I Know
Freitag_RZ.indd 1
Wolfgang Freitag
Freitag · I Know. Modal Epistemology and Scepticism
Scepticism, the view that knowledge is impossible, threatens our conception of ourselves as epistemic subjects as much as it endangers our conception of the external world. The book develops a modal account of knowledge and provides an answer to scepticism based on a detailed examination of the main sceptical argument. It discusses prominent contemporary theories of knowledge, in particular safety and sensitivity theories, and shows that they cannot handle Gettier-type examples of a new kind. An alternative analysis of knowledge in terms of relevantly normal possibilities is developed. The sceptical argument addressed aims to show that we cannot know ordinary things because we cannot rule out that we are in a sceptical scenario. Classical responses, like dogmatism, non-closure theories, and epistemic contextualism, are explored and rejected as unnecessary for a refutation of the sceptical argument. A detailed investigation reveals, first, that the failure to know that we are not in a sceptical scenario does not conflict with ordinary knowledge, but only with knowledge that we know, and, second, that we can indeed know that we are not in a sceptical scenario. It is therefore claimed not only that we know, but also that we know that we know.
I Know. Modal Epistemology and Scepticism
31.01.13 11:51
Wolfgang Freitag · I Know
Wolfgang Freitag
I Know Modal Epistemology and Scepticism
mentis MÜNSTER
Gedruckt mit Unterstützung des Förderungs- und Beihilfefonds Wissenschaft der VG Wort Einbandabbildung: © magann - Fotolia.com
Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http://dnb.dnb.de abrufbar.
Gedruckt auf umweltfreundlichem, chlorfrei gebleichtem ∞ ISO 9706 und alterungsbeständigem Papier
© 2013 mentis Verlag GmbH Eisenbahnstraße 11, 48143 Münster, Germany www.mentis.de Alle Rechte vorbehalten. Dieses Werk sowie einzelne Teile desselben sind urheberrechtlich geschützt. Jede Verwertung in anderen als den gesetzlich zulässigen Fällen ist ohne vorherige Zustimmung des Verlages nicht zulässig. Printed in Germany Einbandgestaltung: Anne Nitsche, Dülmen (www.junit-netzwerk.de) Druck: AZ Druck und Datentechnik GmbH, Kempten ISBN 978-3-89785-774-2
Table of Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9
PART 1 KNOWLEDGE AND MODALITY 1
Two worlds: Scepticism and the modal approach to knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Two worlds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pyrrhonian scepticism . . . . . . . . . . . . . . . . . . . . . . . . . . . Cartesian scepticism . . . . . . . . . . . . . . . . . . . . . . . . . . . . Knowledge and guarantee . . . . . . . . . . . . . . . . . . . . . . . .
. . . . .
15 15 21 27 31
2.1 2.2 2.3 2.4
Modal accounts of knowledge: The failure of closeness analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lewis’s account . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sensitivity and safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Safety, sensitivity, and distant epistemic luck . . . . . . . . . . . Analysis of distant epistemic luck . . . . . . . . . . . . . . . . . . .
41 42 47 53 61
3 3.1 3.2 3.3 3.4
Knowledge, normality, and inference . . . . . . . . . . . . Guarantee and normal conditions . . . . . . . . . . . . . . . . . . . Guarantee, lotteries, and Gettier . . . . . . . . . . . . . . . . . . . . Knowledge and inference . . . . . . . . . . . . . . . . . . . . . . . . . . Inference and closure . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
67 68 75 78 87
Conclusion: The rationale of modal scepticism . . . .
93
1.1 1.2 1.3 1.4 2
PART 2 KNOWLEDGE AND SCEPTICISM 4 4.1 4.2 4.3
The sceptical paradox . . . . . . . . . . . . . . . . . . . . . . . . . The basic form of the sceptical argument . . . . . . . . . . . . . Reformulating the sceptical argument . . . . . . . . . . . . . . . Reactions to the sceptical argument . . . . . . . . . . . . . . . . .
. . . .
101 101 103 106
6
5 5.1 5.2 5.3 6 6.1
Table of Contents
Theoretical anti-scepticism: Concession or confrontation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Non-closure theories . . . . . . . . . . . . . . . . . . . . . . . . . . . . Epistemic contextualism . . . . . . . . . . . . . . . . . . . . . . . . . Neo-Mooreanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . .
109 109 119 135
.
141
6.2 6.3 6.4
Diagnostic anti-scepticism . . . . . . . . . . . . . . . . . . . . . Beyond containment and confrontation: A dilemma for scepticism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scepticism and focus . . . . . . . . . . . . . . . . . . . . . . . . . . . . A pragmatic explanation for alethic scepticism . . . . . . . . Pragmatic scepticism? . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . .
