively direct, non-scientiﬁc, and highly reliable belief-producing method available to many ordinary people, of many backgrounds. Max comes

Defeating the Live Sceptic Head-On

197

to believe, as a non-mortal, that the blue whale is the largest ﬁsh. Later, when he is an adult, everyone is screaming at him that the blue whale is not a ﬁsh; but Max sticks with his belief, even though he understands what the screamers are screaming, he is aware of the liveness of their view, etc. So he is guilty of a serious epistemic vice. Further, he is deﬁnitely following a highly unreliable, unsafe, and insensitive belief-sustaining method. That’s a total of four powerful reasons for denying Max knowledge that the blue whale is the largest ﬁsh. I fail to see any relevant difference between the whales-as-ﬁsh case and the live sceptic’s case; the same four reasons are present in each case. When we persist with our claims to know that ﬁre-engines are red, we are following the same highly unreliable, unsafe, insensitive, and epistemically vicious belief-sustaining method as Max did. The Hank Aaron home run story supports the same conclusion. Jane starts out knowing that Aaron hit 755 home runs. Then Smith and Jones get their good but ultimately misleading evidence that he hit just 753 home runs. Once Jane is a fully-ﬂedged mere mortal with respect to the live 753-homer hypothesis, if she sticks to her 755homer belief she fails to have knowledge. Surely that’s obvious: all these baseball experts are screaming that Aaron hit only 753 homers, the evidence is excellent, Jane can’t do anything to poke any holes in that evidence, she is aware of that evidence and its respected standing, and yet she continues to believe that Aaron hit 755 homers. That’s just not knowledge in my book. (And if it is knowledge, then epistemologists should be embarrassed about spending so much time on something so trivial as knowledge.) So knowledge isn’t that tenacious. Third problem. Suppose I’m completely wrong about everything: the individual continues to have knowledge even in the face of recognized expert screaming. No matter what I argue, one can put one’s foot down and insist that if someone comes to believe, in the normal ways, that the table is red and is the cause of the throbbing pain in one’s shin, then the belief-producing warrant that person has for those beliefs is so outstanding that virtually nothing can cancel its adequacy for knowledge—not recognized expert opinion; not unreliable, unsafe, insensitive belief-sustaining methods; not serious epistemic violations; nothing.64 The beliefs in question are so peculiar, one 64 I have encountered this response to the new sceptic’s conclusion that we don’t know that ﬁre-engines are red. Apparently, the fact that many colour scientists and scientiﬁcally informed philosophers are convinced that ﬁre-engines aren’t red had no effect whatsoever on this fellow. He is a theist.

198

Defeating the Live Sceptic Head-On

might think, that no matter what happened, if one came to believe, truly and in the ordinary way, that the socks are blue, then one knows it even if one recognizes full well that all the experts in heaven and on earth scream otherwise. But surely this is a case of impoverished knowledge! Surely this is a case of the anti-sceptic winning the battle but not the war. The live sceptic might be wrong in holding that the stubborn individuals don’t know, but she is right in her central claim that the overall epistemic quality that these true beliefs have is much less than any of us have ever supposed in our anti-sceptical moments. What she has shown is that even if we mere mortals do retain into mortality the knowledge that common sense says we have, this knowledge is much less epistemically impressive than any antisceptic has ever supposed. Let me be clear: the live sceptic who concedes the existence of knowledge of colours, pains, character traits, and beliefs, but challenges its status, is not saying that all our knowledge is pathetic; no such global claim is being made. She admits that our other knowledge may be epistemically very impressive. Her point is that the knowledge threatened by the live sceptical arguments is of much, much lower quality than anyone has ever suspected, and typically much lower than parts of common-sense knowledge not threatened by contrary live hypotheses. She is targeting a special set of beliefs, ones targeted by the live error theories, and claiming that the warrant possessed by these beliefs is anaemic compared to that had by our other beliefs that amount to knowledge. To that extent, the live sceptic has won. As I have mentioned several times, we can obey the anti-sceptical voices in our heads if we want; there is nothing really stopping us; but we are holding on to a prize not worth having.

