(f‘ .7 u , . - . ..-.v:.
. “A” . 3.... . «

 

. 3”.” an .
.5. v I I . l t
.a :..nn.3... . I.“
. ~ :5.) l‘l .- 1. A . (“9a.

‘1 1......

.533? t

Jim!

to . V
tummy?
um? . a

v n

.1: .2».

.v I 5. 5
.4! f. .1359;
1 ..J.-.n:..llx.:...:

. t.
»

cf.

« vac-“ﬂutvlusnﬂhun
.Jvhmkaﬁﬁuulf
.15... 5:. huh I...
..il¢;ﬂ.u3uuv
93vi..kma.
13-11 In.

nﬁﬁdmv ghmwm...
x

wuu ,

1.? 0 560:; .. .11
35.22.9111!-
II .1 ,

.5. .
a 9;. L5 . . 2%
F3. ..

0:. mm.

'7.-

3 an" apron-mm;

a . ‘2 I5
‘14:.
a

Haunt!!!
($1.23!...
6 1251.6 . 173:5

m...“ 0‘. A. «an... AQHMV'W

 

 

 

ﬁﬂﬂLnﬁﬁuéxﬁﬂu , . ‘ . . ..»§E..E..£ ‘ . _ . . , , , , . airﬁrrl.

":“U~'~._

290}

This is to certify that the

dissertation entitled

MEANINGS, MANNERS, AND SCEPTICISM

presented by

Sarah Black Jones

has been accepted towards fulﬁllment
of the requirements for

Ph . D . degree in Philosophy

 

WCW

Major proféstgrd

Date 12/14/01

 

MS U is an Afﬁrmative Action/Equal Opportunity Institution 0-12771

 

 

 

LlBRARY

Michigan State

University

 

 

PLACE IN RETURN BOX to remove

TO AVOID FINES return on or

MAY BE RECALLED with earlier due (1

this checkout from your record.

before date due.
ate if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cleIFiC/DateDuo.p65-p.15

 

MEANINGS, MANNERS, AND SCEPTICISM
By

Sarah Black Jones

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Philosophy

2001

ABSTRACT
MEANINGS, MANNERS, AND SCEPTICISM
By

Sarah Black Jones

In Meanings, Manners, and Scepticism, I argue that the ordinary language philosopher,
the relevant altemativist, and the contextualist fail to refute the sceptic. I argue that they
fail to refute the sceptic because their arguments critically rest on a number of both
dogmatic and erroneous assumptions. In particular they rest on the dogmatic and
erroneous assumptions that common sense is threatened by skepticism, that ordinary
linguistic behaviour is of epistemic signiﬁcance, and that the sceptic’s departure from the
linguistic norm sufﬁces to prove his sceptical alternatives irrelevant to the truth of
ordinary claims to know. I argue that the ﬁrst assumption is mistaken because it wrongly
assumes that common sense consists not only of ﬁrst-order beliefs about the world, but
also second-order knowledge beliefs about those beliefs. I argue that the second
assumption is mistaken because it conﬂates two logically distinct sets of standards: the
standards of appropriate and the standards of true assertion. Finally, I argue that the third
assumption is mistaken because it rests on a confusion of the kind of standard operative
in ordinary knowledge-discourse with the kind of standard the sceptic appeals to when he
reasons that we can never know. I conclude that appealing to ordinary linguistic
behaviour and habits cannot sufﬁce to prove the sceptic’s hypothesis irrelevant to the

truth of our knowledge claims, though insofar as common sense does not include second-

Sarah Black Jones
order knowledge beliefs the truth of even the most radical skepticism cannot threaten our

ordinary claims to know.

For Julia and James, my mother and father
and
Katharine, my daughter

iv

ACKNOWLEDGMENTS

There a number of people who I wish to thank for having made the completion of this work
possible. Most especially, I would like to thank certain of my teachers, each of whom it was
a great privilege to know: First, and most especially, professor Herbert J. Garelick whose call
to the vocation of philosophy must surely have been Divine, and of whom no words of praise
or appreciation could possibly do justice. To him, I am most grateful for having been his
student, and for the unique and unforgettable privilege of having been in his presence.
Professor Charles J. McKracken whose classical style and commitment to tradition will
forever inﬂuence me. Professor Hal Walsch for his absolute reﬁisal to acquiesce to senseless
dogmatism, and, whose penchant to lapse into Latin and Greek made for the most wonderful
of learned lectures. Professor Richard J. Hall, for the years of his always intellectually
stimulating courses and seminars, and for his encouragement. The members of my dissertation
committee, especially Debrah Nails and Winston Wilkinson for their detailed comments.
Finally, professors Stephen Esquith and Albert Cafagna, the kindest of philosophers, without

whose support I surely could not have completed this work.

3
i
I
!

 

TABLE OF CONTENTS

I. Introduction ........................................................ 1
II. Meanings, Manners, and Scepticism .................................... 20
§ i. The Sceptic ............................................... 27
§ ii. The Argument From Ordinary Language ......................... 34
§ iii. Disentangling: A Conﬂation of Manners and Truth ............... 44
III. Contextualism .................................................... 58
§ i. What is Contextualism? ....................................... 60
§ ii. David Lewis ............................................... 62
§ iii. Stewart Cohen ............................................. 65
§iv. Keith DeRose ............................................. 70
§ v. Contextualism and Ordinary Language ........................... 74
§ vi. DeRose and Invariantism ..................................... 87
§ vii. Conclusion ............................................... 96
IV. Relevant Altemativism .............................................. 99
§ i. Relevant Alternatives and Epistemic Operators .................... 102
§ ii. Does Henry Know? Stine’s Reason to Believe and the Simple Approach
to Relevant Alternatives ..................................... 129
§ iii. Goldman and Perceptual Discrimination ........................ 137
§ iv. Dretske Revisited ......................................... 147
§ v. Relevant Altemativism and Contextualism: Relevant Differences ...... 162
V. The Epistemic Relevance of Common Sense ............................ 171
§ i. Prefatorial Remarks ......................................... 171
§ ii. The Dogmatic Foundations of Epistemic Conventionality ........... 180
§1ii. Common Sense: the Separability of First and Second-Order Beliefs . . . . 187
§ iv. Objections and Replies ..................................... 200
§ v. Concluding Remarks ....................................... 206
Works Cited ....................................................... 214
Works Consulted .................................................... 216

CHAPTER I:
Introduction

In his Outlines of Pyrrhonism, Sextus Empiricus wrote: “Scepticism is an ability, or
mental attitude, which opposes appearances to judgments in any way whatsoever . . . . ”‘
The goal of this sceptical attitude was initially to reach a state of quietude; a state initially
thought to be achieved by discovering the truth. But upon subjecting every belief to
criticism, the Pyrrhonian discovered that he arrived not at certainty and quietude, but
rather, at uncertainty and disquietude.2 This uncertainty was a consequence of what the
Pyrrhonian inquiries did uncover: not truth, but a number of competing possible truths.
For every belief that P, that is, they found alternative beliefs that — P.3 Moreover, for each
of those alternative beliefs that — P, they discovered reasons and evidence. The reasons

and evidence for — P, they asserted, were “equipollent” to those asserted for P. These

reasons, that is, were equal with respect to their probability.

 

‘ Sextus Empiricus. Outlines Onyrrhonism, trans. R.G. Bury, (Cambridge: Harvard University Press,
1933), 7.

2 What they also found was that they arrived at a state of quietude, not through knowledge of the truth,
however, but through the recognition that the search for knowledge of the true nature of reality is futile,
and can end only in frustration and disquietude. Thus did they replace this search for knowledge with the
suspension of judgment about all matters “not evident.”

3 One here, 1 think, begins to see the origins of not only the relevant altemativist’s position, but equally, of
each sceptical and anti-sceptical effort to deal with the relevance of alternatives to one’s belief that P.
Indeed, as we shall see in chapters 11, III, and IV of this work, the question of the relevance of such
alternatives is at the heart of much of the twentieth century attempt to answer the sceptic.

l

The Pyrrhonian sceptic’s recognition of this state of evidentiary “equipollence”4 led the
Pyrrhonian to “suspend judgment” about P altogether:

The phrase “I suspend judgment” we adopt in place of “I am unable to say which
of the objects presented I ought to believe and which I ought to disbelieve,”
Indicating that the objects appear equal to us as regards credibility and
incredibility.s

They argued, therefore, that to assent either to P or to — P would be wholly dogmatic,

given P’s non-evident nature. Such assent is dogmatic, the Pyrrhonian argued, because

there is no criterion sufﬁciently justiﬁed by which we may separate the true from the false;6

no criterion, that is, according to which we can know with certainty which of our beliefs
are true. Why is there no criterion sufﬁcient for distinguishing the true from the false?
The Pyrrhonian summed it up this way:

In order to decide the dispute which has arisen about the criterion, we
must possess an accepted criterion by which we shall be able to judge
the dispute; and in order to possess an accepted criterion, the dispute
about the criterion must ﬁrst be decided. And when the argument
reduces itself to this form of circular reasoning, the discovery of the
criterion becomes impracticable, since we do not allow them [our
opponent] to adopt a criterion by assumption, while if they offer to
judge the criterion by a criterion, we are forced to a regress ad
inﬁnitum.7

 

‘ Outlines, 111.
5 Outlines, 1 15.

‘ The importance of this ancient skeptical argument is evidenwd throughout the history of
philosophy. Indwd, it was just such an objection which Gassendi offered against Cartesianism in the
17th century and which Descartes regarded as the “objection of objections.” CF both “The Fifth Set of
Objections,” in The Philosophical Works of Descartes, trans. Elizabeth Haldane and GT Ross,
(Cambridge: Cambridge University Press, 1955), and Richard Popkin, “The High Road to
Pyrrhonism,” American Philosophical Quarterly 2, (January 1965): l-15.

7 Outlines, 165.

Both the spirit of this sceptical attitude, as well as the consequences which follow from its '
application, have haunted knowledge claims ever since. Indeed, it is precisely the question
of such a criterion which is at the heart both of Descartes’ Meditations as well as the
criticisms leveled against it.8 Wherever there is uncritical assumption, there is also an
arsenal of alternatives, and the question of the criterion of the criterion. This plethora of
alternatives to P, together with the vexing problem of the criterion, is sufﬁcient to expose
the dogmatic nature of any claim to certainty.910 So unrivaled, in fact, is its power both to
disrupt and to throw into chaos all varieties of dogmatic thought, that after its long
absence during the medieval period, and upon its rediscovery (as well, of course, as the
translation of it into Latin and French) at the dawning of the renaissance, the Outlines of
Pyrrhonism (a work which catalogues all of the means of arriving at alternatives, as well
as the insoluble problems for distinguishing amongst them) gave rise to a crisis for all
presumed knowledge; a crisis for dogmatic thought unrivaled in all of history.ll

Upon the Outlines’ rediscovery, there ensued an unparalleled struggle between reason

 

8 See especially, Meditations I and 11, Trans. Elizabeth S. Haldane, and G.R.T. Ross (Cambridge:
Cambridge University Press , 1911).

9 Indeed, the mere possibility of such equipollent alternatives, combined, of course, with the absence
of any defensible criterion, has constituted the chief agnoilogical weapon of the sceptic from the
Pyrrhonians through Hume.

‘° It is interesting to note, that in spite of the now ancient origins of the regress argument (regarding
the criterion), it was precisely to deal with this problem (as inherited by Descartes) that contemporary
theories of knowledge such as coherentisrn, and contextualism were developed. And, as we shall see,
throughout the contemporary period it is dispute over the criterion which provides at least a major
part of the impetus for the development of relevant altemativism, whose central concern is just the
delineation of the relevant from the non-relevant alternatives to knowing.

“ For a very thorough presentation of the Pyrrhonian impact in the Modern period, see Richard
Popkin’s “The High Road to Pyrrhonisrn,” and “Scepticism and Anti-scepticism,” in Popkin, The
High Road to Pyrrhonism, Ed. Richard A. Watson and James E. Force, Indianapolis: Hacket
Publishing Company, 1980).

and belief. This struggle is everywhere evident: from the attack of the French intellectuals
such as Pierre Gassendi, Simon Foucher, and Samuel Sorbiere, on the “new Dogmatism”
of Descartes, to the complete unraveling of “the warp and woof of man’s intellectual
world”‘2 effected by Pierre Bayle’s sceptical Dictionary. ‘3 That struggle’s centuries long
exposition and disputation put an end (at least for the French and Germans) to the former
satisfaction achieved by resting the veracity of either one’s religious or one’s scientiﬁc
knowledge in the lap of speculative reason.” Indeed for the ﬁrst time in nearly a
millennium, dogmatic belief was the subject of relentless and philosophical scrutiny; a
Pyrrhonian scrutiny, the eﬁ'ect of which shook the foundations of western culture to a
degree from which it has not recovered.

What made possible the seismic impact of these ancient arguments? What made
possible the resulting destruction of man’s faith in reason? The modern response to the
arguments of the sceptic was by and large to concede that the sceptic established beyond
any doubt the lack of any foundation in reason for both religious and scientiﬁc knowledge
claims. It was widely acknowledged that by proving that choosing one alternative
description of reality’s nature rather than another has no rational basis, (owing to the lack

of any justiﬁed criterion) the sceptic established the impossibility of knowledge. 1’

 

'2 Popkin, “The High Road to Pyrrhonism,” 25.

'3 Perhaps the most signiﬁcant of which was the impact of Hume’s Pyrrhonism on the rest of the history
of philosophy through its inﬂuence on Kant.

" Again, Popkin identiﬁes the insipid reaction of the British - especially amongst the common sense
philosophers. See, for example, his “Scepticism and Anti-scepticism in the Latter Part of the Eighteenth
Century,” in Popkin, The High Road to Pyrrhonism.

'5 Such reasoning is echoed throughout the history of scepticism.

4

Furthermore, she presented what seemed an irrefutable case for the lack of any criterion
sufﬁciently justiﬁed to determine what is the case and to distinguish it from what is not.
What reason’s application to belief revealed, then, was not truth, but a plethora of
uncertainty. Reason, that is, had established the utter indefensibility of

judging reality to be one way rather than another.

As a consequence, the dogma of a millennium was either (though atypically) rejected in
total or was said to rest (necessarily) on faith. The sceptic’s ruthless and Pyrrhonian
criticism of the claims to lmowledge of belief’ s content resulted, therefore, not only, in the
overt recognition of reason’s inability to defend belief, but also in a proliferation of the
defense of convention. Notable, however, was the absence of convention’s defense by
reason. Equally notable was the concomitant proliferation of claims that only faith could
save belief. Throughout this neo-Pyrrhonian period the primary response to scepticism
was to separate what could be known through reason, from what could be known only
through faith. This is evident, of course, in Pascal, Montaigne, and the other great French
intellectuals of the time. Furthermore, though this strategy of defending the norm (the
common beliefs of the time) preserved to a certain extent much of what was dogmatically
assumed, it did so without any pretense to a rational foundation.“ The Modern defense of
belief acknowledged the agnoilogical power of the sceptical arguments, and the

consequent futility of giving a reasoned defense of belief. Faith, after all, not reason,

 

'5 Here, of course, I am excepting Descartes’ fantastic dogmatism, as well as that of other Modern
rationalists.

could alone save belief.l7

In fact, although nearly every sceptic of the Modern period was -- in spite of her
scepticism -- a devout believer, none pretended to justify or defend those beliefs through
human reason.18 David Hume, of course, represents the culmination of this Pyrrhonian
critique of judgment. Hume went beyond the sceptical conclusion that reason could not
ground belief to argue that belief itself is grounded in something other than reason.l9
Instinct, Hume argued, not reason, is the source and foundation of belief. Believing that
there is an external world, that there are other minds, that the future will be like the past, is
a matter of natural, not rational, necessity.

Nature by an absolute and uncontrollable necessity has required us to judge
as well as to breathe and feel.20

A fundamental consequence of beliefs natural necessity is that no sceptical argument,
however powerﬁil, can effect its abandonment. “Philosophy” Hume acknowledged,
“would render us completely Pyrrhonian, were not nature too strong for it.”21 Hume’s
insight into the necessity of belief is in fact recognized repeatedly in the history of

scepticism. Indeed, the necessity of belief in spite of reason’s opposition is evidenced in

 

'7 There were, of course, exceptions. One might think Descartes among the most notable of such
exceptions. Yet even he, in his response to Gassendi, for example, did not pretend to be able to
overcome the Pyrrhonian attack on the criterion; rather, he simply pointed to the devastating
consequences this attack would have for knowledge if taken seriously.

’8 Montaigne, Pascal, Foucher, Gassendi, Huet, and Sorbieri.

‘9 A point which leads Hume to disagree with the Pyrrhonians about the possibility that we can
suspend belief.

20 David Hume, A Treatise of Human Nature, ed., L. A. Selby-Bigge (Oxford: Oxford University
Press, 1988).

2‘ Hume, Treatise, 183-4

ancient Pyrrhonism, and pronounced unequivocally by Sextus.
And though the both the Academics and the Sceptics say they believe some
things, yet here too the difference between the two philosophies is quite
plain. For the word “believe” has different meanings: it means not to resist
but simply to follow . . . but sometimes it means to assent to a thing of
deliberate choice . . . . our belief is a matter of simple yielding without any
consent . . . .22

Adhering, then, to appearances we live in accordance with the normal rules
of life, undogrnatically, seeing that we cannot remain wholly inactive.23

In spite of the apparent conﬂict, then, between reason and belief; between the examined
and the unexamined; between scepticism and common sense, the sceptic (from Pyrrho to
Hume) has long recognized the impossibility of suspending belief altogether. Belief, after
all, is necessary to existence. (Without the belief that there is an external world, for
example, we would each surely perish. Without the belief in other minds, we would have
no reason for communication. Without the belief in the order of nature we would make
no plans.) The sceptic, therefore, has thematically and consistently maintained a
distinction between the natural, ineradicable, and indispensable, propensity and necessity
to believe (in spite of reason’s inability to confer certainty upon that belief), and the
eradicable and unjustiﬁed judgment regarding the knowability of those beliefs.
Accordingly, she has maintained that we must acquiesce to the natural necessity of belief,
but without the addition of indefensible and unnecessary judgments concerning our
knowledge of beliefs truth or falsity.