141 150 155 165
7 7.1 7.2 7.3 7.4
Scepticism and iterative knowledge . . . . . . . . . . . . . Dialectical options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Higher-order knowledge . . . . . . . . . . . . . . . . . . . . . . . . . Iterative knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Final thoughts on Cartesian scepticism . . . . . . . . . . . . . .
. . . . .
169 170 172 176 181
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
183
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Person Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
185 193
Preface
In spite of the significance of scepticism for epistemology, I have set out writing this book without any intention of dealing with this topic at all, let alone making it the main theme of my explorations. The original plan was to discuss modal approaches to epistemology and the many ways of Gettiering justified beliefs, more or less what is now contained in Chapters 2 and 3. But my studies developed in different directions, all of which appeared to be related to scepticism in some way or other—a development that perhaps betrays my natural inclinations. So scepticism presented itself as the inescapable destiny of the project and I had to accept it as the main theme of the book. My original focus, modal epistemology, connects to this theme in two different ways. Firstly, it evolves from what I consider to be the epistemological conditio humana, the separation between subjective and objective world (ch. 1). Secondly, modal epistemology has grave consequences for the sceptical argument in its various forms. The challenge with respect to this argument is to provide an analysis that points out where the argument is faulty and at the same time explains why it appears to be so convincing nevertheless. This double challenge has often been recognized, but extant responses left me deeply unsatisfied. Even the most sympathetic readings of (neo-)Mooreanism, non-closure theories, contextualism, or relativism were unable to lift the mist in which my thinking appeared to be enshrouded. A radically different and, I think, more natural approach was needed. I explore this alternative in Chapters 4–7 based on a modal conception of knowledge. My views have evolved over a number of years and under the influence of countless discussions, presentations, formulations, and reformulations of the topics and ideas. The most important types of influence on my thinking are probably those which I am unable to fathom. Though I do not recall the individual instances, I remember the occasional sense of deep truth in mere side-remarks, intonations, and gestures made by teachers, colleagues, and friends. I am grateful to those who sidestepped theorizing to advance the right perspective on the phenomenon. I have greatly profited from the fact that, while in the process of writing this book, I have been able to work in the departments of philosophy at Konstanz, and in more recent times, at Bern and Heidelberg. In all three places I have found a great atmosphere and, most importantly, colleagues and friends with whom I enjoyed excellent discussions on topics directly and indirectly related to this book. Amongst the colleagues that must be named are Franz
8
Preface
Huber, Andreas Kemmerling, Holger Leerhoff, Jacob Rosenthal, and Holger Sturm. (Probably there are many others, who will hopefully forgive my forgetfulness.) I want to thank Ede Zimmermann, who made me aware of the linguistic literature on focus. I am also grateful to the participants of various colloquia and conference talks for many valuable comments, criticisms, and suggestions on various parts of the book. The book grew out of my habilitation thesis, which was refereed by Josef Bayer, Thomas Grundmann, Wolfgang Spohn, and Marcel Weber. They all provided me with excellent advice and suggestions in a way this final version has greatly profited from. For the remaining mistakes, of course, I alone am responsible. I am especially grateful to Christopher von Bülow, who corrected the final draft of this book and prepared the index, and to Michael Kienecker from mentis, who was extremely supportive throughout this project. Most important to the book, and to me, has been, and is, and will be, A. Z.— my alphabet. The Baden-Württemberg Stiftung supported much of my research on epistemology in its starting phase. The Zukunftskolleg Konstanz promoted my work in many different ways during my five-year fellowship, in which the bulk of this book was written. An early version of Sections 2.2 and 2.3 has been published as part of the paper “Sensitivity, Safety, and ‘Distant’ Epistemic Luck” in Theoria. I wish to thank the publishers for their kind permission to make use of this material here. Heidelberg, October 2012
Wolfgang Freitag
Introduction
I know. I know that the sun is bigger than the moon, that the sun is shining, that I have two hands, and that they are presently typing on my keyboard. I know that there is a tree in front of the window, that such a window exists, and that the window needs cleaning. I know a lot, and I am not alone in this. Tom knows that Tim is a good squash player, but should better not set foot on a football field. And Tina knows that her new car is green and that it still belongs to the bank. One need not go as far as David Lewis (1996, p. 418) and think that it is “a Moorean fact that we know a lot” to appreciate that we know a lot. To suppose that we do not have any knowledge as a matter of fact is not straightforwardly absurd. Knowledge usually does not come for free, but it is a real achievement. We know what we know because we have perceived things to be that way, because we have been told about them, or because we have read about them. We gain knowledge, at least usually, by means of some empirical investigation. Our experiments may go wrong, the books may be unreliable or faulty, and our informants may be biased. Then we do not know. Barring such deceptive circumstances, however, we do have knowledge. That we know a lot is an ordinary empirical fact, but it is a fact nevertheless. There is also a lot we do not know. We don’t know the number of planets in the universe or whether there is extraterrestrial life. We don’t know whether the last dinosaur died on a Tuesday, nor the outcome of the lottery next Saturday. Much, or even most, of what there is to know we do not know. About many things we will never find out. It will be easy to get to know on Sunday that our lottery ticket has, once again, not won. We simply watch the news or read the paper. By contrast, it would be hard, perhaps even impossible, to discover whether or not the last dinosaur died on a Tuesday. The contents of black holes might be forever concealed from us and so are some of the thoughts of our beloved ones. Although we are not in epistemological paradise, we are epistemological optimists and pride ourselves of epistemic wealth. There are a lot of things that we are able to know—if we care to investigate them and invest enough time and commitment. All this seems utterly trivial, yet a little reflection shows that it is not. Even if it should be true that we can know a lot, this truth is not obvious. If it is a truth it must be rescued from the surging flames of scepticism, which has a strong line of argument in the following. We don’t know, the sceptic says, that we are not brains in vats who are fed with perfectly-ordinary-seeming experiences by a supercomputer, because such a scenario of total deception