13

Concluding Reﬂections The threat of scepticism is what keeps the theory of knowledge going. Barry Stroud, The Signiﬁcance of Philosophical Scepticism

Suppose aliens landed on Earth tomorrow with their super-duper spaceships. That would be incredibly interesting. But it wouldn’t change epistemology. So there are aliens; who cares? Now suppose that the live sceptic is right. Should we care? Sure enough, if she’s right, then in many worlds some really intelligent people get some of their seemingly most secure knowledge sabotaged. That is indeed unexpected. But in this essay have we stumbled across something that should change the way we do any of the important parts of epistemology? Or have we just learned some sexy yet ultimately epistemologically irrelevant facts? The existence of aliens is fascinating, but doesn’t change our theories of knowledge or warrant; should the soundness of the live sceptic’s argument change our theories of knowledge or warrant or other epistemologically important notions? The live sceptic’s argument suggests that it’s facts about liveness and mere mortality that rob us mere mortals of our knowledge. So this essay has brought into prominence the epistemic importance of liveness and mere mortality, which are central to knowledge and justiﬁed belief independently of the truth of live scepticism. This in turn makes us probe epistemic deference further than we might have, as the epistemic importance of liveness is partly constituted by such deference. Obviously the live sceptic’s argument raises questions of the epistemic signiﬁcance of agreement and disagreement, which should be particularly interesting to philosophers. For instance, who are you to disagree with David Lewis? It’s not terribly important which claim you disagree with him about. It might be the claim that there are

200

Concluding Reﬂections

other concrete possible worlds. The problem is that even though he held that P while you hold that :P, you know full well that he was your epistemic superior, even with regard to the topics surrounding P. David Lewis could kick your philosophical ass when it comes to modality—or just about anything else metaphysical for that matter. Not only that: he put much more thought into the question whether P than you ever will. He is your epistemic superior in just about every way when it comes to modality, and yet you say that he’s wrong and you’re right on that very topic. And of course this has nothing in particular to do with David Lewis or you. For instance, Lewis, Kripke, Wright, Williamson, Chalmers, and many other contemporary philosophers can each kick my philosophical ass as well, and yet I disagree with them even though I admit that they can kick my ass. Are we mere mortal philosophers just overwhelmingly and incurably arrogant? Of course, this problem is not limited to philosophical belief. There are many instances in which we believe P even though we know full well that there are many experts who believe the opposite while being our epistemic superiors in just about every relevant way with regard to P and the topics surrounding P. By ‘epistemic superior’ I mean that they are generally more informed than we are on the topics involving P, they have more raw intelligence, they have thought and investigated whether P longer and in more depth than we have, they have thought and investigated the topics surrounding P longer and in more depth than we have, they are just as or even more intellectually careful than we are, they are just as or even more intellectually honest than we are, and, crucially, we admit that they have understood and fairly and thoroughly evaluated the same evidence as we have—and usually more evidence. Needless to say, this issue is highlighted by the live sceptic and needs exploring. The live sceptic also throws some light on the issue of what possibilities must be ruled out in order to know something. In Chapter 7 we looked at the question of if and under what conditions we have to rule out possibilities consistent with P in order to know P. In Chapter 6 we saw that A could know P while B does not, even though in some sense B’s belief in P is epistemically superior to A’s. I have not explored that idea in this essay, but this seems to be an idea worth pursuing.65 65 Finally, if the live sceptic is right, then the ‘norm of assertion’ is probably not knowledge, as Tim Williamson (2000) and many others think. Just because I don’t know