The sceptic, then, does not recommend suspension of mere belief regarding ordinary

 

”Outlines, 141.

”Outlines, l7.

matters (a suspension she would have regarded as impossible), but rather the suspension
only of judgment about the veracity of those beliefs For mere belief — of the sort that is
not accompanied by assent — is itself neither dogmatic nor suspendable. What is both
dogmatic and suspendable, however, (and, therefore, objectionable to the sceptic) is the
kind of “second-order” belief that makes pretensions to know what is believed. It is
judgment, that is, about whether we know the truth of those beliefs that has the sceptic’s
attention. For it is only in judgment, not mere belief, that one goes beyond what is
evident. We must distinguish, therefore, between our direct beliefs about the world (what
henceforth I shall refer to as ﬁrst-order beliefs) and our judgments about those beliefs
(what I call second-order beliefs, indicating thereby a kind of belief which is not about the
world, but rather, about beliefs about the world). The sceptic accepts both what she must
believe as well as the (universal) natural necessity of our unwillingness to give up those
(ﬁrst-order beliefs), but maintains all the while that no matter how unwilling (or incapable)
we may be to give up those beliefs, any claim to [mow what we believe cannot enjoy
rational defense, and never, therefore, can constitute knowledge. Why can our second-
order knowledge beliefs enjoy no rational defense? There is, again, no rationally justiﬁable
criterion for choosing our belief that P over its alternative. There is, that is to say, no
rational justiﬁcation for supposing that P is true while —-P is false.

In the twentieth century the natural necessity (and the consequent universality and
ineradicability) of belief has, however, been widely confused with beliefs epistemic
justiﬁcation. The distinction between judgments about those necessary beliefs (what I am

calling second-order beliefs), and beliefs of the ﬁrst-order has also been largely

overlooked. The necessity of believing that there is a well-ordered external world, for
example, which causes our perceptual experience has been confused with the unnecessary
(and unordinary) belief that we can know the world is thus and so. As a consequence,
contemporary anti-sceptics have not only misunderstood the true object of radical
philosophical scepticism, but also, therefore, have misunderstood the kind of argument
needed to refute such scepticism. This conﬂation and error is clearly evident in the
manner in which three of the most important anti-sceptical movements of the twentieth
century argue against the sceptic: it is evident in the ordinary language philosopher, the
relevant altemativist, and the contextualist’s attempt to refute the sceptic. Indeed, it is
evidenced not only in the particulars of their anti-sceptical arguments, but also in their
conviction that common sense is threatened by scepticism.

It is only on the basis of their failure to give appropriate consideration to the sceptic’s
insistence that mere ﬁrst-order belief about the world is necessary, not in itself
objectionable, and to be distinguished from the unnecessary and objectionable practice of
judging the veracity of those beliefs, that these anti-sceptical philosophers feel compelled
to defend common sense against the speciﬁc threat of philosophical scepticism. For,
correctly understood, these distinctions (which must be correctly understood if the very
.object of the sceptic’s agnoilogical moorings is to be identiﬁed) reveal that the sceptic’s
arguments do not threaten common sense, any more than they threaten the ‘googoos’ and
‘gagas’ of newborns. These basic ﬁrst-order beliefs are not in themselves of any interest
to the sceptic; for such beliefs are based not on reason but nature. Moreover, they are not

threatened, and cannot be threatened by scepticism, because they are not judgments. It is

only judgments whose epistemic credentials can be impugned. The object of scepticism is
not the natural ﬁrst order-beliefs which comprise the content of common- sense, but
rather, scepticism’s object is judgments; judgements about the accuracy of those (and
other) beliefs.24 An accurate understanding of what scepticism does and does not imply,
as well as an accurate understanding of the nature and contents of common sense, reveals,
therefore, the complete inanity and ﬁrtility of each of these contemporary anti-sceptical
refutations.

In the following chapters, I argue that neither these distinctions nor their consequences
have been clearly understood. Had they been understood, we should not in the last
century have had such a plethora of defenses of common sense, defenses erected against
the particular (and supposed) threat of philosophical scepticism. Equally, we should not
have had the particular variety of sceptical reﬁitation so widely embraced in the twentieth
century; that which rests its entire reﬁrtation on mere facts of ordinary behaviour, ordinary
belief, and ordinary language. For the sceptic, I argue, is not concerned with how we
happen ordinarily to behave, speak, or believe; at least not directly. Rather, she is
concerned with how we ought to behave, speak and believe; and then only with respect to
the second-order, for it is only here that epistemological criticism makes sense.
Consequently, how we speak and think is relevant only to the extent that it is revelatory of
the second-order.

Of course, the ordinary language philosopher, as well as her heirs, have assumed that

 

2‘ I can, for example, (not to mention, I most assuredly do) believe that there is an external world, a
world that it is the kind of world in which there are objects, people, and planets, and I can do so
without thereby believing that] know there is such a world.

10

how we happen to speak is relevant to the question of whether we have knowledge. So
too they have assumed that the ordinary is relevant to the question of scepticism’s truth or
falsity. Equally, therefore, they have assumed that the irrelevance of the sceptic’s
alternatives is afait accompli. All that remained to be done, they thought, was to explain
how exactly the sceptic manages to get our attention, to seem so nearly right. The
answer, they argued, lay in the manner in which the sceptic raises, manipulates, or
otherwise distorts the standards of knowing ordinarily in place.

The present work is in part a critical inquiry into the conditions which make possible this
particular variety of anti-sceptical reﬁrtation. I argue that these conditions reside not in
the nature of things, but rather in a number of philosophical (and dogmatic) assumptions.
If what follows achieves nothing else, I hope that it at least makes clear that there are a
number of clearly identiﬁable assumptions, and that these are necessary to support the
variety of anti-sceptical arguments I discuss. Moreover, because each of these assumptions
is neither necessary nor otherwise compelling, I argue they should be rejected.

My inquiry reveals that the kind of anti-sceptical inference made by the ordinary
language philosopher, the relevant altemativist, and the contextualist — from the ordinary
to the epistemic — rests on a number of misunderstandings. Among those
misunderstandings is the already mentioned assumption that the ordinary carries epistemic
and semantic relevance while the extraordinary does not (or, at least not much). It is this
assumption, for example, which allows not only the ordinary language philosopher, but
also the relevant altemativist, and the contextualist to conclude that the sceptic’s

departure from the ordinary sufﬁces to reﬁrte (for all practical purposes) her agnoiological

11

conclusions. This conclusion rests, however, on an additional misunderstanding of the
nature and contents of common sense. It rests on a misunderstanding of the kind of
standards that govern common sense. It rests also, therefore, on a misunderstanding of
the nature and signiﬁcance of the disparity between the sceptic’s queries and our ordinary
linguistic habits, the standards of knowledge she cites, and the relation of both to the
claims of common sense.

A further, and perhaps not less signiﬁcant, result of my inquiries is the discovery
that the relevance of the Cartesian alternatives to knowledge may be much greater than
these anti—sceptical views have acknowledged. Like the ordinary language philosopher,
the contextualist, and the relevant altemativist, however, I too conclude that the Cartesian
alternatives to knowledge are of little signiﬁcance for common sense, though not at all for
the reasons these others have given. Rather, it is because the presumed correlation
between defeating the sceptic and saving common sense is fundamentally mistaken. This
correlation is mistaken because it falsely assumes that the standards that govern common
sense are of the same kind as the standards that govern knowledge. It is mistaken,
moreover, because it falsely attributes to the contents of common sense second-order
knowledge beliefs which are, I argue, anything but common. It is mistaken because it
conﬁrses the kinds of beliefs we do ordinarily hold with the kind of beliefs that we do not;
while it is the latter alone which the sceptic means to confute. And because it is the latter
kind of beliefs the sceptic means to refute, common sense beliefs remain unscathed by
scepticism, even if the sceptic is correct about the kinds of justiﬁcation necessary for

knowledge.

12

There are, of course, different varieties of sceptical argumentation. Equally, there are
different varieties of anti-sceptical argumentation. In this work I am interested only in that
variety of scepticism which argues that because there are always inelirninable alternatives
to what we claim to know all knowledge is impossible. According to this sceptical
argument, the logical possibility of these alternatives, coupled with their inelirninability,
suﬂices to make them relevant. What kinds of alternatives? Alternatives like that I may
now be dreaming; or, that unbeknownst to me, I may be the unwitting victim of an evil
and cunning genius; or that, again unbeknownst to me, I may be an envatted brain having
the experiences I do have not because of an orderly external world, but rather, because of
the stirnulations my envatted brain receives from the super-computer to which it is
connected.

Of course, what each of these alternatives has in common is the fact that each describes
a logically possible alternative to what we ordinarily believe is the cause of our experience.
More importantly, each of these alternatives describes a state of affairs that we cannot
eliminate. Thus the alternatives of philosophical signiﬁcance are those that are not only
logically possible, but also, inelirninable. What makes these alternatives ineliminable is, as
the Pyrrhonian discovered long ago, the fact that we lack any criterion by which we may
judge the one to be the true cause of our experience rather than the other. Lacking any
such criterion, the sceptic concludes we must suspend our judgement regarding the actual
nature of the cause of our experience.

Just as there are a variety of kinds of scepticism, so also, of course, there are any

number of kinds of anti-scepticism. Again, however, I am not here concerned (nor would

13

it be practical to be concerned) to reﬁrte every sort of anti-sceptical argument. Rather, in
what follows, I conﬁne my attention to just one variety of anti-sceptical argument -— that
which infers from the sceptic’s mere linguistic oddity her epistemological irrelevance -
and to three of its particularly important representatives: the ordinary language
philosopher, the relevant altemativist, and the contextualist.25

There are three primary reasons I conﬁne my attention these kinds of anti-sceptical
argument: ﬁrst, each argues speciﬁcally against the kind of sceptical position I am
concerned with; second, each confuses the ﬁrst-order beliefs which are both necessary and
constitutive of common sense for the second-order beliefs which are neither; and third,
each appeals to the same kind of argument — the argument from ordinary language — in its
attempt to reﬁrte the sceptic and to save common sense.

Like the defenders of the religious norm during the renaissance and modern period,
these contemporary anti-sceptics felt compelled to safeguard the doxastic norm; they felt
compelled, that is, to safeguard the contents of common sense. Novel, however, to their
twentieth century approach, is their defense of what we ordinarily believe simply on the
grounds that it is ordinary. Of course, they do not argue explicitly from the mere fact that
our ordinary beliefs are ordinary that they must therefore be correct. Nonetheless,

ultimately the argument of each — surface complexities and dissimilarities aside — does

 

2slt might be thought that I should address Malcolm’s or Putnam’s radical anti-scepticism which
claims that scepticism is self-defeating, that the “dreaming” and “brain-in-the-vat” alternatives can’t
get off the ground. There are at least two reasons why I am not addressing this radical anti-
scepticism: ﬁrst, while I am conﬁdent their arguments, such as Putnam’s which assumes a causal
theory of reference, fall prey to the Pyrrhonian reply that they presuppose question-begging criteria, it
would take another thesis to do justice to their challenge. Second, the anti-sceptical adversaries I am
addressing do not regard these standard sceptical alternatives as self-defeating or self-refuting. In
fact, they take them seriously enough to advance extensive arguments against their relevance.

14

reduce to this form.

Chapters 11, III, and IV of this work are largely devoted to establishing that in spite of
the obvious surface differences in their arguments, the ordinary language philosopher, the
contextualist, and the relevant altemativist’s arguments do in fact reduce to the same kind
of argument. To establish this, I argue that each appeals to the same ﬁrndamental
principle in their anti-sceptical refutations. The principle to which they appeal states that
any challenge to a knowledge claim which violates ordinary linguistic practices is
irrelevant to the truth of that knowledge claim. In chapter 11, I lay the foundations for the
critique of this mode of argumentation. I there argue not only that this principle is
mistaken, but also, that to argue from mere linguistic oddity to epistemic irrelevance is
transparently fallacious. Why? Because to the extent that one assumes the ordinary is
revelatory of the epistemic, one conﬂates two kinds of standards which are in fact
logically distinct: the standards governing appropriate assertion, and the standards
governing true assertion.

In the second chapter I also give numerous illustrations that demonstrate that these
kinds of standards are in fact distinct, and must be distinguished if either common sense or
scepticism is to be clearly understood. I argue that to the extent that the standards are
distinct in kind, the proof that the sceptic violates the one (that of ordinary assertion) is
not sufﬁcient to prove she violates the other (of true assertion). And insofar as each of
these philosophers establishes only that she violates the standards of ordinary assertion,
there arguments do nothing to settle either the question of the truth or falsity of the

sceptic’s conclusions, nor do they settle the truth or falsity of the contents of common

15

sense. I conclude that one cannot reﬁrte the sceptic merely by pointing to the
conventionality of everyday belief, nor by pointing merely to the absence (or violation) of
conventionality in the alternatives to which the sceptic appeals.

With respect, then, not only to the relevant altemativists, but also to the contextualists
and to the ordinary language philosopher, the truth of the sceptic’s claim that knowledge
is unattainable comes down to a question of the relevance of her alternatives. Moreover,
the relevance of those alternatives comes down to how well they ﬁt in with our ordinary
linguistic practices. But why should their relevance be determined by our ordinary
practices? Why should a question of truth be determined by a matter of convention?

In my ﬁfth and ﬁnal chapter I argue that the entire project of protecting the ordinary
rests on a fundamental misunderstanding of what the ordinary entails. In particular, I
argue, it rests on the confused, but widely embraced, belief that the contents of common
sense include not merely ﬁrst-order beliefs about the world, but also second-order beliefs
about those beliefs. That is, it rests on the conviction (largely unexamined) that common
sense includes not only the belief that there is an external world, that one has hands, and
so on, but also the additional set of corresponding beliefs that each of these ﬁrst-order
beliefs is known to be true. For it is only if common sense does include such second-order
beliefs, such judgements about the knowability of the ﬁrst-order, that common sense is (or
can be) threatened by scepticism.

I argue, however, that this view of common sense is in error. One does not in fact
ordinarily believe, in addition to each of one’s ﬁrst-order beliefs about the world, that

each of those beliefs are known. Rather, one simply believes. Moreover, what concerns

16

the sceptic is not what we believe about the world, but rather, whether we can know what
we believe. To the extent, therefore that the ordinary man’s common sense does not
include such second-order beliefs, to the extent that it does not include beliefs about what
we can know, to that same extent are the contents of common sense protected from the
sceptic, and the refutations of radical scepticism (on behalf of common sense)
unnecessary.

Why, one might ask, would the likes of Austin, Lewis, Dretske, Goldman, and DeRose
be mistaken about the nature and contents of ordinary belief? How, after all, could one
get the ordinary wrong? I believe this mistake is explained by a philosophical confusion
over the kinds of standards which govern what we ordinarily say and do, and the kinds of
standards which govern truth. It is a confusion, that is, of the standards of good manners
with those of good truths. This confusion, I believe, is the result of a further confusion
between the strength of our natural convictions and the kind of certainty required for
knowing. Had due attention been given to the Humean (indeed the sceptical) insistence
on the natural necessity of (ﬁrst-order) belief, such conﬁrsion about both the object of
scepticism, and the sceptic’s conception of ordinary belief, might have been avoided. For,
again, it is not mere unavoidable belief about the world that the sceptic attacks. Rather, it
is judgement about belief. Only in judgement do we ﬁnd the kind of dogmatism that the
sceptic attacks. Indeed, from the earliest systematic scepticism, the object of scepticism
has been explicitly deﬁned as the dogmatic beliefs of the philosopher. Thus the object of

the earliest scepticism was directed at the claims of the Eleatics, the Pythagoreans, the

17

stoic, the Epicurean, and the Platonist, for example.26

The goals of this work are, therefore, threefold: ﬁrst, to establish that each of these major
anti-sceptical movements is reducible to an argument from ordinary language; second, to
establish that these arguments from ordinary language are fundamentally ﬂawed; and third,
to establish that because the ordinary language philosopher, the contextualist, and the relevant
altemativist misconstrue both the nature of common sense and the object of scepticism, they
fail to refute the sceptic and to save common sense.

In addition to these thesis, I conclude that even if we grant to the ordinary language
philosopher, the contextualist, and the relevant altemativist that common sense does include
second-order knowledge beliefs, the “knowledge” they allow us to claim is completely
unsatisfactory. For the kind of knowledge their theories afford us rests essentially on ignoring
what is difficult, presupposing what cannot be established by argument, and begging the very
question at issue. Their kind of knowledge is indistinguishable from dogmatic assumption.
It requires neither truth, nor justiﬁcation. This philosopher ﬁnds such knowledge deeply
unsatisfactory. For this philosopher still believes that knowledge is different from mere belief,
that knowing requires more than mere inter-subjective agreement, and that winning an
argument still requires the giving of good reasons. In so far as “knowledge” is (by their lights)
a matter of presupposing, ignoring, and ad hoc excepting, I conclude then that even if we can
“know” according to these contextually sensitive standards — such a victory for “knowledge”

is cheap indeed. For to overcome uncertainty by presupposing, ignoring, and ad hoc

 

”See, for example, the Outlines. Sextus Empiricus makes clear throughout the Outlines that the
subject of sceptical criticism is not those beliefs we are compelled to hold, but rather, those non-
compulsory beliefs held by the philosopher, whom he would call “the dogmatist”.