10
Introduction
is subjectively indistinguishable from an ordinary, non-deceptive scenario in which we have those experiences one would expect to have if one were leading an ordinary life. But if we don’t know that we are not deceived we don’t know what we ordinarily believe to know. So, contrary to our deepest convictions, it appears that we don’t know that the sun is bigger than the moon or that the sun is presently shining. In the light of the possibility of evil demons, of our being envatted brains or victims of the Matrix, how should any knowledge be possible? Scepticism denies the possibility of what I have declared to be an obvious empirical fact: that knowledge is abundant. If scepticism is right we are epistemically blind, unresponsive to any information. Despite its implausibility, however, scepticism cannot be belittled as the ghost haunting the epistemologist’s dreams. It appears to be the dire epistemological reality apparently undeniable when we are most aware, most clearly awake, and not stupefied by the routines of ordinary life. One of the principal, or primary, motives for epistemology is to give an answer to the sceptical challenge. One response would be to accept the sceptical claim that knowledge is impossible and a fortiori to accept that it is not actual. But this is, as I have already claimed, a highly implausible answer. We know a lot, and that we can know a lot is something we know better than the premise of any sceptical argument. Scepticism would involve a major change in our conceptual scheme, and would have deep repercussions not only for what we know but also for what we do. That it conflicts so fundamentally with our ordinary knowledge attributions, and with the possibility thereof, and hence with the epistemic language game we play, is indicative of the fact that scepticism is wrong. Consequently I take the task of epistemology to be conservative, explanatory of our everyday practice, not revisionary in claiming this practice unsound. Thus, our task is to give a non-sceptical answer and protect knowledge from the sceptical argument. My goal in this book is to avert the sceptical impetus and to demonstrate that we can remain epistemological optimists despite all the cognitive difficulties and shortcomings of our species. My main objective therefore is to show that famous arguments for scepticism, in particular the argument from ignorance based on the possibility of such sceptical scenarios as our being envatted brains or mere Matrix existences, can be undermined. To succeed in this task, we have to take the claims and aims of scepticism seriously. We must not hold it lightly, thinking that we can brush it off easily. Otherwise we will miss the lessons scepticism tries to teach us. If we can give a response to it at all, it will only be through a succession of detailed arguments and demonstrations. Undermining the sceptical argument has been one of the major aims in much of the contemporary epistemological debate. Indeed, the sceptical
Introduction
11
challenge can be understood as the starting and focal point of different major trends in epistemology. Non-closure theorists, epistemic contextualists, and neo-Mooreanists (not to mention contrastivists and relativists) all pride themselves on giving the proper response to the sceptical challenge. But all of the mentioned positions are highly contentious and therefore make for problematic anti-sceptical candidates. If any one of them were embraced the corresponding response to scepticism would be conditional upon substantial philosophical commitments. But if scepticism were itself refutable only by recourse to philosophical theory, this would be an important concession to the sceptic. For without knowledge of the respective anti-sceptical theory we would not know that scepticism is false and would not be in a position to claim that we know anything. Not least because of such general considerations, I think that these epistemological theses do not represent the right answer to the sceptical challenge—whatever their merits otherwise might be. I will show that the sceptical argument can be rejected without the adoption of any of these contentious theories, indeed without adopting any contentious view at all. It will turn out that the difficult part is not so much to give the right response to scepticism. The difficult part is to explain why we have been lured into scepticism in the first place, why we consider the sceptical challenge to be a hard problem, one which can be resolved only by the force of philosophical apparatus. It will turn out that there are a number of different factors contributing to the apparent appeal of scepticism, some of a semantic and others of a pragmatic character. My approach to scepticism is therefore largely diagnostic, a demonstration that we might have been led astray at crucial junctures by the sceptic’s reasoning. But the diagnostic approach is accompanied by a constructive account of knowledge. What is it that we possess when we have knowledge? Recent answers suggest that it is a sort of belief which is immune to error. Knowledge is belief that cannot (in a sense to be specified) be false. In this sense, knowledge is necessarily true belief. This is the common credo of modal epistemology, 1 which has become the most prominent, and most successful, approach to the theory of knowledge. It is part of the task of this book to present the overall aims, methods, and strategies of modal epistemology. Modal epistemology has been developed mainly in response to the challenge posed by the Gettier problem, the problem of epistemic luck. In order to be able to rule out epistemic luck, we need some type of necessity 1
When I use the term “modal epistemology”, I refer to a certain type of epistemology, namely one which bases its fundamental term, “knowledge”, on modalities. I therefore deviate from the rather established use which understands “modal” in this context as qualifying the propositions whose knowledge is in question. For such a discipline I would prefer the label “epistemology of modalities”. I hope that no confusion springs from my usage.