Concluding Reﬂections

201

Now return to the epistemic importance of liveness and mere mortality. Why should those facts sabotage our knowledge—especially if they don’t ruin the knowledge of children and adults who aren’t mere mortals? If the live sceptic is right, in virtue of what facts, what truth-makers, is she right? I think that the right picture is probably along the following externalist lines. We are part of an epistemic community that pools its knowledge, reasons, beliefs, concepts, meanings, and warrants. We share an enormous epistemic and mental treasure. Roughly put, if you came to know that water is a compound just by reading it in a book or hearing it from a science teacher, if as a matter of actual fact you came to know in this incredibly easy way that physical fact which took many brilliant people many years to ﬁnd out, then you participate in other epistemic activities of that community even if you’re not expert in them, even if you’re not fully aware of them, and, to a certain extent, even if you don’t want to participate. One way in which this happens is that if such-and-such experts are part of your public knowledge source, then what’s live for them is live for you, possibly subject to exceptions in certain peculiar knowledge-attribution contexts. If the experts pass some of their knowledge on to you, then they must pass on liveness conditions as well. Given that certain hypotheses are live in our intellectual community, we have an epistemic responsibility to them in the sense just articulated. You don’t get to pick and choose; you get the bad with the good. Kripke, Putnam, Kaplan, Burge, and others have argued famously over the last thirty years for a more social theory of meaning, reference, and belief.66 Roughly put, what your words refer to and their particular meaning are transferred to you by your community. The fact that you use ‘Feynman’ to refer to Feynman, the fact that you understand the English sentence ‘Feynman was a physicist’, the fact that you even use that sentence at all as a sentence (instead of that ﬁre-engines are red, there is little—I would say ‘no’—impermissibility in my telling my daughter that ﬁre-engines are red. For one thing, most mere mortals don’t think that they don’t know that ﬁre-engines are red. But this depends on just what ‘the constitutive rule of assertion’ really amounts to. 66 I have in mind Kripke’s and Putnam’s work on reference (Kripke 1980 and Putnam 1975), Kripke’s Wittgensteinian work on meaning (1982), Kaplan and Putnam’s work on ‘consumerism’ (Kaplan 1989 and Putnam 1975), and Burge’s work on belief (1979). Obviously, Wittgensteinians have been indicating similar thoughts for decades, but usually without any arguments rigorous enough or theses precise enough to even attempt to evaluate. Sorry, but that’s the way it’s always struck me!

202

Concluding Reﬂections

using it as an ink pattern), the fact that you use ‘Feynman was a physicist’ to express your belief that Feynman was a physicist, the fact that you have that belief, and even the fact that you mean any truth at all with that sentence, are things you get from your epistemic and linguistic community. Epistemologists have always known about the powerful source of testimony. Crudely put, much of the justiﬁcation you have for your beliefs exists in virtue of justiﬁcation had by your community, which somehow passes over to you as a member of that community. To a large extent, the notions of interest to philosophers—reference, meaning, knowledge, and justiﬁcation— are things we inherit from our community. Philosophers of mind and epistemologists have traditionally heavily favoured the individual over the community. The individual is important, no doubt, but what these contemporary philosophers have shown us is that social relations—in the guise of linguistic, doxastic, and epistemic inﬂuence from one’s community—are crucial as well. Kripke and the rest have focused on the good things we get: knowledge, belief, meaning, and reference. In this essay I have explored another potential community ‘gift’, but one that isn’t so wonderful: the undermining of warrant and the subsequent downfall of knowledge. Even so, I have tried to emphasize that in a real sense the sabotaging of knowledge is in some epistemic sense a gift: those in the sceptical snare are epistemically better off than those out of the loop regarding the live sceptical hypotheses. The live sceptic’s line of thought can be taken to an extreme. One might think that even if someone isn’t aware of the presence of the live but false hypotheses in one’s epistemic community, one’s true belief doesn’t amount to knowledge. The mere membership in the epistemic community with certain live hypotheses is enough to sabotage knowledge obviously inconsistent with those hypotheses. So philosophers or cognitive scientists with no awareness at all of the sceptical hypotheses lose their knowledge. This claim might be correct; I have argued neither for nor against it. But it is a view signiﬁcantly more radical than the one I’ve defended in this essay. I stipulate that the mere mortals are aware of the live status of the rival hypotheses, but perhaps unaware of the details of the reasons why those hypotheses are live. If one is an epistemic externalist, then one will be attracted to the comparison of many of the qualities had by our beliefs—reference, meaning, and warrant especially—with the economic qualities had