18

excepting is not to give a philosophical defense of knowledge, nor is it to refute the sceptic.
Rather, it is to re-embrace the very dogmatism — that is, assent to the non-evident claims —-

that centuries of philosophical scepticism has been designed to end.

19

CHAPTER II:
Meanings, Manners, and Scepticism

In this chapter, I examine the relation between meanings, manners, and scepticism. In
particular, I examine the effectiveness of the ordinary language‘ response to scepticism. I
argue that the ordinary language argument against philosophical scepticism is
fundamentally mistaken, and does not, therefore, reﬁrte scepticism. For the ordinary
language refutation2 of scepticism conﬁises good manners for good truths, and it is only to
the extent that one overlooks this confusion that one is tempted by its anti-sceptical
implications. Once, however, that confusion is made explicit, it is clear that the argument
from ordinary language poses no serious threat to —— let alone refutation of — the
sceptic.

The kind of philosophical sceptic I examine (I suppose there can be many) in what
follows is the one who (like Descartes at the end of his First, and the beginning of his
Second, Meditation) argues that there is nothing we can know: Not that the ﬂowers in
the garden are in bloom, not that there are birds in the garden, nor indeed that there is a

garden.

 

1What I intend by ‘ordinary language’ is somewhat narrow, and shall become clear shortly. Sufﬁce it
for now to say that I intend the following: the position that ordinary language is correct and can be used to
resolve philosophical problems; more speciﬁcally, with respect to scepticism, any proposed refutation of
scepticism which argues from the mere linguistic oddity of the sceptic’s remarks to their supposed
epistemological irrelevance or falsity. (As will become clear, there are reasons to include both the
relevant altemativist, and the contextualist within this genre.) My remarks throughout apply only to this
variety of ordinary language.

20

In particular, I am interested in the philosophical sceptic who argues for this conclusion on
the grounds of competing possible descriptions of the etiology of our experience. It is
the sceptic who argues, for example, that my experience of the ﬂowers in bloom may be
the result of a garden in bloom, but may also be the result of the machinations of an evil
and all powerful genius who delights in deceiving me. It is the sceptic who argues that my
experience of the birds in the garden may be caused by birds in a garden but may also be
caused by the stimulation received by my envatted brain from a super-computer to which
it is attached. It is the sceptic who argues that my experience of the garden itself may be
caused by a garden, but may also be caused by a dream I am now having.

But in thinking over this I remind myself that on many

occasions I have in sleep been deceived by similar illusions,

and in dwelling carefully on this reﬂection, I see so

manifestly that there are no certain indications by which we

may clearly distinguish wakeﬁilness from sleep that I am

lost in astonishment. And my astonishment is such that it is

almost capable of persuading me that I now dream.1
It is the sceptic who argues that because the content of my experience is equally
compatible with these alternatives, yet incompatible with what I ordinarily believe, I can
never come to know. This is the sceptic who interests me, and whose purported
refutation by ordinary standards shall be the focus not only of this chapter, but of the
chapters to come.

What is it about such competing possibilities that grounds the sceptic’s systematic

doubt? His conclusion that we cannot escape eternal ignorance? Why are these

 

3Descartes. rileditation 1, 146.

21

competing possibilities grounds for concluding that I — and you —— can never know?

The sceptic argument is this: knowing has certain necessary conditions. Among these
necessary conditions is the condition of certainty. Because certitude is a necessary
condition of knowing,2 it follows that if one is uncertain of what she believes, she fails to
know.3 Certainty is a condition of knowing that the ﬂowers in the garden are in bloom. It
is a condition of knowing that the birds are in the garden. It is a condition of knowing that
there is a garden. Indeed, certainty, the sceptic argues, is a condition of knowing anything
at all.

The sceptic, however, thinks that the existence of competing possibilities (the issue of all
likelihood or probability aside) undermines certainty. Why? What has certainty to do with
dreaming, vatting, and evil and cunning genius? The sceptic argues that certainty requires
the complete absence of doubt. Here it is useﬁrl to distinguish two kinds of doubt: on the
one hand, there is the doubt of the subject’s state of mind, what we might call ‘subjective
doubt’; on the other hand, there is the fact of the matter — whether there exist factors
which make something dubious — what we might call objective doubt. What kind of doubt
does the sceptic require be absent? Not mere subjective doubt, for it is easy enough to fail
to doubt what seems perfectly ordinary. Rather, it is objective doubt which cannot co-
exist with certainty, and objective doubt follows from objective possibility. Where

objective doubt exists, so also does (objective) uncertainty. And where uncertainty reigns,

 

‘I will not, in this work, look into the further issue of which conditions are sufﬁcient for knowledge.
My present concern is simply the sceptic’s claim that ruling out the Cartesian alternatives to one’s
knowledge claims is a necessary condition of knowing anything at all.

22

I suppose, then, that all the things which I see are false; I
persuade myself that nothing has ever existed of all that my
fallacious memory represents to me; I consider that I
possess no senses; I imagine that body, ﬁgure, extension,
movement, and place are but ﬁctions of my mind. What,
then, can be esteemed as true? Perhaps nothing at all,
unless that there is nothing in the world that is certain.2

In particular, I am interested in the philosophical sceptic who argues for this conclusion on
the grounds of competing possible descriptions of the etiology of our experience. It is
the sceptic who argues, for example, that my experience of the ﬂowers in bloom may be
the result of a garden in bloom, but may also be the result of the machinations of an evil
and all powerﬁrl genius who delights in deceiving me. It is the sceptic who argues that my
experience of the birds in the garden may be caused by birds in a garden but may also be
caused by the stimulation received by my envatted brain from a super-computer to which
it is attached. It is the sceptic who argues that my experience of the garden itself may be
caused by a garden, but may also be caused by a dream I am now having.

But in thinking over this I remind myself that on many

occasions I have in sleep been deceived by similar illusions,

and in dwelling careﬁrlly on this reﬂection, I see so

manifestly that there are no certain indications by which we

may clearly distinguish wakeﬁrlness from sleep that I am

lost in astonishment. And my astonishment is such that it is
almost capable of persuading me that I now dream.3

It is the sceptic who argues that because the content of my experience is equally

 

2Rene Descartes, Meditations on First Philosophy, Trans. Elizabeth Haldane and GT Ross, The
Philosophical Works of Descartes (Cambridge: Cambridge University Press, 1911), 149.

3Descartes, Meditation I, 146.

23

compatible with these alternatives, yet incompatible with what I ordinarily believe, I can
never come to know. lhis is the sceptic who interests me, and whose purported
refutation by ordinary standards shall be the focus not only of this chapter, but of the
chapters to come.

What is it about such competing possibilities that grounds the sceptic’s systematic
doubt? His conclusion that we cannot escape eternal ignorance? Why are these
competing possibilities grounds for concluding that I — and you — can never know?

The sceptic argument is this: knowing has certain necessary conditions. Among these
necessary conditions is the condition of certainty. Because certitude is a necessary
condition of knowing,‘ it follows that if one is uncertain of what she believes, she fails to
know.’ Certainty is a condition of knowing that the ﬂowers in the garden are in bloom. It
is a condition of knowing that the birds are in the garden. It is a condition of knowing that
there is a garden. Indeed, certainty, the sceptic argues, is a condition of knowing anything
at all.

The sceptic, however, thinks that the existence of competing possibilities (the issue of all
likelihood or probability aside) undermines certainty. Why? What has certainty to do with
dreaming, vatting, and evil and cunning genius? The sceptic argues that certainty requires
the complete absence of doubt. Here it is useﬁrl to distinguish two kinds of doubt: on the

one hand, there is the doubt of the subject’s state of mind, what we might call ‘subjective

 

‘I will not, in this work, look into the further issue of which conditions are suﬂicient for knowledge.
My present concern is simply the sceptic’s claim that ruling out the Cartesian alternatives to one’s
knowledge claims is a necessary condition of knowing anything at all.

24

doubt’; on the other hand, there is the fact of the matter - whether there exist factors
which make something dubious — what we might call objective doubt. What kind of doubt
does the sceptic require be absent? Not mere subjective doubt, for it is easy enough to fail
to doubt what seems perfectly ordinary. Rather, it is objective doubt which cannot co-
exist with certainty, and objective doubt follows from objective possibility. Where
objective doubt exists, so also does (objective) uncertainty. And where uncertainty reigns,
so also does ignorance.

What, then, is ‘objective possibility’? What is objectively possible? Anything, I should
think, which does not involve a logical contradiction. It is, for example, objectively
possible that I am now dreaming, however unlikely it may be either that I am, or that I will
(subjectively) seriously entertain the possibility. Equally, it is objectively possible (that is,
there is no contradiction in supposing) that I am an envatted brain, or a victim of an evil
and cunning genius. Indeed, no matter what my peculiar subjective response to these
possibilities is, they are possibilities. In order, then, for me to know what I believe, I must
be able to establish that I am not merely dreaming, not an envatted brain, and not a victim
of evil genius. Unless I can rule these out, I cannot be objectively certain. To the extent
that I cannot rule them out, these alternatives undermine my certainty. They undermine
my certainty because they are compatible with all my evidence and yet incompatible with
what I believe. They cannot be ruled out because from the contents of my experience I
cannot discriminate between the possibilities. For all I know, I am now dreaming, or a
brain in a vat, or a victim of ingenious deception.

There are, however, those who reject this harsh evaluation of our cognitive powers.

25

Among the most strongly opposed are the defenders of ordinary language. These
guardians of the linguistic norm regard the sceptic as more of a trickster than a serious
epistemological rival. The sceptic, they think, is seduced into disbelief by estrangement
from ordinary linguistic practices. They believe, moreover, that proving that it is
(perfectly) ordinary to say we know sufﬁces to refute the sceptic. For what is ordinary,
they argue, is manifestly correct.

In what follows, I offer a critical examination of this argument from ordinary language.
I argue that the ordinary language reﬁrtation of scepticism rests on the uncritical Principle
of Epistemic Conventionality: the principle according to which we may infer the
irrelevance of the sceptic’s doubts simply ﬁom their unconventionality.’ I argue that this
principle which constitutes the foundation of this kind of argument is ambiguous. On one
interpretation the principle is true, but ineffective; on the other interpretation, it is false.
According to that interpretation which makes it true, it is trivial. On the other hand,
according to that interpretation which makes it a weapon against scepticism, it is false. In
fact the principle seems effective against the sceptic only when one overlooks its
(illegitimate) conﬂation of two logically distinct sets of standards; the standards of
appropriate and the standards of true assertion.6

In this chapter, I offer numerous illustrations of the logical separability of these

standards, and conclude that the ordinary language advocate believes that mere facts about

 

’As I will argue in chapters III, IV, and V, the arguments of the relevant altemativist and the
contextualist also rest on this principle.

“My understanding of this distinction as well as the work it can do in defending the sceptic was
very much inﬂuenced by Barry Stroud’s The Signiﬁcance of Philosophical Scepticism (Oxford:
Clarendon Press, 1984): in particular chapter II.

26

ordinary language can refute the sceptic only because the ordinary language advocate
overlooks the logical independence of two kinds of standards. The ordinary language
advocate believes falsely that the sceptic’s challenges to our knowledge claims are
irrelevant to their truth, but only because she believes — also falsely —— that the
standards of appropriate assertion (which underlie our knowledge claims) are identical

with the standards of truth (which do not).

Section 1': The Sceptic

There are three essential elements in the argument of the (at least my) philosophical
sceptic: First, the sceptic claims that objective certainty is a necessary condition of
knowing; second the sceptic claims that ruling out those alternatives explanations of the
cause of our experience which are equally compatible with that experience is a necessary
condition of objective certainty; third, and ﬁnally, the sceptic claims that because we
cannot eliminate those alternatives (nor, therefore, distinguish the true from the false with
respect to these causal explanations), we cannot meet the conditions necessary for
knowledge. In other words, we cannot be objectively certain. The ineliminability of the
Cartesian alternatives undermines the possibility of objective certainty, and therefore, the
possibility of knowledge.

The sceptic, of course, at least the modern sceptic7, does not expect you to give up your

ordinary beliefs about the world ( that the world which is the cause of your experience is

 

7As a guide to living well, the ancient sceptics approach the issue of suspending belief somewhat
differently. See especially Sextus Empiricus, Outlines of Pyrrhonism, Book 1, Trans. R.G. Bury.

27

an external world, that the objects of your experience exist, and that you are among those
objects). The abandonment of ordinary belief is not the point of the sceptic’s reasoning.8
(And so, objecting to the sceptic on the grounds that it would be impossible to get by
without these beliefs is quite without point.) Rather, the sceptic expects only that you
abandon your judgement that you know what you believe; for that judgement, he argues,
must, inevitably, prove false. This is not to say that what you believe (e.g., that there are
trees outside your window) is false. Rather, it is only to say that your belief that you know
(if you possess such belief) is false, false according to the objective standards of knowing.
Take, for example, my claim to lmow certain things about my dog Beau. Beau, at the
moment, appears to be a real dog. Indeed, he appears to be a perfect specimen of the only
breed which can boast of being “gentlemen in white fur”; a Pyrenees, that is, of Siberian
ancestry and majestic appearance. And, as is characteristic of his breed, he has soft dark
eyes, and appears to be drooling (somewhat unmajestically) as he lounges about. The
question for the sceptic — the question the sceptic wishes us to entertain — is this:
appearances and beliefs aside, what do I know? Do I know these things which appear to
me to be so obviously the case? What would it take for me to know, for example, that I
have a dog, of Pyrenean origin, with soft dark eyes, and incessant (and unmajestic) drool?
Of course, undeniably —— indubitably — it seems perfectly evident that I have a dog, a

real dog, that his name is Beau, that he lounges about, and so on. Indeed, at times I even

 

8 One might think the sceptic does argue we should abandon belief. Strictly speaking, however,
the sceptic argues only against judgement about the veracity of belief, not belief in itself. For further
explanation of this often overlooked feature of the sceptic’s actual intentions, see chapters I and V.

28

say of myself that I [mow these things. (As, for example, when I am asked how I know he
is a real or pure-bred Pyrenees, I point to his double dew claws.) But what do I know?
Do I know, for example, that what appears now to be a dog is a dog? Do I know that I
have a dog? Do I know that I now sit in a room, and he in it with me? Do I know that the
clear liquid which appears to run from his mouth is drool? Real drool? Do I know his
double dew claws are real? Do I know this dog, this room, this drool, these dew claws are
not merely dreamt, vatted, or the results of computer stimulation? The sceptic argues I do
not — and can not — no matter how unwilling I may be to withdraw my beliefs. For
knowing, the sceptic reminds me, requires certainty. And so long as I cannot prove I am
not now dreaming, so long as I cannot prove that I am not now vatting, so long as I
cannot not prove that I am not now the unwitting victim of a cunning and evil genius,
certain is something I am not. Q.E.D.

Might it not be that the room I believe I am in is really only a ﬁgment of my, or
someone else’s, imagination? Might the room-appearance not be caused by a dream I am
now having? Might not this hypothesis explain my experience equally well? (Descartes’
familiar argument that there are no convincing signs distinguishing waking and dreaming
proves powerful here.) Or, might not the appearance of my room be caused by a super
computer which, in virtue of its connections to my envatted brain, stimulates this
appearance in me? How would the experience I am now having differ from the experience
I would have if either of these alternative possibilities obtained? The sceptic’s answer is:
not at all. Indeed, that is precisely the sceptic’s point, not to mention the well-spring of

the sceptic’s power.

29

And what of my perception of Beau, my (apparent) dog? I see what I take to be his soft
eyes, his clean white ﬁrr, and yes, his drool. But might not all of this too be caused by a
dream I am now having? “How often in sleep have I dreamt of Beau . . .”? Might not
these perceptions be the deliverances not of my eyes, but of my artiﬁcially stimulated and
envatted brain? The answer, of course, is yes. They might be. After all, these alternatives
are objectively possible.9 And moreover, and perhaps most importantly, each element of
my present experience would not differ, would not differ at all, in any detail, if it were
caused not by a real room, dog, eyes and drool, but by a dream, a computer, or an evil (a
very evil) genius. For, ex hypothesi, the content would — in each case — be precisely the
same; the same as it would be if the cause of the content were the things I take to be real.
I would perceive my familiar room, with all of its familiar ﬁrrnishings, the same
architectural ornaments. I would perceive the same white fur, dark eyes, and drool. And,
importantly, I would believe the same things.

From such objectively possible, and yet competing, perceptual equivalents follows the
doubt from which the sceptic derives her claim that we cannot know. Though, of course,
it is only if these competing alternatives are also ineliminable that they threaten my claims
to know; as noted, it is the fact that they are inelirninable that undermines my certainty,
since it is the extent to which I cannot know I am not now dreaming, vatting, or deceived,

that the certainty required for knowing cannot be mine.

 

9Because there is general agreement amongst the philosophers that I will discuss that — according
to the sceptic’s standards of elimination — these are both possible and ineliminable, I will not spend
time arguing that they in fact are. Instead, because the real question becomes whether or not objective
possibility and ineliminability stand in the way of knowing, my focus will be on this question, and how
one might go about answering it.

30

One may, of course, object that (at least some) such alternatives really are eliminable,
can be reﬁned by ordinary evidence, and hence cannot threaten what I (seem to) take
myself to know. Indeed, within the ordinary language camp, there are a number of such
protestations; Austin’s is a paradigm example. Distinguishing waking states from dream
states, he argued, is not mysterious at all. One need merely appeal to well established
methods: “There are recognized ways of distinguishing between waking and dreaming”. ‘0
Unfortunately, Austin does not tell us what these methods are, nor how they answer the
sceptic. One can assume, however, that the methods in question are the methods one
ordinarily employs when one checks to see if one’s child is asleep, or even if one is oneself
asleep (methods like seeing if one’s eyes are open, if one is dressed, if the sun is up, and so
on).