12
Introduction
concerning the truth of beliefs; where a belief is necessarily true, it cannot be luckily true, at least not in the way relevant for knowledge. Although modal epistemology has been developed primarily in response to the Gettier problem, we can also see it as a response to the substantial worries of scepticism, to the problem of our beliefs’ being possibly false. Accordingly, the book falls into two parts. The first part provides a positive account of knowledge in modal terms. Chapter 1 describes the basic starting point by distinguishing the objective world (consisting of the facts) and the subjective world (composed of an agent’s experiences and doxastic states). In Chapter 2, I develop the basic idea of contemporary modal conceptions of knowledge, in particular safety theories, sensitivity theories, and Lewis’s account, and show that they are crucially defective. As a reaction, I develop an alternative in Chapter 3, and extend the modal analysis to inferential knowledge. Chapters 4 to 6 are dedicated to the discussion and the undermining of the sceptical argument as presented above. After a precise formulation of the argument and its aims (ch. 4), I will discuss the three most prominent reactions to it, namely non-closure theories, epistemic contextualism, and neo-Mooreanism (ch. 5). This will prepare the stage for my own solution in Chapter 6, where it will be shown that the sceptical conclusion rests on premises for which there is no consistent reading such that they are both plausible. Even if my argument against the sceptic about knowledge is successful, however, there is room for a sceptical comeback, according to which, although knowledge is possible, knowing that one knows is not. As a consequence of such higher-order scepticism, we would have no way of knowing that our beliefs are epistemically proper. Chapter 7, however, is an attempt to show that we can resist this further sceptical argument. I hope thereby to arrive at what I conceive as perhaps the fundamental anti-sceptical issue: to secure not only that there is an epistemic condition worthy of being called “knowledge”, but that this condition itself can be known to be present.
PART 1 KNOWLEDGE AND MODALITY
This first part is concerned with the basic idea of modal epistemology. According to this approach, knowledge is belief which is based on some restricted form of necessity, i. e., on truth in all members of a certain set of possible worlds. The central task of modal epistemology is therefore to provide a non-trivial, well-motivated, and revealing characterisation of the relevant set. The heuristic strategy for coping with this task is to give a constructive definition of knowledge in response to our intuitions. Prominent roles will be played by two different kinds of problem cases, namely Gettier cases and lottery cases, which yield constraints on a successful account of knowledge. As will emerge, the Gettier problem will be of particular importance, since a successful response to it involves the right determination of the set of possibilities involved. I will first, in Chapter 1, discuss the general form which a modal account must take, and introduce what I consider the crucial notion involved, namely the notion of what I call “guarantee”. My goal is then to review prominent accounts of the guarantee relation: Robert Nozick’s sensitivity account, David Lewis’s account, and Duncan Pritchard’s safety characterisation. This will be the content of Chapter 2. I claim that they are indeed successful with respect to the lottery puzzle. They also promise to deal with the Gettier problem, but it will be shown that while they are successful with regard to Gettier cases of the traditional type, they founder on a new kind of Gettier problem. In response to the failure of these analyses I will, in Chapter 3, give what I consider a new type of modal account, an account of guarantee in terms of a normal condition clause, and show that—and how—this new type of account operates with Gettier problems and lottery problems.
14
Knowledge and Modality
I will start with a general description of the relation between two worlds: the ‘objective world’, the world as it is, and the ‘subjective world’, the world as it appears to a given epistemic agent. This will provide us with insights into the two fundamental epistemic predicaments and hence furnish us with two different forms of sceptical challenges, which I will label ‘Cartesian’ and ‘Pyrrhonian’, respectively. The basic idea of modal epistemology will then be developed in response to the former.