Concluding Reﬂections

203

by our goods and services. The monetary value of a good can change as a result of factors quite distant from that good. Analogously, certain changes that are quite extrinsic to our minds and bodies can have profound effects on the linguistic, epistemic, and cognitive values of our mental states. But those effects are present only if you have a membership card for the community in which those changes occur.67 I think the arguments for, and particulars of, live scepticism are of signiﬁcant philosophical interest even if they don’t change the way we construct theories of knowledge and warrant. However, I suspect that they should affect how we proceed in building a theory of knowledge and epistemic value, via a closer investigation of the ‘social’ nature of epistemic warrant. Obviously, some of the issues to be pursued concern the kinds and nature of other epistemic consequences of our membership in our epistemic communities. There is something simple, clean, hypnotizing, and beautiful about the classic arguments for radical scepticism. They can mesmerize just about anyone, regardless of philosophical or scientiﬁc background. To a certain extent, they keep philosophers employed. The arguments for live scepticism may never be simple, clean, hypnotizing, or beautiful; they will never constitute a jobs programme for philosophers. But they do make us view scepticism with new eyes. 67

Burge has explored these issues with regard to belief content (e.g. 1979, 1986, 1989).

References Brown, J. (2005), ‘Adapt or Die: The Death of Invariantism?’, Philosophical Quarterly. Burge, T. (1979), ‘Individualism and the Mental’, in French, Uehling Jr., and Wettstein (eds.), Midwest Studies in Philosophy: Studies in Metaphysics, iv (Minneapolis: University of Minnesota Press), 73–121. —— (1986), ‘Intellectual Norms and the Foundations of Mind’, Journal of Philosophy 83: 697–720. —— (1989), ‘Wherein is Language Social?’, in George (ed.), Reﬂections on Chomsky (Oxford: Blackwell). Byrne, A., and Hilbert, D. (2003), ‘Color Realism and Color Science’, Behavioral and Brain Sciences 26: 3–21. Churchland, P. M. (1989), A Neurocomputational Perspective: The Nature of Mind and the Structure of Science (Cambridge, Mass.: MIT Press). Churchland, P. S. (1986), Neurophilosophy: Toward a Uniﬁed Science of the Mind-Brain (Cambridge, Mass.: MIT Press). Cohen, S. (1988), ‘How to be a Fallibilist’, in Tomberlin (ed.), Philosophical Perspectives, ii (Atascadero, Calif.: Ridgeview Publishing Company), 91–123. —— (1999), ‘Contextualism, Skepticism, and the Structure of Reasons’, in Tomberlin (ed.), Philosophical Perspectives, xiii (Oxford: Blackwell), 57–89. Conee, E. (2001), ‘Heeding Misleading Evidence’, Philosophical Studies 103: 99–120. Dennett, D. (1978), Brainstorms (Montgomery, Vt.: MIT Press). DeRose, K. (1992), ‘Contextualism and Knowledge Attributions’, Philosophy and Phenomenological Research 52: 913–29. —— (1995), ‘Solving the Skeptical Problem’, Philosophical Review 104: 1–52. —— (1999), ‘Contextualism: An Explanation and Defence’, in Greco and Sosa (eds.), The Blackwell Guide to Epistemology (Oxford: Blackwell). —— (2002), ‘Assertion, Knowledge, and Context’, Philosophical Review 111: 167–203. —— (2005), ‘The Ordinary Language Basis for Contextualism and the New Invariantism’, Philosophical Quarterly. —— (forthcoming), ‘Sosa, Safety, Sensitivity, and Skeptical Hypotheses’, in Greco (ed.), Sosa and his Critics (Oxford: Blackwell). Doris, J. (1998), ‘Persons, Situations and Virtue Ethics’, Nouˆs 32: 504–30. —— (2002), Lack of Character (Cambridge: Cambridge University Press).