The problem, of course, is that one cannot assume the veracity of such methods without
begging the question against the sceptic. The sceptic, no doubt, will protest that using any
such methods to verify one’s belief that one is awake assumes the veracity of the very
thing in question, namely, one’s perceptions. One cannot -- without circularity -- appeal
to these perceptions to justify the conclusion that one is not asleep; these perceptions
might themselves be caused by dreams, vatting, or evil and cunning genius. In the absence
of a criterion by which one can prove one is not dreaming, vatting, or the victim of

ingenious deception in the ﬁrst place, it will be of no (non-question-begging) use to appeal

to these ordinary methods to show that one isn’t. It is, after all, precisely the ordinary that

 

loJ.L. Austin, “Other Minds”, J.L. Austin Philosophical Papers, ed. J .0. Urmson, and G.J.
Warnock, (Oxford: Oxford University Press, 1979) 87.

31

is in question.

Moreover, long before the likes of Descartes, the ancient Greeks”, recognized the
problem of the criterion of the criterion.12 Even if one had a criterion for ruling out having
merely dreamt, or believed by deceit, how could one know this criterion is not itself the
result of dreaming, or deception? The problem is quite obvious, and answering it lands us
either in an inﬁnite regress, a circular justiﬁcation, or an arbitrary stopping point.13 As
Sextus Empiricus states in his Outlines:

. . . he who prefers one impression to another, or one
‘circumstance’ to another, does so either uncritically and
without proof, or critically and with proof; but he can do
this neither without these means (for then he would be
discredited) nor with them. For if he is to pass judgement
on the impressions he must certainly judge them by a
criterion; this criterion, then, he will declare to be true, or
else false. But if false, he will be discredited; whereas if he
shall declare it to be true, he will be stating that the criterion
is true either without proof or with proof. But if without
proof, he will be discredited; and if with proof, it will
certainly be necessary for the proof also to be true, to avoid
being discredited. Shall he, then, affirm the truth of the
proof adopted to establish the criterion after having judged
it or without having judged it? If without judging, he will
be discredited; but if after judging, plainly he will say he has
judged it by a criterion; and of that criterion we shall ask for
a proof, and of that proof again a criterion. For the proof
always requires a criterion to conﬁrm it, and the criterion

 

”Sextus Empiricus, Outlines, Book IV, 67-7. Since Sextus was primarily a compiler of the sceptical
arguments of others, this argument predates the Outlines.

”Stroud offers an effective discussion of essentially this problem in his Signiﬁcance of
Philosophical Scepticism, though he leaves its ancient origin out of the discussion.

13For further reading here, the debate between the foundationalists and coherentists is of course
relevant. See, for example, Lawrence Bonjour, “A Critique of Foundationalism,” in The Theory of
Knowledge: Classic and Contemporary Readings. Ed., Louis P. Pojman (Belmont, CA., Wadsworth,
I993).

32

also a proof to demonstrate its truth; and neither can a proof
be sound without the previous existence of a true criterion
nor can the criterion be true without the previous
conﬁrmation of the proof. So in this way both the criterion
and the proof are involved in the circular process of
reasoning, and thereby both are found to be untrustworthy;
for since each of them is dependent on the credibility of the
other, the one is lacking in credibility just as much as the
other. 1‘

Even if Austin (and Moore, and Bouwsma,ls and countless others) were correct in
thinking that to distinguish wakefulness from sleep is a simple, ordinary, matter, what
about vatting ? What about an evil and cunning genius? Surely these are no ordinary
matters. What kind of criterion could possibly enable me to distinguish these possible
causes ﬁ'om the causes I think are responsible for my experience? The evidential states are
identical. For, ex hypothesi, my vision of Beau can be cleverly produced in my envatted
brain by a super computer. Even the calm look of his soft dark eyes, the clean white ﬁir,
even the drool. All of these perceptions might be caused by clever wiring which
stimulates precisely those brain states (and thereby simulates the experiences) I would
have if Beau were a real dog with real drool. Likewise, each of these perceptions might
be brought about in me through the clever machinations of an evil genius, leaving me no

perceptual or evidential difference; might, that is, in the sense that there is no logical

contradiction involved in supposing it to be so.

 

l‘Sextus Empiricus, Outlines Book 1, 68-9. (This argument is also found throughout the Outlines.)

15See, for example, G.E. Moore, “Proof of an External World,” G. E. Moore: Philosophical Papers
(New York, Collier Books, 1962); and OK. Bouwsma, “Descartes’ Evil Genius.” Bouwsma ’s
Philosophical Essays (Lincoln, University of Nebraska, 1965).

33

Consequently, if a man can prefer one impression to another
neither without a proof and a criterion nor with them, then
the different impressions due to the differing conditions will
admit of no settlement; so that as a result of this Mode also
we are brought to suspend judgement regarding the nature
of external realities“

Section ii: The Argument From Ordinary Language

Admittedly, the sceptic’s claim that knowing that we are not dreaming, vatting, or
victims of evil and cunning genius is a necessary condition of knowledge is odd. In fact,
of course, it seems perfectly absurd. Is it not absurd to think that I fail to know of the
existence of my dog because it is possible that I might be deceived in one of these ways?
Because my perception of him might be caused by a super-computer, or an evil genius?
How often, after all, do I thus scrutinize my evidence before I give Beau his daily treat?
How often do I stop to consider the competing possible etiologies of my experience
before I wipe away his drool? How often does anyone scrutinize the basis of her certainty
before claiming to know? Absurd indeed! Indubitably so.

In what follows I examine the view that the sceptic’s hypothesis is absurd; her
hypothesis, that is, that because I might now be deceived in some undetectable way, I
cannot know. Speciﬁcally, I examine the inference ﬁom the sceptic’s mere absurdity to
the conclusion that her assertions must also be irrelevant or false. The proponent of this
view argues that because of the degree to which the sceptic violates ordinary language her

doubts are irrelevant. Moreover, she claims, the

sceptic’s violation of ordinary language is sufficient to prove her claim that we fail to

 

l“Sextus, Outlines, Book 1, 69.

34

know is false, her doubts innocuous. ‘7

J. L. Austin’s “Other Minds” is paradigmatic of the kind of anti-sceptical argument I call
the argument from ordinary language. I shall, therefore, refer to his ‘reﬁrtation’
throughout. Austin argues that doubting our ordinary claims to know, for Cartesian-like
reasons, is inappropriate. More importantly, he argues that the sceptic’s raising of the
Cartesian alternative possibilities to what we know is itself inappropriate. Austin supports
this contention with some observations about our ordinary linguistic practices. Of
particular interest are his remarks about what ordinarily happens when one makes a claim
to know.

Austin observes, for example, that ordinarily when a claim to know has been made, a
number of rules or conventions are observed. The claim to know that “There is a
goldﬁnch in the garden,” for example, is one such knowledge claim, and — ordinarily —
when it is made, a number of linguistic conventions are observed. For instance, unless one
has some speciﬁc reason to doubt what has been claimed, one accepts it. And, moreover,
one does not doubt unless one has in mind, “and can specify,” some deﬁnite lack.18

If, for example, I have a reason to doubt your claim to see a bird in the garden, I might

point out that the evidence you offer (e.g., that what you perceive has several

 

”Representatives of the view I have in mind here include Austin, Wittgenstein, and Malcolm. No
one of them individually represents this view in whole. (Austin, for example, does not explicitly state
that the sceptical thesis is false, at least not in the material we look at, though be strongly suggests, if
not implies, this. And Wittgenstein too would hesitate to say that this view is false. Both he and
Austin would likely argue it is senseless, rather than false. Malcolm, on the other hand, explicitly
states this view is false. Yet the spirit of what I here call the ordinary language argument is an
essential part of the work of each.

l“Austin, “Other Minds,” 84.

35

characteristics unique to the Carduelis) is compatible with the perceived things being a
species of large moth, or some other winged non-Carduelis. Moreover, having speciﬁed
such a deﬁnite lack in your evidence, if you go on to explain why the perceived entity
could not be other than you said, I would not — ordinarily -— go on to insist on more.
Instead, I would recognize that “enough” evidence had been given when, “relative to
current intents and purposes,”the evidence “le[aves] no room for an alternative competing
description of it.”19 Hence, if in response to my query, for example, you cite your training
in omithology, your specialization in distinguishing small birds from other winged non-
birds, and your extensive ﬁeld experience, it would be unreasonable and inappropriate for
me to protest: “That is not enough.” As Austin observes, “Enough is enough. It doesn’t
mean everything. . . . It does not mean enough to show it isn’t a stuffed goldﬁnch,” at
least, not with respect to the conventions governing ordinary linguistic situations.20
Suppose, with these conventions in mind, that you and I travel abroad to England.
Suppose, moreover, that while there we attend a garden party. In the garden, the trees
appear to be in bloom, and several small species of birds appear to be perched about the
foliage. Austin’s view is this: were you to exclaim, in this context, “Look, there is a
goldﬁnch in the greenery,” my most appropriate response — assuming my familiarity with

the conventions governing ordinary discourse — in the absence of any speciﬁc reason to

 

19Austin, “Other Minds,” 84.This appeal of Austin’s to rules implicit in ordinary conversation, as
well as his emphasis on context, anticipates the appeal of many later Contextualists. See, for
example, my discussion of Lewis in chapter 111.

20Austin, “Other Minds,” 84. (To Rich Hall I owe the following: Austin appears here to anticipate
the denial of deductive closure; a denial found in a number of later drinkers; Nozick and Dretske, for

example.) Austin also anticipates the contextualists and the relevant altemativists insofar as he argues
against scepticism on the grounds that some alternatives to our knowledge claims are not relevant.

36

doubt your knowledge claim, is to express my agreement with you, or at the very least, to
refrain from antagonistic challenge. I might, for example, say “Indeed!” Or, not myself
being familiar with this particular species of British bird, I might query, “How do you
know?” Or, “How can you tell?” Or, “How do you know it is not a woodpecker?” In
either case, my response is perfectly ordinary, and, therefore, in making it I violate no
obvious linguistic norms. Austin would approve.

The sceptic, however, asserts that neither you, nor 1, nor anyone else can know anything
at all. The sceptic reasons that you, for example, cannot know any empirical proposition
because certainty is a condition of knowing, a necessary condition, and so long as you do
not know you are not now dreaming, vatting, or deceived by a strange and malevolent
genius, certain is something you are (certainly) not. Certainty requires the ability to rule
out such alternatives, but -- lacking a criterion -- you cannot: whatever your evidence that
there is a ﬁnch in the foliage, that same evidence supports the hypothesis that you are now
dreaming that there is a ﬁnch in the foliage, drearrring that you and I are in a garden,
dreaming that you are an ornithologist. Equally, your evidence that there is a ﬁnch in the
foliage supports the ﬁrrther hypothesis that your brain is envatted. So also it supports the
hypothesis that you are merely the unwitting victim of an evil genius of the Cartesian
variety.

Suppose that the sceptic appears at the garden gate. Suppose ﬁrrther that, upon
approaching, the sceptic overhears our discussion concerning the goldﬁnch and your claim

to know. How might the sceptic respond? It is quite improbable that the sceptic would

37

accept the literal content21

of your claim to know there is a goldﬁnch in the greenery, since
you imply, thereby, that you are certain it is so. Rather, one imagines the sceptic,
Pyrrhonian opposer that she is, querying thus: “Are you certain it is a real goldﬁnch?”
“After all, for all you know (‘which’, slyly she murmurs, ‘is nothing’), you might be
dreaming you see a goldﬁnch, you might be vatting you are an ornithologist, you might be
the unwitting victim of evil and cunning genius who deceives you into believing you are in
a garden, when all the while you are not.” In articulating her view that you cannot be
certain that what you perceive is a real goldﬁnch, the sceptic here plants the seeds of
doubt. For, even though you may steadfastly adhere to your belief that you do perceive a
goldﬁnch, a real goldﬁnch, and that you are an ornithologist, it is clear that you do not
know it is so.22 As the sceptic argues, clearly you are not certain it is so, and cannot be
certain unless you can also rule out the alternative possibilities regarding the etiology of
what you now experience. Lacking certainty, you also lack knowledge; necessarily so.

What should we make of the sceptic’s contention that your claim to know is ill-made? It
seems that the reservations which ground her uncertainty differ from the more ordinary
doubts expressed, for example, by my suggesting that the bird is perhaps a woodpecker,
rather than a goldﬁnch. It is quite unremarkable for someone to challenge your

knowledge claim by suggesting that perhaps you have mistaken one variety of bird for

another. Contrariwise, to challenge your claim to know on the grounds that you may be

 

2|The intended content has rather different implications from the literal content. This distinction is
covered to some extent in Chapter IV’s discussion of Dretske, but will be taken up in greater depth in
chapter V.

22Whether you ought to claim that you know this to be so is a different matter.

38

dreaming, or vatting, or deceived by an evil and cunning genius, is extraordinary. Indeed,
Austin suggests (as do the later contextualists and relevant altemativists), the degree to
which the sceptic’s alternative hypotheses depart from the norm suﬁices to make them not
merely unusual, but irrelevant and false.

Austin reminds us that “knowing it is a real goldﬁnch is not in question in the ordinary
case.”23 In the ordinary case — in the context of an ordinary conversation — one does
not doubt the reality of a purported perception. One does not raise the possibility that the
perception might be unreal. One certainly does not raise the possibility that the purported
knower might be dreaming, vatting, or the victim of an evil genius. But why not? Why
does one not raise such doubts in contexts like the garden party? And, furthermore, of
what signiﬁcance is it that one ordinarily does not raise such doubts?

With respect to the ﬁrst question, why one ordinarily does not raise such doubts, Austin
maintains that one does not raise these kinds of doubts because doing so would violate a
number of linguistic conventions. What are these conventions? Austin argues that doubts
concerning the reality of a perception — your perception of the goldﬁnch, for example, or
my perception of Beau -— can appropriately be raised (and seemingly therefore relevant“)
only when two conditions are met:

First, the reason for doubting must describe a possible state

of affairs.
Second, there must be a reason to think the possibility

 

23Austin, “Other Minds,” 88. (My emphases.)

2"The question of the relation of epistemic relevance to appropriateness is somewhat ambiguous in
Austin, hence I say “seemingly”. I address the importance of this ambiguity to the ordinary language
refutation in section in of the present chapter.

39

described actually obtains.”
These conditions are dictated by the norms of ordinary linguistic behavior.

Irnportantly, meeting the ﬁrst condition is not (always) sufficient for meeting the second.
Thus, though it may be possible that your perception of the ﬁnch in the garden is a causal
result of stirnulations to your envatted brain, there may be no reason to think it is so.
Likewise, though it may be possible that my perception of Beau’s soft dark eyes is the
causal result of a dream I am now having, there may be no reason to think it is so. In the
absence of such reasons, however, convention decrees that any challenge to my
knowledge claims based on these Cartesian possibilities is inappropriate. Hence when I
claim to know that Beau is near, you ought not, unless you have reason to think it so,
challenge my claim on the grounds that I may be deceived in one of these ways.

The sceptic’s suggestion that you might be dreaming, vatting, or deceived by an evil
genius, can be raised only if she has a reason to think it is so.26 She cannot
(appropriately) challenge your knowledge claims on the grounds of mere possibility
alone. Thus, if she is to say that you do not know that it is a (real) goldﬁnch causing what
you perceive, because it might be stuffed, she must have a reason to think the goldﬁnch is
stuffed. Indeed, Austin seems to imply, the sceptic’s raising of the more important and

speciﬁcally Cartesian possibilities that you may be dreaming, vatting, or in some other way

 

2’Austin, “Other Minds,” 87 . See also 98-99. I should note here that with these two conditions,
Austin anticipates central elements of both the Contextualists as well as the Relevant Altemativists.
See, for example, my discussion of Stine in chapter IV, and also of Dretske’s “Pragmatic Dimension
of Knowledge,” also in chapter IV.

26Austin, “Other Minds,” 98-99.

40

deceived, are irrelevant to your claim to know; irrelevant, that is, unless she can specify
some reason to think that you are in some such way deceived.

What follows, it seems, is that unless the sceptic can produce a reason for doubting
(beyond mere possibility) the genuineness of your perceptual experience, she cannot
challenge your claims to know: not your claim to know that the ﬂowers in the garden are
in bloom, nor your claim to know that the birds are in the garden, nor indeed your claim to
know that there is a garden, a garden in which all this “blooming buzzing confusion” takes
place. She is, Austin thinks, effectively silenced; silenced by the conventions which govern
appropriate assertion.27

Return for a moment to the garden. Suppose we are sitting amongst what appear to be
diverse species of trees, other ﬂora, and birds. The lighting appears to be adequate. Our
vision is good. We’ve not had much to drink. What reason could the sceptic have to
think we might be dreaming? What reason could she have to think we might be vatting?
What reason could she have to think we might be victims of evil and cunning genius? It
appears she has none. Surely, at least, that is Austin’s opinion.28 Yet, in the absence of
some such reason, her peculiar doubts cannot be raised.

That the sceptic has, in the garden, no reason to think our experience is phoney in any of

these ways seems true. I have, after all, deﬁned the situation as ordinary, and if the

 

27The question of whether she is not only silenced, but refuted, is considered in the next section of
this chapter.

2“Of course, if the sceptic is correct in describing the sceptical alternatives as equally compatible
with all the evidence from experience, then it may be that we at least have as much reason for
believing we are dreaming, vatting, or deceived by an evil genius, as we do for believing the cause of
our experience is an external world much like the one we ordinarily believe is the actual cause of our

experience.