Wolfgang Freitag
Freitag · I Know. Modal Epistemology and Scepticism
Scepticism, the view that knowledge is impossible, threatens our conception of ourselves as epistemic subjects as much as it endangers our conception of the external world. The book develops a modal account of knowledge and provides an answer to scepticism based on a detailed examination of the main sceptical argument. It discusses prominent contemporary theories of knowledge, in particular safety and sensitivity theories, and shows that they cannot handle Gettier-type examples of a new kind. An alternative analysis of knowledge in terms of relevantly normal possibilities is developed. The sceptical argument addressed aims to show that we cannot know ordinary things because we cannot rule out that we are in a sceptical scenario. Classical responses, like dogmatism, non-closure theories, and epistemic contextualism, are explored and rejected as unnecessary for a refutation of the sceptical argument. A detailed investigation reveals, first, that the failure to know that we are not in a sceptical scenario does not conflict with ordinary knowledge, but only with knowledge that we know, and, second, that we can indeed know that we are not in a sceptical scenario. It is therefore claimed not only that we know, but also that we know that we know.
I Know. Modal Epistemology and Scepticism
31.01.13 11:51
Wolfgang Freitag · I Know
Wolfgang Freitag
I Know Modal Epistemology and Scepticism
mentis MÜNSTER
Gedruckt mit Unterstützung des Förderungs- und Beihilfefonds Wissenschaft der VG Wort Einbandabbildung: © magann - Fotolia.com
Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http://dnb.dnb.de abrufbar.
Gedruckt auf umweltfreundlichem, chlorfrei gebleichtem ∞ ISO 9706 und alterungsbeständigem Papier
© 2013 mentis Verlag GmbH Eisenbahnstraße 11, 48143 Münster, Germany www.mentis.de Alle Rechte vorbehalten. Dieses Werk sowie einzelne Teile desselben sind urheberrechtlich geschützt. Jede Verwertung in anderen als den gesetzlich zulässigen Fällen ist ohne vorherige Zustimmung des Verlages nicht zulässig. Printed in Germany Einbandgestaltung: Anne Nitsche, Dülmen (www.junit-netzwerk.de) Druck: AZ Druck und Datentechnik GmbH, Kempten ISBN 978-3-89785-774-2
Table of Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9
PART 1 KNOWLEDGE AND MODALITY 1
Two worlds: Scepticism and the modal approach to knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Two worlds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pyrrhonian scepticism . . . . . . . . . . . . . . . . . . . . . . . . . . . Cartesian scepticism . . . . . . . . . . . . . . . . . . . . . . . . . . . . Knowledge and guarantee . . . . . . . . . . . . . . . . . . . . . . . .
. . . . .
15 15 21 27 31
2.1 2.2 2.3 2.4
Modal accounts of knowledge: The failure of closeness analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lewis’s account . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sensitivity and safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Safety, sensitivity, and distant epistemic luck . . . . . . . . . . . Analysis of distant epistemic luck . . . . . . . . . . . . . . . . . . .
41 42 47 53 61
3 3.1 3.2 3.3 3.4
Knowledge, normality, and inference . . . . . . . . . . . . Guarantee and normal conditions . . . . . . . . . . . . . . . . . . . Guarantee, lotteries, and Gettier . . . . . . . . . . . . . . . . . . . . Knowledge and inference . . . . . . . . . . . . . . . . . . . . . . . . . . Inference and closure . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
67 68 75 78 87
Conclusion: The rationale of modal scepticism . . . .
93
1.1 1.2 1.3 1.4 2
PART 2 KNOWLEDGE AND SCEPTICISM 4 4.1 4.2 4.3
The sceptical paradox . . . . . . . . . . . . . . . . . . . . . . . . . The basic form of the sceptical argument . . . . . . . . . . . . . Reformulating the sceptical argument . . . . . . . . . . . . . . . Reactions to the sceptical argument . . . . . . . . . . . . . . . . .
. . . .
101 101 103 106
6
5 5.1 5.2 5.3 6 6.1
Table of Contents
Theoretical anti-scepticism: Concession or confrontation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Non-closure theories . . . . . . . . . . . . . . . . . . . . . . . . . . . . Epistemic contextualism . . . . . . . . . . . . . . . . . . . . . . . . . Neo-Mooreanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . .
109 109 119 135
.
141
6.2 6.3 6.4
Diagnostic anti-scepticism . . . . . . . . . . . . . . . . . . . . . Beyond containment and confrontation: A dilemma for scepticism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scepticism and focus . . . . . . . . . . . . . . . . . . . . . . . . . . . . A pragmatic explanation for alethic scepticism . . . . . . . . Pragmatic scepticism? . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . .
141 150 155 165
7 7.1 7.2 7.3 7.4
Scepticism and iterative knowledge . . . . . . . . . . . . . Dialectical options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Higher-order knowledge . . . . . . . . . . . . . . . . . . . . . . . . . Iterative knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Final thoughts on Cartesian scepticism . . . . . . . . . . . . . .
. . . . .
169 170 172 176 181
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
183
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Person Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
185 193
Preface
In spite of the significance of scepticism for epistemology, I have set out writing this book without any intention of dealing with this topic at all, let alone making it the main theme of my explorations. The original plan was to discuss modal approaches to epistemology and the many ways of Gettiering justified beliefs, more or less what is now contained in Chapters 2 and 3. But my studies developed in different directions, all of which appeared to be related to scepticism in some way or other—a development that perhaps betrays my natural inclinations. So scepticism presented itself as the inescapable destiny of the project and I had to accept it as the main theme of the book. My original focus, modal epistemology, connects to this theme in two different ways. Firstly, it evolves from what I consider to be the epistemological conditio humana, the separation between subjective and objective world (ch. 1). Secondly, modal epistemology has grave consequences for the sceptical argument in its various forms. The challenge with respect to this argument is to provide an analysis that points out where the argument is faulty and at the same time explains why it appears to be so convincing nevertheless. This double challenge has often been recognized, but extant responses left me deeply unsatisfied. Even the most sympathetic readings of (neo-)Mooreanism, non-closure theories, contextualism, or relativism were unable to lift the mist in which my thinking appeared to be enshrouded. A radically different and, I think, more natural approach was needed. I explore this alternative in Chapters 4–7 based on a modal conception of knowledge. My views have evolved over a number of years and under the influence of countless discussions, presentations, formulations, and reformulations of the topics and ideas. The most important types of influence on my thinking are probably those which I am unable to fathom. Though I do not recall the individual instances, I remember the occasional sense of deep truth in mere side-remarks, intonations, and gestures made by teachers, colleagues, and friends. I am grateful to those who sidestepped theorizing to advance the right perspective on the phenomenon. I have greatly profited from the fact that, while in the process of writing this book, I have been able to work in the departments of philosophy at Konstanz, and in more recent times, at Bern and Heidelberg. In all three places I have found a great atmosphere and, most importantly, colleagues and friends with whom I enjoyed excellent discussions on topics directly and indirectly related to this book. Amongst the colleagues that must be named are Franz
8
Preface
Huber, Andreas Kemmerling, Holger Leerhoff, Jacob Rosenthal, and Holger Sturm. (Probably there are many others, who will hopefully forgive my forgetfulness.) I want to thank Ede Zimmermann, who made me aware of the linguistic literature on focus. I am also grateful to the participants of various colloquia and conference talks for many valuable comments, criticisms, and suggestions on various parts of the book. The book grew out of my habilitation thesis, which was refereed by Josef Bayer, Thomas Grundmann, Wolfgang Spohn, and Marcel Weber. They all provided me with excellent advice and suggestions in a way this final version has greatly profited from. For the remaining mistakes, of course, I alone am responsible. I am especially grateful to Christopher von Bülow, who corrected the final draft of this book and prepared the index, and to Michael Kienecker from mentis, who was extremely supportive throughout this project. Most important to the book, and to me, has been, and is, and will be, A. Z.— my alphabet. The Baden-Württemberg Stiftung supported much of my research on epistemology in its starting phase. The Zukunftskolleg Konstanz promoted my work in many different ways during my five-year fellowship, in which the bulk of this book was written. An early version of Sections 2.2 and 2.3 has been published as part of the paper “Sensitivity, Safety, and ‘Distant’ Epistemic Luck” in Theoria. I wish to thank the publishers for their kind permission to make use of this material here. Heidelberg, October 2012
Wolfgang Freitag
Introduction
I know. I know that the sun is bigger than the moon, that the sun is shining, that I have two hands, and that they are presently typing on my keyboard. I know that there is a tree in front of the window, that such a window exists, and that the window needs cleaning. I know a lot, and I am not alone in this. Tom knows that Tim is a good squash player, but should better not set foot on a football field. And Tina knows that her new car is green and that it still belongs to the bank. One need not go as far as David Lewis (1996, p. 418) and think that it is “a Moorean fact that we know a lot” to appreciate that we know a lot. To suppose that we do not have any knowledge as a matter of fact is not straightforwardly absurd. Knowledge usually does not come for free, but it is a real achievement. We know what we know because we have perceived things to be that way, because we have been told about them, or because we have read about them. We gain knowledge, at least usually, by means of some empirical investigation. Our experiments may go wrong, the books may be unreliable or faulty, and our informants may be biased. Then we do not know. Barring such deceptive circumstances, however, we do have knowledge. That we know a lot is an ordinary empirical fact, but it is a fact nevertheless. There is also a lot we do not know. We don’t know the number of planets in the universe or whether there is extraterrestrial life. We don’t know whether the last dinosaur died on a Tuesday, nor the outcome of the lottery next Saturday. Much, or even most, of what there is to know we do not know. About many things we will never find out. It will be easy to get to know on Sunday that our lottery ticket has, once again, not won. We simply watch the news or read the paper. By contrast, it would be hard, perhaps even impossible, to discover whether or not the last dinosaur died on a Tuesday. The contents of black holes might be forever concealed from us and so are some of the thoughts of our beloved ones. Although we are not in epistemological paradise, we are epistemological optimists and pride ourselves of epistemic wealth. There are a lot of things that we are able to know—if we care to investigate them and invest enough time and commitment. All this seems utterly trivial, yet a little reflection shows that it is not. Even if it should be true that we can know a lot, this truth is not obvious. If it is a truth it must be rescued from the surging flames of scepticism, which has a strong line of argument in the following. We don’t know, the sceptic says, that we are not brains in vats who are fed with perfectly-ordinary-seeming experiences by a supercomputer, because such a scenario of total deception