References

205

Dretske, F. (1970), ‘Epistemic Operators’, Journal of Philosophy 69: 1007–23. —— (1981), ‘The Pragmatic Dimension of Knowledge’, Philosophical Studies 40: 363–78. —— (1991), ‘Two Conceptions of Knowledge: Rational vs. Reliable Belief ’, Grazer Philosophische Studien 4: 15–30. Feyerabend, P. (1963), ‘Materialism and the Mind–Body Problem’, Review of Metaphysics 17: 49–66. Frances, B. (1999), ‘Contradictory Belief and Epistemic Closure Principles’, Mind and Language 14: 203–26. —— (forthcoming), ‘When a Skeptical Hypothesis is Live’, Nouˆs. Ginet, C. (1980), ‘Knowing Less by Knowing More’, in French, Uehling Jr., and Wettstein, (eds.), Midwest Studies in Philosophy Studies in Epistemology, v (Minneapolis: University of Minnesota Press). Goldman, A. (1975), ‘Discrimination and Perceptual Knowledge’, Journal of Philosophy 73: 771–91. Greco, J. (2002), ‘How to Reid Moore’, Philosophical Quarterly 52: 544–63. Hardin, L. (1988), Color for Philosophers: Unweaving the Rainbow (Indianapolis: Hackett). Harman, G. (1980), ‘Reasoning and Evidence One Does Not Possess’, in French, Uehling Jr., and Wettstein (eds.), Midwest Studies in Philosophy Studies in Epistemology, v (Minneapolis: University of Minnesota Press). —— (1999), ‘Moral Philosophy Meets Social Psychology: Virtue Ethics and the Fundamental Attribution Error’, Proceedings of the Aristotelian Society 99: 315–31. —— (2000), ‘The Nonexistence of Character Traits’, Proceedings of the Aristotelian Society 100: 223–6. Heller, M. (1999), ‘The Proper Role for Contextualism in an Anti-Luck Epistemology’, in Tomberlin (ed.), Philosophical Perspectives, xiii (Oxford: Blackwell), 115–29. Hyman, J. (2003), ‘Pains and Places’, Philosophy 78: 5–24. Kaplan, D. (1989), ‘Afterthoughts’, in Almog, Wettstein, and Perry (eds.), Themes from Kaplan (New York: Oxford University Press), 565–614. Kripke, S. (1979), ‘A Puzzle about Belief ’, in Margalit (ed.), Meaning and Use, (Dordrecht: D. Reidel), 239–83. Reprinted in Salmon and Soames (eds.), Propositions and Attitudes (New York: Oxford University Press, 1988), 102–48. —— (1980), Naming and Necessity (Cambridge, Mass.: Harvard University Press). —— (1982), Wittgenstein on Rules and Private Language (Cambridge, Mass.: Harvard University Press). Lewis, D. (1996), ‘Elusive Knowledge’, Australasian Journal of Philosophy 74: 549–67.