41

situation is truly ordinary, it will not (it seems) be the case that there is reason to think we
are dreaming, vatting, or deceived. An ordinary situation, by deﬁnition, seems to be one
where such concerns are absent. What follows, in Austin’s opinion, is that ordinarily —-
which is most of the time — the sceptic will lack (conventional) justiﬁcation for raising her
peculiar — and most extraordinary — doubts.

Can she raise them elsewhere?29 Can she, for example, raise her doubts where
circumstances suggest she has reason to think these alternative possibilities might obtain?
Suppose, for example, that you and I, having had a bit too much to drink, ﬁnd ourselves
slumbering on the garden benches. And suppose that upon waking in the garden, I
exclaim, “That goldﬁnch is bothering me,” implying, of course, that there is a real
goldﬁnch nearby, and (seemingly) that I believe that I know this. The sceptic, returning to
the garden and overbearing my exclamation, might query thus: “Are you sure you are fully
awake? You had rather a lot to drink last night. Perhaps you have just been dreaming of
a goldﬁnch?” And, she might continue, “I assure you, I heard no goldﬁnch as I
approached, but only the faint whimper of what seemed to be a slumbering voice.”

Applying Austin’s conventional criteria of appropriate doubt in these circumstances
reveals that the sceptic’s doubt (about whether I might merely be dreaming) is — in the
face of my recent intoxication and slumber — appropriate (and, seemingly, therefore
relevant). For these latter conditions — my recent intoxication and slumber — give her

reason to think I might have been dreaming. Indeed, in this case, her reason to think I

 

29One must be clear to distinguish here the question of whether she can raise such doubts from the
question of whether she should raise such doubts. Austin does not do this, and this is a serious
omission on his part. More on this forthwith.

42

might have just been dreaming is the reason she raises the possibility.” Likewise, if upon
awaking in the garden, we learn that our host — heretofore unknown to us — is an avid
taxidermist who collects and stuffs goldﬁnches to display in his greenery, the sceptic’s
query regarding our certainty that the thing we identiﬁed last evening as a (real) goldﬁnch
was not a stuffed goldﬁnch may be appropriate. After all, our somewhat peculiar host is
in the habit of scattering stuffed goldﬁnches about his garden. Here again, it is not merely
possible that we may have been deceived in our perceptions, there is some reason to think
it so. This being so, the sceptic falters not in doubting our claims to know.31

There exist, then, contexts in which (at least some of) the sceptic’s doubts about our
claims to know can be appropriately raised.32 What Austin suggests, however, is that the
contexts — such as those described - where it is appropriate for the sceptic to raise
these speciﬁcally Cartesian doubts are never present in the ordinary case. In particular,

although in the ordinary case it is always possible that I may be dreaming, vatting, or the

 

30So as not to mislead or misrepresent the sceptic, I should mention that though according to
Austin’s criteria, the sceptic’s doubts are appropriate in this context, this is precisely the kind of
context in which a global sceptic would not raise her doubts. At least, it is not the kind of context
from which she would derive her universal doubt. The sceptic’s power lies in ﬁnding reason to doubt
even in the ideal cases, though the a-typical cases are important in so far as they play a role in
revealing the objective conditions of knowing, a role Austin (as well as the contextualists and the
relevant altemativists) fails to note.

31Though in the case of the stuffed ﬁnch I might, of course, be able to discover the ﬁnch is stuﬁ’ed,
unlike the case where I rrright be dreaming, there is an important relation between these kinds of
cases; the conditions of the possibility of certain doubts ever being relevant to my claims to know.

32Of course, conﬁning the sceptic’s doubts to the atypical misses the point of the sceptic’s doubt.
For her power resides in ﬁnding typical cases, actually better than typical cases, where perceptual
conditions and the like are ideal, and then showing that even under such ideal conditions, one cannot
know. Nonetheless, even when conﬁned to atypical cases, the sceptic’s doubts have universal
signiﬁcance, for the fact that her doubts are ever relevant reveals something important about the
general conditions of the applicability of our concept of knowledge.

43

victim of evil and cunning genius, ordinarily (which is to say most of the time) there is not
any reason to think that I am. What follows is that it is inappropriate (most of the time)
for the sceptic to challenge my claims to know. Hence the fact that there are always a
number of possible competing descriptions of the etiology of our experiences does not —
contrary to what the sceptic has argued — undermine our knowledge claims. At least, it

cannot prevent us from saying that we know, most of the time.33

Section iii: Disentangling: a conﬂation of manners and truth

Austin argues that the sceptic’s Cartesian doubts are, more often than not,
inappropriate. He argues that the two conditions of appropriate doubt are not often met.
Austin’s conditions, again, are that the reason for doubt must express a possible state of
affairs, and that there must be a reason to think the state of affairs expressed by that doubt
obtains. The problem for the sceptic, then, is that although the ﬁrst condition may be met,
this second condition is not. In ordinary circumstances, there is no reason to think of you
or me that we might be dreaming, vatting, or deceived by evil and cunning genius. What
follows, Austin argues, is that it is inappropriate for the sceptic to raise her doubts against
our ordinary claims to know. Her doubts are inappropriate in all but the most unusual

circumstances (e.g., when the subject of the purported knowledge is inebriated, or half

 

33Austin’s satisfaction with this silencing of the sceptic rests on (among other things) a
misunderstanding, or overlooking, of the transcendental signiﬁcance of the applicability of the
sceptic’s doubts to such cases. Though I must reserve any sufficient argument about this for another
place, the mere fact that the sceptic’s doubts are at times applicable suggests that the requirement of
certainty may be built in to our concept of knowledge.

44

asleep). Such doubts cannot be raised appropriately unless conditions (either of the
perceiver or the environment) render not only possible, but to some extent likely, that one
or more of the competing etiological descriptions is true.

One must concede that the conditions described by Austin as necessary for the
appropriate assertion of doubt (hereafter, the conditions of appropriate assertion) do
represent how we ordinarily expect others to behave in the context of ordinary
conversation. Therefore, one must also concede that the sceptic behaves inappropriately
when she raises these Cartesian challenges in the midst of ordinary conversation. As I
confessed earlier, her doubts are contextually absurd. What, after all, could be more
absurd than to challenge my claim to know that my dog exists on the grounds that it is
logically possible that I may be a mere envatted brain? Surely the unlucky individual who
puts forth such an assertion is likely to wind up in a straight-jacket.

And how could I deny that I these hands and this body are mine, were it
not perhaps that I compare myself to certain persons, devoid of sense,
whose cerebella are so troubled and clouded by the violent vapours of
black bile, that they constantly assure us that they are kings when they are
really quite poor, or that they are clothed in purple when they are really
without covering, or who imagine that they have an earthenware head or
are nothing but pumpkins or are made of glass. But they are mad, and I
should not be any less insane were I to follow examples so extravagant.”

Nonetheless, one must be cautious. One must be cautious not to infer from the
sceptic’s absurdity more than one is entitled. One must be cautious , that is, not to infer

too much from the fact that the sceptic’s behavior is inappropriate. For, though the

sceptic’s behavior — her articulation of Cartesian doubt — may be absurd, may violate

 

3‘Descartes, Meditation I, 145. (le emphases)

45

the conventions which govern appropriate assertion, it does not follow, at least not
straightaway and without further argument, that her conclusions are false, irrelevant, or
epistemically misguided.

It is here —-— with respect to the epistemic relevance of the sceptic’s doubts, and with
respect to the truth or falsity of her conclusions -— that one must consider the second
question: what is the significance of the sceptic’s absurdity? What is the signiﬁcance of
her (or anyone else’s) violating the conventions governing appropriate assertion? And, in
particular, what —— if any — is the epistemic signiﬁcance of so violating these standards?

Austin has argued only that the sceptic’s doubts are inappropriate, since — when raised
—— they are not supported by any reason to think they obtain. Asking for the epistemic
signiﬁcance of the sceptic’s violation is another way of asking how Austin’s observations
about ordinary linguistic practices are supposed to refute the sceptic. The sceptic, after
all, is concerned with knowledge, not —— it seems — appropriate (linguistic) behavior or
manners. Moreover, Austin, it seems, wishes not merely to comment on the sceptic’s
manners, but to reﬁrte her.

Without further argument, argument establishing a link between what is linguistically
inappropriate and what is false, however, the sceptic remains unrefuted; whether ill-
mannered or not.” It is precisely here that Austin’s anti-sceptical suggestions go too far,
beyond the evidence of ordinary linguistic behavior and convention. For, though such

ordinary convention may, in all but the most extraordinary contexts, effectively silence the

 

35Equally problematic is any inference from what is linguistically appropriate to what is true. I will
discuss this at length in chapter V.

46

sceptic, they do not prove her false. They silence, but they do not refute.

Consider what reasons Austin has given for rejecting scepticism. Austin has stated what
are the standards for appropriately raising doubts to another’s claim to know. He has
observed that in ordinary conversation two standards are conventionally observed. He has
argued that the sceptic violates these standards of appropriate assertion when she
challenges our ordinary knowledge claims. From this violation of conventional standards,
however, Austin infers also the irrelevance of the sceptic’s doubts. But, irrelevance with
respect to what? Irrelevance in what sense? Are the sceptic’s hypotheses irrelevant to
the truth of our knowledge claims? Or to something else? In what follows, I argue, it is
only the latter, not the former, which Austin can justiﬁedly infer.

Austin’s conclusion — that the doubts raised by the sceptic are irrelevant — is
somewhat ambiguous, and derives its ambiguity from the equally ambiguous principle on
which it rests: If a challenge to a knowledge claim violates, when raised, the conventional
standards for appropriate assertion, it is irrelevant to that knowledge claim.36 Let us call
this principle the Principle of Epistemic Conventionality. Straightaway, this principle
begs for an answer to the question: irrelevant in what sense? Irrelevant to what? There
are two reasonable interpretations of this principle. There are also, therefore, two
reasonable interpretations of Austin’s conclusion that the sceptic’s doubts are irrelevant to

our ordinary knowledge claims. What I ﬁnd, however, is that each interpretation proves

 

36In this, we shall see, Austin anticipates the central arguments of both the later Contextualists as
well as the Relevant Altemativists. Their use of this Principle is discussed in chapters III and IV of
the present work.

47

incapable of reﬁtting philosophical scepticism.

The ﬁrst interpretation of the Principle of Epistemic Conventionality follows
immediately ﬁ'om the evidence Austin gives for concluding that the sceptic’s doubts are
absurd. According to this interpretation, when it is asserted that the doubts of the sceptic
are irrelevant, on the grounds that they violate conventional standards for appropriate
assertion, their irrelevance is conﬁned to the context of good manners. The sceptic’s
violation, that is, makes her comments embarrassing, annoying, and irrelevant to the goals
of ordinary conversation (goals which are generally limited to following convention and
effectively communicating), goals aided by observing the (conversational) manners each
duly educated person learns as she acquires the English (and, presumably, any other)
language.37

If this ﬁrst interpretation of the principle is the correct interpretation, Austin’s
‘refutation’ of the sceptic is seriously deﬁcient. For such an interpretation of the principle
does no more than to bar the sceptic ﬁom speaking at (and perhaps attending) parties. It
silences her, but it does not refute her. Refutations concern truth and falsehood.
Refuting someone requires showing that what she says is false, not merely odd,
linguistically aberrant, or ill-mannered. In fact, because establishing that what the sceptic

says is false requires showing that she has violated, not merely the standards of

 

”I do not intend to argue at any length that these are the goals of ordinary conversation. For now,
however, perhaps it shall sufﬁce to merely indicate what I mean by ordinary conversation. Suppose,
for example, I take my car in for an oil change. The attendant tells me that my oil is a quart low. I
ask her how she knows this. She shows me the dip stick. My interest in this ordinary conversation
about my oil is not Cartesian certainty, but keeping my car on the road.

48

appropriate assertion, but also the standards of true assertion, appealing to her violation
of the former cannot suﬂice to refute her, unless there is a necessary connection between
speaking inappropriately and speaking falsely.

But does there exist such a connection between the standards governing appropriate and
those governing true assertion? Only if there is such a connection can the Principle of
Epistemic Conventionality, and, therefore, the argument from ordinary language, reﬁne,
and not merely silence, the sceptic. The question of the connection between these two
kinds of standards leads to the second interpretation of the principle of Epistemic
Conventionality. According to the second interpretation, the sceptic’s linguistic
transgressions sufﬁce to refute her, not merely prove her ill-mannered. According to that
interpretation, the kind of irrelevance the sceptic has is epistemic. But this second
interpretation rests on the presumed existence of a necessary connection between those
standards governing what it is appropriate to say and those governing what is true. I turn
now to that presumed connection.

There are, in fact, compelling reasons to believe that the standards governing
appropriate assertion are not only not strongly connected, but are in fact logically
independent. If the standards are logically independent, then, rather than giving one
compelling reason to reject scepticism, one will have compelling reason to conclude that
showing the sceptic’s expressions of doubt are inappropriate or ill-mannered — which is
all the ﬁrst interpretation of the principle can justify —- does not sufﬁce to refute her.

Moreover, and consequently, if there is no such strong connection, then only the ﬁrst

49

interpretation of the Principle — according to which the sceptic’s irrelevance is conﬁned
to the context of manners — is justiﬁed. But this interpretation, though it may justify
shunning or ignoring the sceptic for pragmatic considerations, will not do so on semantic
ones. That is, it will not do so on the basis of precisely the grounds necessary to refute,
rather than merely silence.

Let me then establish the logical distance between the standards of appropriate and true
assertion. I begin by considering the following: Suppose that having outdone ourselves at
the garden party, we ﬁnd ourselves resting up at the Opera. Suppose, moreover, that
while taking in La Boheme, I notice not only that the gentleman next to me is wearing a
taupe, but also that his dental hygiene is so poor that I can discern, without effort, the
particular herbs which have seasoned his dinner. Should I mention to this fellow that his
toupe has slipped and his teeth need brushing? It is perfectly obvious that my so doing
would be completely inappropriate. But, what follows from my mere inappropriateness?
Does it follow that the statements I make are false? Are they irrelevant to the truth about
his toupe or his teeth? Surely not. His taupe has, after all, slipped, and his teeth are laden
with basil.

One might argue that the sceptic’s queries are much like my (proposed) undiplomatic
and ill-mannered behavior toward the stranger at the opera. Indeed, Austin and his
ordinary language companions would, no doubt, appreciate the analogy. Importantly,
however, just as it does not follow from my ill-manneredness alone that what I assert is

false, so also it does not follow from the sceptic’s ill-manneredness alone that what she

50

asserts is false, no matter how many conventions (governing appropriate assertion) she
may twist or violate.

Consider another example.38 Suppose my good ﬁiend Rodger invites me to a party, and
I accept. Suppose, moreover, that just prior to my arriving at the party I have parted
company with my friend John. Suppose that I have just encountered John buying supplies
for Rodger’s party, to which he assured me he is going forthwith. Now, suppose further
that, after I have arrived at the party, the host (Rodger) asks of me “Is John going to be at
the party?” And suppose, ﬁnally, that I respond, “Yes. I know he will. I just saw him at
the store and he assured me he would be here forthwith.” In the case just described, I
think, the conditions for appropriately asserting that I know are met. I have, after all, just
seen John. John did in fact assure me he was on his way to the party. (Perhaps earlier in
the day I even helped John pick out a jacket for the party.) But, from the appropriateness
of my assertion alone, does it follow that I know John will be at the party? Does it follow
that my knowledge claim is true? Surely not. For, in spite of all my appropriateness,
consider what could happen. John, upon leaving the corner store, might be struck by a
car. Indeed, for all I know, he might be struck by a meteorite. In fact, he may be so
struck that he is unable to go to the party and ﬁnds instead he must go to the hospital
where he lapses into a coma and dies.

Suppose that John is struck by a meteorite, does lapse into a coma, and dies. Suppose,

moreover, that upon my leaving the party, the host — having been informed of John’s

 

38I borrow the structure of this example (and much detail) from Barry Stroud’s Signiﬁcance of
Philosophical Scepticism, 68-71.

51

misfortune —- approaches me, not to thank me for having attended, nor to console me on
the loss of my friend, but only to berate me for the false arrogance I expressed in saying of
John that I knew he would be there. After all, he points out, I knew no such thing. A
necessary condition of knowing John would be at the party is knowing he would not be
struck down by a meteorite. And this, he exclaims, I clearly did not know.

In such circumstances, surely one must concede that the host’s beration of me is wholly
inappropriate and inexplicably rude. Yet, does it follow from his ill-manneredness alone
that what he says is false? Did I in fact [wow that John would be at the party? Did I know
John would not be struck down by a meteorite? Surely not. After all, John was not at the
party, and he was struck down by a meteorite.39

What I wish to suggest here, of course, is that the party host is much like the sceptic.
The sceptic too behaves rather inappropriately at parties. At the garden party, the sceptic
claimed that you failed to know there was a goldﬁnch in the greenery because you did not
know that you were not dreaming, vatting, or deceived by an evil and cunning genius. In
so doing, the sceptic violated our ordinary linguistic conventions. Her suggestion,
therefore, is rather like Rodger’s suggestion that I failed to know that John would be at
the party because I did not know that he would not be struck down by a meteorite. And,
just as it did not follow from the inappropriateness of Rodger’s comments that what he
said was mistaken, so also it does not follow from the sceptic’s inappropriateness that

what she says is mistaken.

 

3”I assume here that it is also a necessary condition of knowing that what is believed be true.