10
Introduction
is subjectively indistinguishable from an ordinary, non-deceptive scenario in which we have those experiences one would expect to have if one were leading an ordinary life. But if we don’t know that we are not deceived we don’t know what we ordinarily believe to know. So, contrary to our deepest convictions, it appears that we don’t know that the sun is bigger than the moon or that the sun is presently shining. In the light of the possibility of evil demons, of our being envatted brains or victims of the Matrix, how should any knowledge be possible? Scepticism denies the possibility of what I have declared to be an obvious empirical fact: that knowledge is abundant. If scepticism is right we are epistemically blind, unresponsive to any information. Despite its implausibility, however, scepticism cannot be belittled as the ghost haunting the epistemologist’s dreams. It appears to be the dire epistemological reality apparently undeniable when we are most aware, most clearly awake, and not stupefied by the routines of ordinary life. One of the principal, or primary, motives for epistemology is to give an answer to the sceptical challenge. One response would be to accept the sceptical claim that knowledge is impossible and a fortiori to accept that it is not actual. But this is, as I have already claimed, a highly implausible answer. We know a lot, and that we can know a lot is something we know better than the premise of any sceptical argument. Scepticism would involve a major change in our conceptual scheme, and would have deep repercussions not only for what we know but also for what we do. That it conflicts so fundamentally with our ordinary knowledge attributions, and with the possibility thereof, and hence with the epistemic language game we play, is indicative of the fact that scepticism is wrong. Consequently I take the task of epistemology to be conservative, explanatory of our everyday practice, not revisionary in claiming this practice unsound. Thus, our task is to give a non-sceptical answer and protect knowledge from the sceptical argument. My goal in this book is to avert the sceptical impetus and to demonstrate that we can remain epistemological optimists despite all the cognitive difficulties and shortcomings of our species. My main objective therefore is to show that famous arguments for scepticism, in particular the argument from ignorance based on the possibility of such sceptical scenarios as our being envatted brains or mere Matrix existences, can be undermined. To succeed in this task, we have to take the claims and aims of scepticism seriously. We must not hold it lightly, thinking that we can brush it off easily. Otherwise we will miss the lessons scepticism tries to teach us. If we can give a response to it at all, it will only be through a succession of detailed arguments and demonstrations. Undermining the sceptical argument has been one of the major aims in much of the contemporary epistemological debate. Indeed, the sceptical
Introduction
11
challenge can be understood as the starting and focal point of different major trends in epistemology. Non-closure theorists, epistemic contextualists, and neo-Mooreanists (not to mention contrastivists and relativists) all pride themselves on giving the proper response to the sceptical challenge. But all of the mentioned positions are highly contentious and therefore make for problematic anti-sceptical candidates. If any one of them were embraced the corresponding response to scepticism would be conditional upon substantial philosophical commitments. But if scepticism were itself refutable only by recourse to philosophical theory, this would be an important concession to the sceptic. For without knowledge of the respective anti-sceptical theory we would not know that scepticism is false and would not be in a position to claim that we know anything. Not least because of such general considerations, I think that these epistemological theses do not represent the right answer to the sceptical challenge—whatever their merits otherwise might be. I will show that the sceptical argument can be rejected without the adoption of any of these contentious theories, indeed without adopting any contentious view at all. It will turn out that the difficult part is not so much to give the right response to scepticism. The difficult part is to explain why we have been lured into scepticism in the first place, why we consider the sceptical challenge to be a hard problem, one which can be resolved only by the force of philosophical apparatus. It will turn out that there are a number of different factors contributing to the apparent appeal of scepticism, some of a semantic and others of a pragmatic character. My approach to scepticism is therefore largely diagnostic, a demonstration that we might have been led astray at crucial junctures by the sceptic’s reasoning. But the diagnostic approach is accompanied by a constructive account of knowledge. What is it that we possess when we have knowledge? Recent answers suggest that it is a sort of belief which is immune to error. Knowledge is belief that cannot (in a sense to be specified) be false. In this sense, knowledge is necessarily true belief. This is the common credo of modal epistemology, 1 which has become the most prominent, and most successful, approach to the theory of knowledge. It is part of the task of this book to present the overall aims, methods, and strategies of modal epistemology. Modal epistemology has been developed mainly in response to the challenge posed by the Gettier problem, the problem of epistemic luck. In order to be able to rule out epistemic luck, we need some type of necessity 1
When I use the term “modal epistemology”, I refer to a certain type of epistemology, namely one which bases its fundamental term, “knowledge”, on modalities. I therefore deviate from the rather established use which understands “modal” in this context as qualifying the propositions whose knowledge is in question. For such a discipline I would prefer the label “epistemology of modalities”. I hope that no confusion springs from my usage.