206

References

Lycan, W. (1977), ‘Evidence One Does Not Possess’, Australasian Journal of Philosophy 55: 114–26. —— (2001), ‘Moore Against the New Skeptics’, Philosophical Studies 103: 35–53. Moore, G. E. (1962), Philosophical Papers (New York: Collier Books). Nozick, R. (1981), Philosophical Explanations (Cambridge, Mass.: Harvard University Press). Pritchard, D. (2002), ‘Recent Work on Skepticism’, American Philosophical Quarterly 39: 215–56. Putnam, H. (1975), ‘The Meaning of ‘‘Meaning’’ ’, in his Mind, Language, and Reality (Cambridge: Cambridge University Press), 215–71. Quine, W. V. O. (1960), Word and Object (Cambridge, Mass.: MIT Press). —— (1985), ‘States of Mind’, Journal of Philosophy 82: 5–8. Rorty, R. (1970), ‘In Defence of Eliminative Materialism’, Review of Metaphysics 24: 112–21. Sosa, E. (1999), ‘How to Defeat Opposition to Moore’, in Tomberlin (ed.), Philosophical Perspectives, xiii (Oxford: Blackwell), 141–54. —— (2000), ‘Skepticism and Contextualism’, in Sosa and Villanueva (eds.), Philosophical Issues, x (Oxford: Blackwell), 1–18. Stich, S. (1983), From Folk Psychology to Cognitive Science: The Case Against Belief (Cambridge, Mass.: MIT Press). —— (1996), Deconstructing the Mind (New York: Oxford University Press). Stine, G. (1976), ‘Skepticism, Relevant Alternatives, and Deductive Closure’, Philosophical Studies 29: 249–61. Stroud, B. (1984), The Signiﬁcance of Philosophical Scepticism (Oxford: Oxford University Press). Williamson, T. (2000), Knowledge and its Limits (Oxford: Oxford University Press).

Index acceptance 35; see also belief Ambiguity Solution 180–2 anti-sceptical strategies 128–32 arrogance 44, 49, 148–9, 199–200 Bad Reasoning Solution 162–4 Beall, J. C. 60 belief 35, 53–7, 77, 99–103, 143–4, 160–1, 180–2 Bloomfield, Paul x bootstrap sceptical arguments, see meta-sceptical hypotheses brain in a vat hypothesis 3–4, 46, 64–5, 74, 97, 188–9; see also scepticism, live versus traditional; scepticism, traditional brain in a vat society 93–7 Brown, Jessica 148 Burge, Tyler 201–2, 203 Byrne, Alex 70 CA factors 123–7 Chalmers, David x character traits 71 Churchland, Patricia 38, 41–3, 174 Churchland, Paul 38, 41–3, 174 closure principles 25–6, 109–13, 182–3; see also ruling out cognitive science 39–40 Cohen, Stewart 15, 123 colour 69–71, 76, 98–9 common sense 57–9, 70, 87, 156–77 communities, epistemic 82, 201 Comparison Solution 164–77 Conee, Earl 92 consumerism 201–2 content, belief 91, 201–2 context 122–5, 129–30; see also knowledge attributions

contextualism 15–16, 97, 127–8, 131–2 CS factors 122–3, 125–7 Defeated-Threat Strategy 129, 131, 132–3, 185–98 defeater 86, 91–3; see also dinosaur story; Senator Smith story; whales-fish story degrees of belief 60–2 Dennett, Daniel 38, 42–3, 174, 181 DeRose, Keith 15, 110–11, 117, 123, 186 dinosaur story 1–2, 98, 125–7 Disarmed-Threat Strategy 129, 130–1, 133, 137–52 Doris, John 71 Double Standard Solution 156–60 Dretske, Fred 15, 26, 186, 192, 193 eliminativism 30–1, 35, 36–41, 42, 45–6, 53–7, 68, 75–6, 119, 171 epistemic blame 61–2, 196; see also arrogance; rationality; reasonable belief epistemic threats 4, 122–7, 131, 134 error theories 57–9, 74–5, 169–70; see also character traits; colour; eliminativism; pain location evidence 124; see also warrant that one does not possess 59–60 experts 18–19, 22, 45, 49–50, 62–3, 88, 148–9, 199–200; see also Mere Mortal Premiss; mere mortality externalism, epistemic 51–3 Extreme Externalism Solution 192–5 Feyerabend, Paul 38, 41–3