52

What surely does follow, however, from these examples, is that there is no necessary
connection between ﬁrlﬁlling (or failing to fulﬁll) the conditions of appropriate assertion,
on the one hand, and fulﬁlling (or failing to ﬁrlﬁll) the conditions of true assertion, on the
other. These examples clearly establish the logical independence of these two sets of
conditions: the host’s remarks are both absurd (inappropriate) and true, and my remarks
are both appropriate and false. What follows is that one cannot infer from the
inappropriateness of the sceptic’s behavior alone that what she says is false. In particular,
one cannot infer from the inappropriateness of her objections that they are irrelevant to the
truth of our knowledge claims. Merely establishing that she violates linguistic conventions
which govern appropriate assertion cannot, therefore, establish that she violates
conventions governing true assertion, and hence though she may be silenced, she most
assuredly is not refuted. The epistemic relevance of the sceptic’s poor manners is nil.

One might object to my examples on the grounds that in them the conditions of
appropriate assertion are not in fact fulﬁlled. Failing to fulﬁll these, one might argue, they
cannot establish the logical independence of this set of conditions from that of the
conditions governing true assertion. Thus, for example, one might argue that the
conditions of appropriately asserting that John would be at the party include knowing that
John would not be struck down by a meteorite. And, the objection might continue, since
clearly I did not know that John would not be struck down by a meteorite, I ought not to
have asserted that I knew he would be at the party. At the very least, one might argue, it

was not appropriate for me to do so. Thus, one might argue, the illustration was not one

53

where the conditions of appropriate assertion were fulﬁlled, and hence not one where the
latter where fulﬁlled and the conditions of true assertion were not.

There are a number of things one can say about this objection. First, and perhaps most
damaging to its proponent, unless the objection is made during a record-breaking meteor
shower, it sets the standards of appropriate assertion incredibly high. Making knowledge
that my ﬁiend will not be struck down by a meteorite a condition of appropriately
asserting I know he will be at the party implies the inappropriateness, and irrelevance, of
nearly every knowledge claim one might ordinarily make.40 Moreover, setting the
standards of appropriate assertion this high amounts to giving the sceptic her cake and
eating it with her. (Surely, the standards of true assertion will not be lower or easier to
fulﬁll than the standards of appropriate assertion.) For precisely this last reason, not to
mention the general counter-intuitiveness of the objection, this is an objection no ordinary
language philosopher would make. In fact, it is more likely an objection a sceptic would
make, and hence serves no purpose in the defense of an ordinary language refutation of
scepticism.

One must then recognize a logical distinction between the conditions of appropriate
assertion and the conditions of truth. One must recognize that this distinction is such that
the one set of conditions can be fulﬁlled while the other is not. Thus I may, while at the

opera, make a true though inappropriate assertion, just as you may, at the party, make an

 

”Many, of course, have argued that the sceptic is at fault for doing precisely this. I shall discuss the
error of this view in chapter V. Sufﬁce it for now to point out that the sceptic’s conclusion are not
directed at what we may appropriately assert, but rather, at what we may justiﬁedly believe.

54

inappropriate but true assertion. The two sets of conditions are logically distinct.
Because they are logically distinct, one cannot infer from the fulﬁllment (or lack thereof)
of the one set, the fulﬁllment (or lack thereof) of the other.

What, then, should one make of Austin’s claim that because the sceptic’s doubts are
inappropriate, they are also irrelevant to our claims to know? I have argued that this
conclusion is based on the Principle of Epistemic Conventionality, the Principle that if
raising Cartesian doubts violates the conventions governing appropriate assertion, then
those doubts are irrelevant to the knowledge claim against which they are directed. I
argued that the conclusion that the sceptic’s doubts are irrelevant must be based on this
principle since it is only by thus linking inappropriateness to irrelevance that the argument
from ordinary language can move from facts about ordinary discourse alone to
conclusions about relevance. In particular, it is only if these two sets of conditions are so
linked that the ordinary language philosopher can infer the irrelevance of the sceptic’s
hypotheses from their mere inappropriateness

I am now in a better position to answer the question: what is the epistemic signiﬁcance
of the sceptic’s violating the conventions of appropriate assertion? Is her violating these
conditions sufﬁcient reason to conclude that her doubts are epistemically irrelevant;
irrelevant to the truth of our claims to know? Is the sceptic’s inappropriateness grounds
for concluding her hypotheses are philosophically insigniﬁcant or that her sceptical
conclusions are false? Clearly not. From their inappropriateness alone, nothing follows

about their truth.

55

Of the two possible interpretations of the Principle of Episterrric Conventionality, then,
the ﬁrst is true, but ineffective; the second false. According to the ﬁrst,one may conclude
from the inappropriateness of the sceptic’s assertions only that they are ill-mannered; not
irrelevant. But on this interpretation, the principle is philosophically inept. It offers no
guidance on the issue of truth or falsehood, and, in particular, it justiﬁes no inferences
about the truth or falsehood of the sceptic’s conclusions. It justiﬁes no inference regarding
the epistemic merits of her doubts. It is, however, to the epistemic merits of her doubts
that any proposed reﬁrtation must be directed. Whether she is ill-mannered, whether her
doubts seem absurd, or out of place, the question of her correctness, the question of
whether one can in fact know anything, is a question of truth. It is a question about the
truth of our knowledge claims, and hence a question about the objective conditions which
must be met if one is to know. It is, therefore, to no avail that the argument from ordinary
language lists the multitude of linguistic rules the sceptic violates. Only the second
interpretation of the principle is sufﬁcient to reﬁrte the sceptic. Yet this interpretation is
neither justiﬁed nor true. Manners and truths stand independent. Hence, that
interpretation that alone could justify the conclusion that the sceptic’s doubts are
epistemically irrelevant— is false; based as it is on the false assumption that ill-
manneredness alone is semantically signiﬁcant.

If, therefore, the ordinary language philosopher believes she has reﬁtted the sceptic (as
opposed to having merely silenced her), she has confused good manners for good truths.

For it is only by illegitimately conﬂating the standards governing appropriate assertion

56

with those governing true assertion that she concludes from the mere inappropriateness of
the sceptic’s hypotheses their epistemic irrelevance. The sceptic makes no claim to share
her linguistic habits with the ordinary; nor should she. Her concern is not pragmatic, but
semantic. She is interested not in what is conventional, but in what is true.

In both her inference from linguistic conventions to epistemic truths, and in her
underlying conﬂation of the standards governing appropriate and true assertion, the
ordinary language philosopher anticipates the themes — as well as the ﬁrndamental errors
—- of the anti-sceptical philosophers of my next two chapters; the contextualists and the
relevant altemativist. Like the ordinary language philosopher, both the contextualists and
the relevant altemativists base their reﬁrtations of the sceptic on the contextual
inappropriateness of the sceptic’s doubts. To the extent to which they do base their
refutations on mere facts of ordinary linguistic behaviour, they can be viewed as the
contemporary representatives of the ordinary language movement. Unfortunately, to
precisely this same extent, they too fail; both in their attempts to refute the sceptic, and in
their attempts to save — from the specter of scepticism — what they take to be
common sense. It is to these contemporary defenders, and their likenesses to the ordinary

language philosopher I now turn.

57

CHAPTER HI
Contextualism
In this chapter, I look at a relatively new position in epistemology called

‘contextualism.’l The contextualist, like the ordinary language philosopher, seeks to
defend the veracity of our common sense knowledge claims, and, thereby, to save our
“ordinary” knowledge. The contextualist seeks also to defeat the ever-present threat of
epistemological scepticism. For the contextualist thinks that scepticism will continue to
threaten our knowledge claims unless we can locate the source of its mysterious power. In
her attempt to explain this mysterious power, the contextualist goes further than the
ordinary language philosopher. The contextualist argues that the power of scepticism is
located in a kind of (contextual) truth about the sceptic’s position. The contextualist ﬁnds
it possible both to defeat the sceptic and to concede (a contextual) victory to the sceptic
because, unlike the ordinary language philosopher’s, the contextualist’s refutation of
scepticism emerges from a semantical theory of knowledge according to which there are a
number of correct analyses of what it means ‘to know’. In fact, it seems that on the
contextualist’s account of knowledge, there are as many correct deﬁnitions of knowledge

as there are knowledge-contexts. For knowledge, on the contextualist’s analysis, is a

 

1The origin of the term ‘contextualism’ is, I believe, Peter Unger’s Philosophical Relativity, though,
as we shall ﬁnd, the position properly called ‘contextualism’ predates Unger’s book. See, for
example, David Lewis, “Scorekeeping in a Language Game,” Journal of Philosophical Logic 8,
(1979): 339-359. And, in contextualist theories of justiﬁcation, David Annis, “A Contextualist Theory
of Epistemic Justiﬁcation,” American Philosophical Quarterly, 15, (1978): 105-112.

58

contextually-sensitive term. Hence, according to the contextualist, because of this
semantic diversity, the sceptic’s denial of our claims to know is compatible with a
victorious common sense.2 The contextualist argues, therefore, that the mysterious hold
the sceptic exercises on phiIOSOphy can be explained in a way which protects ordinary
knowledge.

In the following pages, I argue that the contextualist’s successes in both refuting the
sceptic and saving common sense rest on the same ﬁrndamental error as did the ordinary
language philosopher’s of chapter 11: For like the ordinary language philosopher, the
contextualist appears to succeed only to the extent that the she fails to separate the
standards of appropriate assertion from the standards of true assertion, and to the extent
that this failure goes unnoticed. Once these standards are distinguished, however, her
arguments succumb to the same difﬁculties. For both her refutation of the sceptic and her
conclusion that we do know those things we ordinarily say that we know, rest on her
conﬂation of these logically independent sets of standards. Morever, since the
contextualist grounds her standards of ordinary knowing only in the logically
independent standards of appropriate assertion, the failure of her victory for ordinary
knowledge is a fait accompli. As a consequence, the objective standards of knowledge

remain, in spite of the contextualist’s efforts, undetermined.

 

2I ﬁnd, however, that the nature and extent of this victory, though conceded by all contextualists
(see, for example, David Annis, (1978), Stewart Cohen, “Knowledge, Context, and Social Standards,”
Synthese, 73, (1987): 3-26; and Keith DeRose, “Solving The Sceptical Problem,” The Philosophical
Review, 104, (1995): 1-52; and David Lewis “Scorekeeping” ), is viewed somewhat differently by
each.

59

I argue, moreover, that even if one ignores the contextualist’s failure to distinguish the
logically independent standards for appropriate and true assertion, and therefore limit
one’s scrutiny to other aspects of her argument, the contextualist remains unvictorious.
For her victory for common-sense knowledge is accomplished only through a series of
conversational rules; rules which sanction the presupposing, ignoring, and assuming of
whatever is necessary to give our ordinary ‘knowledge’ contextual veracity. But, I argue,
if our ordinary ‘knowledge’ is veracious only to the extent that we presuppose, ignore,

and assume, the victory here is deeply unsatisfactory.

Section 1': What is Contextualism?

What is Contextualism? As a clearly recognizable theory in epistemology,3
contextualism emerged recently. Initially, the position was neither clearly deﬁned nor
understood.‘ This is due not only to the late appearance of the term ‘contextualism,’ but
also to the difﬁculty of distinguishing the contextualist’s position from that of another
relatively new position, that of the relevant altemativist.s The Contextualist position

does, however, emerge as a response to a very familiar tradition in epistemology. Indeed,

 

3Contextualisrn has long been familiar and discussed at much greater length in ethical theory.
‘See, for example, Cohen, “Knowledge,” which will be discussed at some length here.

’See, for example, Fred Drestke, “Epistemic Operators,” Journal of Philosophy, 67, (1970): 1007-
23; and Alvin Goldman, “Discrimination and Perceptual Knowledge,” Journal of Philosophy, 73,
(1976): 771-91; and Gail Stine, “Skepticism, Relevant Alternatives, and Deductive Closure,”
Philosophical Studies, 29, (1976): 249-61. Since much of chapter IV shall be devoted to the details
of the relevant altemativist position, and to what extent various relevant altemativists are
contextualist, I will not discuss that here.

60

as David Annis notes“, it emerges as an alternative response to the regress argument;7 a
response that claims to avoid the diﬁiculties of both coherentism and foundationalism.

The clearest and earliest attempt to deﬁne contextualism is Unger’s Philosophical
Relativity.8 Unger deﬁnes contextualism by contrasting it with a position he calls
‘invariantism’. Invariantism, in brief, is the familiar view that the necessary conditions of
knowing remain constant, rigid, inﬂexible with regard to idiosyncracies of both subject and
object. By contrast, the contextualist position is characteristically dynamic, pliable, and
ﬂexible regarding the necessary conditions of knowing, especially with respect to the
conversational context in which a knowledge claim ”is made. ‘Knowledge’, the
contextualist argues, is a contextually sensitive term: it changes both in meaning and in
truth conditions as the environment -- both internal and external to the subject -- alters.
Thus, when S is said to know that p, whether she knows that p or not is a function of the
context in which she claims to know.

Unger correctly identiﬁes a contextually sensitive semantics as the most fundamental
feature of any contextualist position. But there are a number of other features which are
particularly worth noting as well. The contextualists, for example, are united by their
common goal to reﬁrte the sceptic. Indeed, this, I think, is the motivation behind their

contextually-sensitive semantics. Moreover, like the ordinary language philosophers, the

 

6Annis, “Contextualist,”105.

7Though I will not attempt to adjudicate who has the better solution to the regress argument, see:
Robert Audi, “Contemporary Foundationalism,” The Theory of Knowledge, ed. Lewis Pojman,
(Belmont, CA: Wadsworth Publishing Company, 1993), 206-214.

8Unger, Philosophical Relativity (Minneapolis: University of Minnesota Press, 1984).

61

contextualists wish to rescue the knowledge assertions of common sense from the
perceived threat of global scepticism. They believe, moreover, that with their
contextualist semantics they can explain both the (contextually circumscribed) victory of
the sceptic, and the enigmatic sway this position has held. Finally, they believe this
explanation -- in terms of a contextual semantics -- proves that the sceptic’s success is
limited in such a way that it cannot implicate our ordinary claims to know.

In the following page I look at a variety of contextualist positions. The ﬁrst contextualist
position I discuss is that of David Lewis.9 Lewis is suitably ﬁrst in my discussion both
because his, to my knowledge, is the earliest contextualist thesis, and because it is a near

paradigm of contextualist reasoning.

Section ii: David Lewis

In “Scorekeeping in a Language Game”, Lewis argues that in any conversation a number
of things are presupposed. Thus, for example, if in the course of an ordinary
conversation, I assert that “The table-top is ﬂat”, it is presupposed (lest I be made
incorrect) that the standard of ﬂatness according to which I assert that the table top is
ﬂat, is not the absolute standard according to which nothing could ever be ﬂatter than the
table top I now describe. 1° Thus, owing to the necessary presuppositions, my claim is

true. Moreover, Lewis observes, these conversational presuppositions are both dynamic

 

9Lewis, “Scorekeeping.”

10This example is used throughout contextualist literature, though I believe its origin is Peter Unger’s
Ignorance.

62

and rule-governed: that is, they change in accordance with rules. Thus, were you to
assert that “The table-top is not actually flat”, the presupposed standard of ﬂatness would
change, moving more in the direction of the absolute standard, according to which that
which is ﬂat is not at all bumpy (under microscopic examination or otherwise), and
according to which your assertion is true.

The course of these changing presuppositions is charted by the changing contextual
needs of the conversation. And, again, this change is rule-govemed. Of the rules
governing the changing conversational presuppositions, the most important is what Lewis
calls The Rule of Accommodation. The Rule of Accommodation is as follows:

Ifat time t something is said that requires presupposition p and if p is not

presupposed just before t, then ceteris paribus and within certain limits,

that presupposition comes into existence at t.11
What does this mean? Lewis argues it means that when, for example, I claim to know that
p (that the table is ﬂat, for example), then, if my claim to know would be false in the
absence of certain presuppositions (for example, the presupposition that the standard of

”’2 at

ﬂatness is relatively low), those truth-bringing presuppositions “spring into existence
t, thereby making my claim to know true.
Because, however, the Rule of Accommodation allows the presuppositions within a

conversation to change as needed, it follows also from the Rule that if at t1 you assert that

I fail to know p, then whatever presuppositions necessary to make your statement true

 

l‘Lewis, “Scorkeeping,” 2.

I’Lewis, “Scorekeeping,” l. The full quotation is, “say something that requires a missing
presupposition and straightaway that presupposition springs into existence . . . . “

63

(for example, the presupposition that ﬂat is an absolute term) spring also (at t1) into
existence. And so it goes. When, and to what extent, is a sentence true? That, Lewis
claims, “depends on the context”:
What is true enough on one occasion is not true enough on another. The
standards of precision in force are diﬂ’erent from one conversation to
another and may change in the course of a single conversation. ‘3
Thus Lewis maintains, “it is not as easy as one might think to say something that is

”1‘ or false: As Lewis states, the Rule of Accommodation is simply an

unacceptable,
instance of the following more general rule:

Ifat time t something is said that requires component Sn . . . if what is said
is to be true . . . then at t the score component Sn [comes into existence].15

One might think that having the relevant presuppositions in place which make your claim
(that I fail to know that p) true, my claim (to know that p, at t) would thereby be rendered
false. Not so, according to the contextualist: for the presuppositions operative at t and t1
are different: the context has changed. In fact, the contextually sensitive presuppositions
allow both my claim to know at t and your claim at t1 that I fail to know, to be true. For
“ceteris paribus and within certain limits” whatever presuppositions are necessary “spring
into existence”. And presuppositions, on a contextualist’s account, do not discriminate.
Indeed at the heart of the contextually sensitive semantics is the claim that seemingly

contradictory knowledge-claims may both be true: both my claim to know and your claim

 

l3Lewis, “Scorkeeping,” 114.
l‘Lewis, “Scorekeeping,” l.
I’Lewis, “Scorekeeping,” 9.

64

that I fail to know, for example.