12
Introduction
concerning the truth of beliefs; where a belief is necessarily true, it cannot be luckily true, at least not in the way relevant for knowledge. Although modal epistemology has been developed primarily in response to the Gettier problem, we can also see it as a response to the substantial worries of scepticism, to the problem of our beliefs’ being possibly false. Accordingly, the book falls into two parts. The first part provides a positive account of knowledge in modal terms. Chapter 1 describes the basic starting point by distinguishing the objective world (consisting of the facts) and the subjective world (composed of an agent’s experiences and doxastic states). In Chapter 2, I develop the basic idea of contemporary modal conceptions of knowledge, in particular safety theories, sensitivity theories, and Lewis’s account, and show that they are crucially defective. As a reaction, I develop an alternative in Chapter 3, and extend the modal analysis to inferential knowledge. Chapters 4 to 6 are dedicated to the discussion and the undermining of the sceptical argument as presented above. After a precise formulation of the argument and its aims (ch. 4), I will discuss the three most prominent reactions to it, namely non-closure theories, epistemic contextualism, and neo-Mooreanism (ch. 5). This will prepare the stage for my own solution in Chapter 6, where it will be shown that the sceptical conclusion rests on premises for which there is no consistent reading such that they are both plausible. Even if my argument against the sceptic about knowledge is successful, however, there is room for a sceptical comeback, according to which, although knowledge is possible, knowing that one knows is not. As a consequence of such higher-order scepticism, we would have no way of knowing that our beliefs are epistemically proper. Chapter 7, however, is an attempt to show that we can resist this further sceptical argument. I hope thereby to arrive at what I conceive as perhaps the fundamental anti-sceptical issue: to secure not only that there is an epistemic condition worthy of being called “knowledge”, but that this condition itself can be known to be present.
PART 1 KNOWLEDGE AND MODALITY
This first part is concerned with the basic idea of modal epistemology. According to this approach, knowledge is belief which is based on some restricted form of necessity, i. e., on truth in all members of a certain set of possible worlds. The central task of modal epistemology is therefore to provide a non-trivial, well-motivated, and revealing characterisation of the relevant set. The heuristic strategy for coping with this task is to give a constructive definition of knowledge in response to our intuitions. Prominent roles will be played by two different kinds of problem cases, namely Gettier cases and lottery cases, which yield constraints on a successful account of knowledge. As will emerge, the Gettier problem will be of particular importance, since a successful response to it involves the right determination of the set of possibilities involved. I will first, in Chapter 1, discuss the general form which a modal account must take, and introduce what I consider the crucial notion involved, namely the notion of what I call “guarantee”. My goal is then to review prominent accounts of the guarantee relation: Robert Nozick’s sensitivity account, David Lewis’s account, and Duncan Pritchard’s safety characterisation. This will be the content of Chapter 2. I claim that they are indeed successful with respect to the lottery puzzle. They also promise to deal with the Gettier problem, but it will be shown that while they are successful with regard to Gettier cases of the traditional type, they founder on a new kind of Gettier problem. In response to the failure of these analyses I will, in Chapter 3, give what I consider a new type of modal account, an account of guarantee in terms of a normal condition clause, and show that—and how—this new type of account operates with Gettier problems and lottery problems.
14
Knowledge and Modality
I will start with a general description of the relation between two worlds: the ‘objective world’, the world as it is, and the ‘subjective world’, the world as it appears to a given epistemic agent. This will provide us with insights into the two fundamental epistemic predicaments and hence furnish us with two different forms of sceptical challenges, which I will label ‘Cartesian’ and ‘Pyrrhonian’, respectively. The basic idea of modal epistemology will then be developed in response to the former.