208

Index

folk epistemology 26–7, 118–20; see also common sense folk psychology 38–9; see also eliminativism Frances, Margaret x Fumerton, Richard x Gettier cases 85–6, 122; see also Harman, Gilbert Ginet, Carl 92 Goldberg, Sanford x Goldman, Alvin 15 Greco, John x, 164, 176 Hank Aaron story 146–8, 197 Hardin, Larry 70 Harman, Gilbert 59–60, 71, 85–6, 92 Hilbert, David 70 Hume, David 164 Hyman, John 71 internalism, epistemic 51–3; see also externalism, epistemic introspection 55; see also ruling out belief-eliminativism; ruling out pain-eliminativism justification, see knowledge; warrant Kaplan, David 201–2 KK principle 106 knowledge 26–7, 35, 82–5; see also ruling out; scepticism; warrant knowledge attributions 88–91, 96–7, 144–52, 173–4; see also CA factors; context; contextualism; CS factors; Disarmed-Threat Strategy; Practicality Solution; Set-Aside Solution; subject factors Kripke, Saul 36, 103, 201–2

Lewis, David 15 Limited Scope Solution 178–80 Live Hypothesis Principle 24–5, 28, 32, 47, 130–1 liveness 8–9, 10, 18–19, 24, 41, 42, 45–6, 48, 199 Lycan, William 92, 176 McGonigal, Andrew x, 195 McGrath, Sarah x Melia, Joseph x mere mortality 11, 19–23, 53–5, 199 Metaphysical Distance Solution 154–5 Mere Mortal Premiss 25, 32, 37, 41, 43, 53–5, 57, 130–1 metaphysics 58, 170 meta-sceptical hypotheses 64–9 Millican, Peter x Modesty Principle 24, 32, 47, 130–1 Moore, G. E. 110–11, 164–77 Nelson, Mark x No-Threat Strategy 129, 130–1, 132, 153–84 Nozick, Robert 25, 186 one-off sceptical results 97–103 pain, see ruling out paineliminativism pain location 71, 80, 99 paradoxes 34, 176 Practicality Solution 144–52 precedents for live scepticism, see one-off sceptical results Pritchard, Duncan x, 28 purple fish story 87–91, 93 Putnam, Hilary 201–2 Quine, W. V. O. 38, 42

Index

209

‘rather than’ 141–2 rationality 49, 60 reasonable belief 49; see also arrogance; rationality; warrant Reid, Thomas 164 relevant alternatives 15–16, 23–4; see also CS factors; knowledge attributions; subject factors reliabilism 192–5; see also externalism, epistemic; internalism, epistemic; Extreme Externalism Solution Robustly Live Solution 160–2 Rorty, Richard 38, 42–3 ruling out beliefeliminativism 53–7 ruling out pain-eliminativism 53–7 ruling out 9, 11-12, 19, 45, 94–6, 106–13

persistent 116–17; see also scepticism, and knowledge traditional 3–5, 25–6, 46, 47, 96, 97, 127, 157 science 71–2, 79, 156; see also Double Standard Solution Senator Smith story 91–3, 98, 191 Set-Aside Solution 137–44 Sosa, Ernest 186, 190–1 Stich, Stephen 36, 38, 41–3, 174 Stine, Gail 15 Stoneham, Tom x Stroud, Barry 110–11 subject factors 122–7, 131, 134 surgery on spine story 99

Safety–Sensitivity Solution 186–91 sceptical conclusions 64–6, 73–4, 76–97 sceptical hypotheses 3–4, 64–72, 73–6 sceptical solution 73–105 scepticism 64–72, 73–4, 75–6 and knowledge 26–7, 67, 114–20, 151–2, 197–8 live 31–4, 43–4, 60, 61, 83–5, 86, 104–5, 202 live versus traditional 132–5

universal scepticism 106–13

Tenacity Solution 195–8 Testibility-Indeterminacy Solution 182–4 trick socks story 98–9

virtues, epistemic 196 warrant 26, 49, 50, 82–4, 114–15; see also defeater; knowledge; rationality water-element story 10–11, 12, 134, 138–43, 189 whales-fish story 79–81, 196–7 Williamson, Timothy 200 Wittgenstein, Ludwig 36