The above considerations have, of course, important implications for Lewis’ attempt to
reach the contextualist’s goals: the contextually sensitive presuppositions allow that both
the claims to know in ordinary conversations and the sceptical claim (in philosophical
conversations) that we fail to know are true. Moreover, the sceptic’s persuasiveness (her
power) can be explained by reference to this contextually-sensitive semantics. And, of
course, the ordinary man’s claims to know can be explained by reference to the same.

Furthermore, the sceptical victory can be contextually confined: for though what may be
truly said in one context (for example my claim to know that p in an ordinary context) may
be only falsely said in another context (my claim to know that P in a philosophical
context), since the contexts and correlative presuppositions are different, my failure to
know in the sceptic’s context of high standards, does not undermine my success in
knowing in the average man’s more ordinary context, where the standards of knowing are
lower. Instances of this rule, which governs the change of conversational presuppositions,

according to Lewis, are prevalent in ordinary conversation. ‘6

Section iii: Stewart Cohen
Stewart Cohen is another philosopher of whom I think contextualism is the right

classiﬁcation.‘7 The construction of his contextualist semantics differs from Lewis’,

 

l“Lewis, “Scorekeeping,” 347.

"Although Cohen seems to align himself with the Relevant Altemativists (see, for example p. 583),
my sense is that is so because he is familiar with the RA position (as he notes, p. 580) but lacks
familiarity with the term ‘contextualism’. Indeed, to my knowledge, contextualism in epistemology
had just begun to emerge as a distinct category shortly prior to Cohen’s paper (see Unger’s 1984, and

65

however, in important ways. Like Lewis, Cohen wants to refute the sceptic. And, as is
characteristic of contextualists, he wants to save the claims to know of the ordinary
person. But, unlike Lewis, Cohen’s 1986 paper, rather than looking to any speciﬁc
conversational rules to ﬁnd the well-springs of contextualisrrr, ﬁnds them in the necessary
conditions of the concept of knowledge itsel .18 Cohen argues that it is a necessary
condition of our concept of knowledge that the putative knower possess “good reasons.”19
Having good reasons, Cohen claims, entails the ability to defeat those defeaters to one’s
good reasons which are contextually important. But there are several kinds of “good
reasons,” and Cohen argues, not all of them are required for knowledge.

Cohen distinguishes three kinds of good reasons: there are
a. Subjectively good reasons
b. Intersubjectively good reasons, and
c. Ideally good reasons.”

Cohen deﬁnes each kind of good reason in terms of the absence of different kinds of

defeaters: One possesses subjectively good reasons for her belief that p, just in case she

 

my earlier comment on this), while the relevant altemativist position was fairly well known much
earlier. In fact Cohen discusses the RA positions of Dretske, Harman, and Goldman at p. 583 in
“Knowledge”. Supposing for the sake of argument, however, that Cohen was cognizant of
‘contextualism’ and hence consciously chose, for whatever reason, not to call his position by that
name, there is plenty of textual evidence in his “Knowledge” to support my claim that he falls
squarely within the contextualist school. Moreover, Cohen clearly identiﬁes his position as a
contextualist one in later papers: see, for example, his “Contextualisrn, Skepticism, and the Structure
of Reasons,” Philosophical Perspectives 13, (1999).

l”This is somewhat of an over simpliﬁcation. The well-springs of contextualism for Cohen are
located in both the concept, and the standards of reasoning of the relevant social group. But it is the
concept which speciﬁes what is necessary for knowing, and Cohen goes outside the concept only in
deﬁning what those necessary conditions entail.

l9Cohen, “Knowledge,” 574.

”Cohen, “Knowledge,” 575.

66

possesses no subjectively evident defeaters to those reasons; one possesses
intersubjectively good reasons just in case there are no intersubjectively evident defeaters
to her reasons; and, ﬁnally, one possesses ideally good reasons just in case there are no
defeaters (evident or otherwise) to her “good reasons.” 2‘

Cohen argues that knowledge “entails” having good reasons.22 But which kind of good
reasons does knowledge entail? Cohen argues that though subjectively good reasons are
entailed by knowledge, they are not a sufﬁcient condition of knowing.” His reasons here
are as expected: Given that subjectively good reasons are present whenever one possesses
no defeaters to those reasons which she ﬁnds evident, if subjectively good reasons were
all that were required for knowledge, then even where there were numerous and widely
recognized defeaters to what the subject claimed to know, so long as she did not
recognize them as such, she could be said to know. Surely, on any adequate account of
knowledge, this will not do.

What about ideally good reasons? Requiring of the putative knower that she possess
ideally good reasons surely would solve the problem of the standards for knowledge being
too low: if the possession of ideally good reasons were requisite for knowledge, no one
could rightly claim to know whether she possessed even one defeater to her good reasons

which she could not rule out. This, of course, is a condition of knowing most hospitable

to the sceptic. But, Cohen rejects this very strong requirement that ideally good reasons

 

21Cohen, “Knowledge,” 576.
21’Coherr, “Knowledge,” 575.

23Cohen, “Knowledge,” 577.

67

are necessary for knowing. He argues instead that knowledge entails only the possession
of intersubjectively good reasons.” But what grounds has Cohen for dismissing the
possibility that knowledge entails ideally good reasons? And what reason has he for
thinking that knowledge entails merely intersubjectively good reasons?”

Cohen argues that it is the intersubjectively good, and not the ideally good, reasons
which are required for knowledge because whereas the former are dictated by the
“intersubjective standards of reasoning” the latter are not.” He supports this claim by
appealing to our ordinary sense of when we say of someone that she either knows, or fails
to know. He argues that we withhold knowledge only up to a certain level of
intersubjective opacity; that is, when a subject possesses a defeater which is
intersubjectively opaque — not evident as a defeater to the group — then we do not
withhold knowledge ﬁ'om the subject, but rather, we readily assent to her claim to know.
Contrariwise, when a subject possesses an intersubjectively evident defeater, and -- in spite
of the presence of this intersubjectively evident defeater, which she fails to recognize as
such, claims to know that p, we readily acknowledge her failure to know. 27

Having appealed to the intersubjective standards of reasoning in deﬁning what is
required for knowledge, Cohen turns to the question: How are these standards set? It is

here that Cohen’s contextualism is most apparent. Cohen argues that the standards of

 

2‘Cohen, “Knowledge,” 578.

2"There are a number of obvious objections to this requirement. I will discuss these in the latter part
of this chapter.

26Cohen, “knowledge,” 57 8.
27Cohen, “Knowledge,” 578.

68

reasoning are affected by two factors: a. The “normal reasoning powers of the

”28 and b. The context of the attribution. ” He concludes,

attributor’s social group,
therefore, like David Lewis, that whether a given knowledge attribution is true depends
upon both the context in which it is uttered and the social group relevant in the context of
the attribution.

Once again, therefore, I ﬁnd that the semantics of knowing is a contextually sensitive
matter. Again, as in Lewis’ contextualism, both I, in my claim to know that the table top
is ﬂat, and you, in your claim that I fail to know the table top is ﬂat, are correct. I am
correct because the acuity of my group is lower than yours, making any defeaters to my
“good reasons” intersubjectively opaque; and you are correct because your group, being a
little brighter than most, ﬁnds the defeaters to my good reasons quite evident and
indefeasible. As Cohen asserts:

The truth-value of a knowledge attribution will depend on the status of the

defeater the subject possesses, relative to the standards that apply in the

context of the attribution.”
We come, therefore, to a thoroughly contextualist semantics of ‘knowledge’: a semantics
according to which the necessary conditions vary, not only from group to group, but from
context to context as well. Once again, therefore, the contextualist —- in this case Cohen

—— meets his contextualist goals by contextualizing the necessary conditions of knowledge.

And again, because the conditions of knowing are contextually sensitive (sensitive to the

 

2“Cohen, “Knowledge,” 579
29Cohen, “Knowledge,” 579.

”Cohen, “Knowledge,” 579.

69

relative acuity of the subject’s social group), both the sceptic and the ordinary person are
correct: The sceptic is correct that the ordinary person fails to know, and the ordinary
person is correct her claims to know. For what is intersubjectively opaque or evident
differs from group to group, and so, therefore, do the conditions of knowing.

Section iv: Keith DeRose

Let us turn now to a more recent contextualist,”l Keith DeRose. Of the DeRose papers I
will discuss or refer to, each is more recent than those of Lewis and Cohen. Moreover,
perhaps owing to the advantage of time, DeRose recognizes his view as a
‘contextualist’one, and titles his 1992 paper accordingly: “Contextualism and Knowledge
Attributions“32 It is his 1995 paper “Solving the Sceptical Problem,” however, that is
most relevant for us here.” In this paper, DeRose, like Lewis and Cohen, attempts to
defeat the sceptic. Like Lewis and Cohen, he defends a contextualist account of knowing.
And again, like the latter, in an effort to both defeat scepticism and explain its enigmatic
power, he defends the knowledge-claims of ordinary folk. Further still, DeRose, like the
other contextualists, argues that his solution establishes the “compatibility” of common
sense knowledge claims and global scepticism.

Unlike Lewis and Cohen, however, DeRose has the advantage of a contextualist

 

31I leave out some other very ﬁne examples of contextualists, due only to a limitation of space and a
desire to avoid redundancy. See, for example, David Annis, “A Contextualist theory of Justiﬁcation,”
American Philosophical Quarterly 15, (1978).

32Keith DeRose, “Contextualism and Knowledge Attributions,” Philosophy and Phenomenological
Research, L11 (1992): 916-929. I shall refer to the 1992 paper infrequently. This decision is justiﬁed
both by considerations of length and of the greater complexity of his 1992 paper.

33I will discuss his “Contextualism” at some length in chapter IV.

70

tradition on which to fall back. In fact, he argues that though the main lines of the
“traditional” contextualist solution are correct, they lack the power to explain the
sceptic’s persuasiveness.” Though DeRose endorses their basic contextualist strategy of
arguing that the sceptic succeeds only at different contextual standards, he argues that the
account of Lewis, for example, offers no real insight into why it is we are so tempted by
the sceptical claim that we fail to know, as for example, when the sceptic asserts we fail to
know we are not mere Brains in vats.35 It is in this area, of providing a real explanation of
the persuasiveness of the sceptical argument (the argument from ignorance, in this case),
that DeRose thinks his account is superior.

DeRose, however, does not make his efforts alone. In his superior attempt both to
explain and defeat the sceptic he draws heavily on Nozick’s subjunctive conditional
analysis of knowledge. 3" DeRose believes that an adequate explanation of the power of
the sceptic’s argument must locate the “mechanism” or “conversational rule” by which the
sceptic “manipulates the standards” of knowledge: This, he thinks, Lewis failed to do (in
his 1979 paper). Indeed, DeRose believes it is only because the sceptic so manipulates

the standards of knowing that she succeeds at all.37 But what is the mechanism by which

 

”DeRose, “Solving,” 7 rr

35Speciﬁcally, DeRose argues that Lewis’ rule of Accommodation can afford no explanation of why
it is that the argument of the complex sceptic is more convincing than the mere assertion of what he
calls “the simple sceptic” that we fail to know.

3"Robert Nozick, Philosophical Explanation, (Cambridge, Harvard University Press, 1981): 167-228

37DeRose (“Solving”) claims that the earlier contextualists also maintain that the sceptic succeeds
only by “raising the standards” of knowledge. This is somewhat innaccurate and needs to be
clariﬁed. Neither Lewis or Cohen argue that the sceptic succeeds by manipulating the standards
(unfairly or otherwise). Rather, as in the case of Lewis, it is by the rule of accommodation -- a rule
which operates equally in the spheres of common and philosophical discourse) that the sceptic’s

71

the sceptic does this? It is in locating this supposed mechanism by which the sceptic
manipulates the standards of knowledge that DeRose appeals to Nozick’s subjunctive
conditional analysis of ‘knowledge’. DeRose argues that Nozick’s claim that a belief
must be “sensitive” to count as knowledge provides the insight needed for a more
advanced contextualism.

A belief is sensitive, on Nozick’s account, iff we would not hold that belief if it were
false. Equivalently, a belief is insensitive iff we would continue to believe it even if it were
false.38 DeRose argues that we in fact have a tendency to believe that a belief must be
sensitive if it is to constitute knowledge. And, it is this tendency, he claims, that is
responsible for the extent to which we are psychologically overpowered by the sceptical
argument. However, though we are persuaded by our tendency to believe sensitivity is
required for knowledge, sensitivity is not a part of our concept of knowledge, and,
therefore, is not ordinarily required for knowing.

DeRose parts company here with Nozick when it comes to building the sensitivity
requirement into the concept of knowledge itself: Nozick argues that sensitivity is a
necessary condition of the concept of knowing, and that this explains why the sceptic’s
claim that we don ’t know that -H (where H is a sceptical hypothesis to the effect that we

may be mere brains in vats, for example) is so powerful. DeRose rejects this explanation

 

claims are made true. And, in the case of Cohen, it is in virtue of the fact that what is
intersubjectively opaque to the average person is not opaque, but evident, to the sceptic, that the
sceptic correctly asserts that we fail to know. Perhaps these are not so different from DeRose’s
explanation, but the detail is worth noting, especially in light of the somewhat hostile tone DeRose
takes toward the sceptic; which is absent in both Lewis and Cohen.

3”DeRose, “Solving,” 18.

72

in favour of another. He argues, in contradistinction to Nozick, that sensitivity is not a
part of our concept of knowledge, and therefore cannot be the explanation for the
persuasiveness of the sceptic’s argument from ignorance. The sensitivity requirement ,he
argues, is only brought into play by the sceptic’s manipulation of a conversational rule, a
rule that when manipulated raises the standards of knowing to a level at which sensitivity
is required. And it is only at this level, in the unusual context of sceptical manipulation,
that the sceptic correctly claims that we know nothing.
This rule, manipulated by the sceptic, to raise the standards of knowing is as follows:
When it is asserted that some subject knows (or does not know) that p, the
standards for knowledge (how good an epistemic position one must be in
to count as knowing) tend to be raised, if need be, to such a level as to
require S’s belief in that particular p to be sensitive in order to count as
knowledge. 3’
DeRose calls this “The Rule of Sensitivity.”40 And, he claims, it is only at the artiﬁcially
high standards brought into play by the rule of sensitivity and the philosophical context of
sceptical discourse, that the sceptic succeeds in arguing that we fail to know. For,
according to DeRose, because sensitivity is not a part of our ordinary concept of knowing,
the sceptic’s success is limited to this kind of contextual bubble wherein the rule of
sensitivity is brought into play. Thus DeRose concludes:

At these high standards, the sceptic truthfully asserts that we don’t know

that -H [that we are not brains in vats, for example],"1
though

 

”DeRose, “Solving,” 36.
”DeRose, “Solving,” 36.
"DeRose, “Solving,” 37.

73

to the extent that he does succeed, he does so only by raising the standards

for knowledge and so the success of his argument has no tendency to show

that our ordinary claims to know are in any way defective. ‘2
This, he thinks is the correct explanation of the sceptic’s (mitigated) success. Thus, he
concludes, like Lewis and Cohen, our ordinary knowledge claims are true and compatible
with scepticism. The sceptic succeeds only in the unusual context -- created by the sceptic
-- where the rule of sensitivity drives up the standards for knowledge; while common

sense succeeds in the more ordinary contexts of everyday conversation, a level where the

standards for knowing are admittedly low.

Section v: Contextualism and Ordinary Language

I have stated both the way in which DeRose’s contextualism attempts to defeat the
sceptic and save our ordinary claims to know by describing a semantics that is contextually
sensitive, and the extent to which the contextually sensitive semantics of Lewis and Cohen
do the same. It is time now, however, to ask for some justiﬁcation of this proposed
semantic theory. One must ask for the justiﬁcation not only of the contextualist’s
semantics, but also, to the extent that it differs, her justiﬁcation for the standards she
argues are the “normal” necessary conditions of knowing, and according to which the
sceptic cannot defeat us. It is here, with respect to the ground of her position, that I ﬁnd
that the Contextualists fall victim to the same kind of problem as that suffered by the
ordinary language philosopher of chapter II.

I argued in chapter H that Austin, for example, in his attempt to argue that the sceptical

 

“DeRose, “Solving,” 6.

74

hypotheses are not relevant to our ordinary knowledge claims, relied on the Principle of
Epistemic Conventionality. According to that principle, whenever a sceptical objection is
made (to the eﬁ’ect that we cannot meet the objective conditions of knowing), if it violates
the conditions of appropriate assertion, that challenge can be ignored. The conditions of
appropriate assertion, I showed, are deﬁned by facts about usage: the things we ordinarily
say about knowledge, the kinds of objections we ordinarily make to knowledge claims,
and the kinds of responses we ordinarily both give and accept (or disregard) to queries
about our claims to know. This set of linguistic facts about what we ordinarily do was
thought to make irrelevant any objection, or claim, that was non-ordinary. Yet, by
showing that the conditions of appropriate assertion (claim, doubt, etc.) are logically
independent of the conditions of true assertion, I showed also that it is only to the extent
that she conﬂates these independent standards that she succeeds in refuting the sceptic and
saves ordinary knowledge.

The contextualists -- in the course of arguing for their most important conclusions --
make nearly the same illegitimate conﬂation as the ordinary language philosopher. And
though the details to some extent vary, the principles involved are much the same. The
extent that the contextualist relies on the same kind of principles in her effort to save the
linguistic norm, therefore, the failure of her purported successes is a fait accompli .

That such a formidable error is made by the contextualists is quite evident in David
Lewis’ discussion of the Rule of Accommodation. Lewis argued that, in accordance
with the Rule of Accommodation, the presuppositions necessary to make a claim true

“spring into existence” as needed. Thus, he claimed, though the sceptic’s denials of our

75

claims to know are true, our individual ordinary knowledge claims are not thereby
rendered false: in fact they remain true in spite of their correctly asserted falsehood!

Lewis’ explanation for this exceedingly strange phenomenon is the Rule of
Accommodation. He argued that this rule (which is merely an instance of a general
schema operative in ordinary conversations) makes those presupposition needed to make
the attributor’s (or the denier’s) knowledge claim true a part of the conversational
presuppositions, ceteris paribus of course, and “within certain limits.” And, moreover,
the rule-govemance of these changes in the contextually sensitive presuppositions is what
makes possible both an explanation of the sceptic’s success and the vindication of our
ordinary knowledge-speak from her sceptical clutches. For, as I noted, our knowledge
claims made at t remain true, even if the sceptic correctly claims at t1 that our knowledge
claims are false: This is the magic of a contextually-sensitive semantics.

In what sense, however, are the changing presuppositions of our conversations “rule-
govemed?” In what sense is the Rule of Accommodation a rule. There are at least two
possible answers here: 1. The rule is a rule in the sense that it prescribes what can and
can not be assumed in a conversation, or 2. The rule is a rule because it describes the
procedures followed in ordinary conversation; that is, it describes the presuppositions we
happen to make. It is, however, only if it is a rule in the first sense that Lewis’
description of how the presuppositions change in a conversation can save knowledge. For
if the “rule” is merely descriptive, it offers no guidance, and certainly no standards,
according to which a claim to know is either true or false. It offers no more guidance than

would rules which describe how we happen to reason. (The gambler’s fallacy, afﬁrming

76

the consequent, and ad hominens come to mind.) Rather, it tells us merely what we
happen to do, how presuppositions happen to get made in a conversation, not how they
ought to get made. ‘3

But is Lewis’ rule a rule in the ﬁrst sense? Clearly it is not. This is evident if we
consider the origin of the rule: the origin is not the analysis of our concept of knowledge.
Indeed, Lewis offers none."4 Lewis’ rule does not even purport to prescribe any
conditions which must be followed for the changing of presuppositions to be legitimate."5
He tells us merely that this rule - the Rule of Accomodation - can be seen to be operative
in many instances in conversations; that it happens to operate in a way which favours
neither scepticism nor common sense.

The only “support” Lewis offers for his proposed rule is that it can be seen to be
operating in conversation. The rule, it seems, has its origin in our ordinary everyday
practices, as this is the only origin Lewis describes. This would go a long way to
explaining its non-prescriptive nature. But if the origin of the rule of accommodation just
is our practices, it seems that the rule itself is not very rule-like at all: if it is “discovered”
simply by paying attention to what we ordinarily do, to what happens when assertions and
counter-assertions are made, and is merely descriptive thereof, it provides no basis for the

objective prohibition or prescription of any conversational practice. For, as I noted in

 

”Indeed, it is a pattern throughout contextualist literature to ignore this important distinction; the
distinction between what is the case and what ought to be the case.

“Lewis does, to some extent, offer such an analysis in his later paper “Elusive Knowledge.” I shall
address that shortly.

4‘Again, in his “Elusive Knowledge,” he does attempt a list of both prohibitive and prescriptive
rules. I discuss these shortly.

77

chapter II, examples abound of knowledge-claims and practices that violate the standards
of true assertion, but not the standards of appropriate assertion, and also of knowledge-
claims that do not violate the standards of appropriate assertion, but which are clearly
false. And since we are here interested in the rules governing knowledge, the rule of
Accommodation, since it is grounded merely in what we happen to do, brings us no closer
to an objective criterion of knowing than at the end of chapter 11. The extent to which this
rule is merely descriptive, it offers no guidance, but merely narration. ‘6

Having argued that Lewis’ Rule of Accommodation fails to be anything but descriptive
in character, in fairness, I should note that Lewis’ more recent paper, his 1996 “Elusive
Knowledge,” at least appears not to make this mistake. In this later paper Lewis again
argues for a contextualist position, but his approach seems much more reﬁned, both with
respect to his earlier paper and with respect to other contextualists. Lewis, here, attempts
not only to provide a number of types of rules, he also goes so far as to argue that
knowledge is infallible. He argues, however, that in spite of the infallibility of
knowledge, we can save common sense knowledge claims.

How does Lewis propose to save ordinary knowledge? He begins by offering a
deﬁnition of knowledge and then providing a number of rules that ﬁrlly explain its
meaning. His deﬁnition of knowledge is as follows:

S knows that P iﬁ’ P holds in every possibility left uneliminated by S’s evidence.

Or, equivalently,

 

“Again, I think my analogy with how we happen to reason is instructive. For surely a mere
description of how we happen commonly to reason does not reveal how we ought to reason.

78

S knows that P iff S’s evidence eliminates every possibility in which not-P."7

This deﬁnition of knowledge looks surprisingly congenial to the sceptic. After all, it is
precisely because the sceptic holds that there are uneliminated possibilities in which not-P,
that she holds S fails to know that P. Lewis, of course, is aware of this difficulty and
proceeds to resolve it by a special deﬁnition of what it means to require of every
possibility that it be eliminated.

He argues that the term ‘every’ is “normally” restricted to a certain domain. Thus, for
example, when at a party one asserts “all the glasses are empty”, it is understood that the
assertion is conﬁned to a limited domain, namely, the glasses at the party. The glasses
outside of the party (in the rest of the world) are understood to be irrelevant to the truth
of the claim, as they are outside the limited domain."8

Lewis argues, by analogy, that this is the case with knowledge. Hence, he asserts,
“when we say that every possibility must be eliminated, we can ignore some of the
possibilities, for they are irrelevant to the truth of what is said.”"9 For, “what does it mean
to say every possibility in which not-p is eliminated? An idiom of quantiﬁcation like
‘every’ is normally restricted to some limited domain?”o Feeling the contextualist need to
keep the sceptic at a safe distance, Lewis proceeds straightaway to modiﬁr his initial

deﬁnition of knowledge dramatically as follows:

 

4’Lewis, “Elusive Knowledge,” 551. (My emphases.)
4"Lewis, “Elusive Knowledge,” 553.

"Lewis, “Elusive Knowledge,” 553. (My emphasis.) One would like to exclaim here “So much for the
infallibility of knowledge!”

”Lewis, “Elusive Knowledge,” 553. My emphasis.

79

S knows that P iff S’s evidence eliminates every possibility in which not-P
— Psst! — except for those possibilities that we are properly ignoring."

The remainder of Lewis’ paper is devoted largely to clarifying this sotto voce proviso.
That clariﬁcation takes the form of an enumeration of rules, rules which specify what may
and may not be “properly ignored”.

Among the more outstanding of these rules are the rules of Actuality, Resemblance,
Reliability, Conversation, and Attention. The Rule of Actuality simply stipulates that we
may not ignore the possibility that actuality exists. In other words, “actuality is always a
relevant alternative”.’2 The Rule of Resemblance is a prohibitive rule: it states that, if
“one possibility saliently resembles another . . . then, if one of them may not be properly
ignored, neither may the other.” ’3

Unlike the Rule of Accommodation of his 1979 paper, these two rules do offer guidance
concerning what may and may not be presupposed in the course of a knowledge
conversation. Thus, to some extent, they appear more rule-like. We now know, for
example, that we may not simply presuppose that actuality is false, nor of anything
saliently resembling it, that it is so. This helps to place restrictions on both what the
ordinary person may be said to know, and on what the sceptic may claim is not known.

To this extent, the rules of “Elusive Knowledge” seem a great improvement on those of

“Scorekeeping.”

 

"Lewis, “Elusive Knowledge,” 554. (One can nearly hear the sceptic gasp! “Infallibility indeedl”)

”Lewis, “Elusive Knowledge,” 554. (What Lewis means by ‘actuality’ is merely what we ordinarily
take to be the case about the world.)

”Lewis, “Elusive Knowledge,” 556

80

Ironically, however, these rules — taken together — seem to place the sceptic at
something of an advantage. For clearly the sceptic’s hypothesis that I may be a mere BIV,
or that I may be dreaming, are compelling precisely because they saliently resemble what
Lewis calls “actuality.” Thus, it seems, according to the rule of Resemblance, the ordinary
person cannot “properly ignore” the sceptical hypotheses, and furthermore, given Lewis’
deﬁnition of knowledge, she cannot claim to know. (Though, of course, the Rule of
Actuality ensures also that the sceptic may not properly ignore the possibility that actuality
isnue)

Lewis, of course, recognizes the problem his Rule of Resemblance seems to pose, 5" and
in the interest of avoiding scepticism and saving ordinary knowledge, he makes an
admittedly ad hoc exception to the rule. In conceding that the rule could apply to the
possibility in which the subject is “radically deceived by a demon” and, therefore, favour
the sceptic, he remarks:

That will be so even if W [the uneliminated possibility] is in other respects
very dissimilar to actuality. . . . Plainly we dare not apply the Rules of
Actuality and Resemblance to conclude that any such W is a relevant
alternative — that would be capitulation to scepticism. The Rule of
Resemblance was never meant to apply to this resemblance! We seem to
have an ad hoc exception to the Rule . . . . What would be better, though,

would be to ﬁnd a way to reforrnulate the Rule so as to get the needed
exception without ad hocery. I do not know how to do this.”5

 

“Actually, though he addresses one problem this rule may pose, it is not the one I mention. He
notes that the rule of resemblance may make relevant the sceptical possibilities simply because they
resemble all possibilities which are uneliminated by the evidence. (See: 557.) In fact he nowhere
addresses the problem I mention - the salient resemblance of the sceptical possibilities to actuality,
though he solves both it and the one he does discuss with his ad hoc exception to the rule of
resemblance.

”Lewis, “Elusive Knowledge,” 556-7.

81

The eﬂ’ort of Lewis to save knowledge with these two rules seems, then, to succeed only
insofar as we accept his ad hocery.

What about his other rules? Do they offer some improvement over the earlier Rule of
Accommodation? These too seem, at ﬁrst glance, to hold out some hope for providing
truly prescriptive (“r ”) rules according to which our claims to know can be not only
made and judged, but vindicated. The Rules of Reliability and Conversation —— like those
of Actuality and Resemblance —— seek to save our ordinary ‘knowledge’ ﬁom the
sceptical hypotheses. And, unlike the Rule of Accommodation, (though like the Rules of
Resemblance and Attention) these rules oﬂer guidance regarding what is — and what is
not -— acceptable to presuppose when making claims to know.

The Rule of Reliability, for example, allows us to “assume that perception, memory, and
testimony are . . . fairly reliable.” Indeed, “we are entitled,” Lewis says, “to take them for
granted.” “We may properly presuppose that they work without a glitch.” Thus, he
says, we may assume, albeit defeasibly, that the possibilities “in which they fail may be
properly ignored.”56 The Rule of Conversation, moreover, speciﬁes that we are “entitled

”

to ignore —— defeasibly — what those around us ignore. ’7 Thus, for example, if the
members of a group I happen to be conversing with ignore all but the most obvious
defeaters to our knowledge claims, I too am entitled — without argument or ﬁrrther

justiﬁcation — to ignore them.

These “rules”, however, are not argued for by Lewis, but merely assumed. And though

 

”Lewis, “Elusive Knowledge,” 558.
”Lewis, “Elusive Knowledge,” 559.

82

assuming perception is reliable, and ignoring all those sceptical possibilities those around
us ignore, should surely help us to defeat the sceptic, one is struck by the fact that one
wins (defeats the sceptic) only by begging the questions at issue. After all, is it not at the
very heart of the sceptic’s argument that we may be misled by our perceptions, that we
may be wrongly ignoring certain uneliminated possibilities which undermine our claims to
know? Moreover, are mere presumptions rule-like at all? It seems the hope held out for
grounding our ordinary claims to know in real rules is dashed, even in this later paper of
Lewis. For though these rules are indeed more rule-like, their basis — assuming and
question-begging — is ﬁrndamentally ﬂawed. And so, though not subject to the same
criticisms as the Rule of Accommodation, they are particularly inept at moving
contextualism into a more positive light.

There remains, however, one ﬁnal “rule” to examine: The Rule of Attention. This rule,
like the others of Lewis discussed above, is meant to ﬁll out the sotto proviso of Lewis’
deﬁnition of knowledge and, thereby, to clarify the conditions of knowing. The Rule of
Attention states that what is and is not ignored is a feature of a particular conversational
context.’8 It states, moreover, that

No matter how far fetched a possibility may be . . . if we are not now
ignoring it, then for us now it is a relevant alternative.”59 “Do some
epistemology. Let your fantasies rip. Find uneliminated possibilities of

error everywhere. Now that you are attending to them . . . you are no
longer ignoring them, properly or otherwise.”

 

”Lewis, “Elusive Knowledge,” 559.
”Lewis, “Elusive Knowledge,” 559.

6‘)Lewis, “Elusive Knowledge,” 559.

83

What is the nature of this rule? Is it, for example, dictated by the necessary conditions
of knowledge? Is it merely descriptive of what happens when a lonely sceptic seeks our
conversation? It is not clear. Surely the rule is not of much help in saving our knowledge
claims. For the more uneliminated possibilities we must attend to, the less likely it is we
will rightly claim to know. Thus, at least if a verbose sceptic is around, our knowledge
claims fall asunder. Though perhaps, in Lewis’ view the other rules (of Actuality,
Resemblance, Reliability, and Conversation) are enough to save our knowledge from the
sceptical possibilities. (After all, what more could one who seeks to save knowledge hope
for than to be entitled to ignore whatever unelirrrinated possibilities of error those around
him ignore?) In fact, rather than functioning to safeguard our ordinary knowledge claims,
the Rule of Attention’s most important function appears to be to explain how it is that the
sceptic has such power, or in Lewis’ words, how it is that knowledge can be so “elusive.”

It is this rule, Lewis argues, that enables epistemology to “destroy itself?” making
immediately relevant any considered hypothesis, no matter how bizarre, and hence making
improper the ignoring of the sceptical hypotheses, at least when they are before our
attention. It is because of this rule — the Rule of Attention — that the sceptic’s raising of
her sceptical hypotheses —- that we may be mere BIVs, or idle dreamers -- undermines our
claims to know. For given that knowing requires the ruling out of every uneliminated
possibility we are not proper in ignoring, we will never fulﬁll the necessary conditions of
knowledge, at least not when the sceptic is present.

Importantly, this last rule is a contextualist one: it makes the sceptical alternatives

alternatives relevant only in those contexts in which we attend to them. This

84

contextualized semantics, therefore, once again allows not only that the sceptic has some
degree of success, but at the same time, safeguards our ordinary knowledge claims. And
it is for this reason — the reason of contextualism — that Lewis can say:

That is how epistemology destroys itself. But it does so only temporarily.

The pastime of epistemology [of looking more closely at these strange

alternatives than we ordinarily do6‘] does not plunge us forevermore into

its special context. We can still do a lot of proper ignoring, a lot of

knowing, and a lot of true ascribing of knowledge, to ourselves and others,

the rest of the time.62
Thus, “Elusive Knowledge,” provides us with a kind of answer to how, in spite of the
sceptic’s victory, we can correctly claim to know.

But how satisfactory is this answer? And, has it improved upon the answer he gave in
1979? I turned to Lewis’ 1996 paper after ﬁnding dissatisfaction with his earlier
“Scorekeeping”. I was dissatisﬁed there because, though there are two kinds of rules —
the merely descriptive and the prescriptive — it seemed that Lewis’ Rule of
Accommodation ﬁt only the former category. This was problematic because rules which
are descriptive merely entail, in themselves, nothing about what is objectively correct:
nothing regarding what we ought or ought not do. As I stated there, if such rules are
narrative merely, based only on the descriptions of the way we happen to behave, they are
not justiﬁcatory. For it never follows from the mere facts of verbal behaviour alone that

one is behaving as she ought to behave. As I have argued, one can easily behave as others

do, and in that sense meet the conditions of intersubjective appropriateness, while at the

 

6|Lewis, “Elusive Knowledge,” 565.
”Lewis, “Elusive Knowledge,” 559. (My emphasis.)

85

same time violate the conditions of objective truth.‘S3

Has Lewis managed in this more recent paper to overcome those difficulties? It is
unclear. For, on the one hand, he gives no explicit reference to the source of these rules
or to their justiﬁcation. And, on the other hand, even to the extent which we accept them
without justiﬁcation, they seem somewhat ad hoc and unsatisfactory. Thus, though he
does not locate the origin of these rules in facts about ordinary language (given that he
does not locate them at all), and hence may escape any claim that he errs in the same
manner as the ordinary language philosopher of chapter II, at the same time, he seems not
to improve upon the problems of his earlier paper. I am left without any vision as to the
purported justiﬁcation of these rules, and hence somewhat bewildered as to their real
nature, and what epistemological work they can do.

In addition to bewilderment, I am left in a state of dissatisfaction: for it is only to the
extent that he either begs the question (as with the Rule of Reliability), makes assumptions
(as with the Rule of Conversation), or makes ad hoc exceptions (as with the Rule of
Resemblance) that his rules manage to save our “knowledge.” In fact the diﬂiculties I
mention are so apparent and overwhelming, I almost forget to ask, “What happened to the
infallibility of knowledge?”; an infallibility Lewis argued was essential to knowledge.
Indeed, Lewis himself seems to recognize the unsatisfactory nature of his “solution” when,
near the end of his paper, he asserts:

I have to grant, in general, that knowledge just by presupposing and
ignoring is knowledge; but it is an especially elusive sort of knowledge, and

 

”See the examples of chapter H, pp. 15-20.

86

consequently an unclaimable sort of knowledge.“

Both the early and later paper of Lewis are unsuccessﬁrl, then, in providing a solution to
the contextualist’s sceptical problem: the problem of explaining the success of the sceptic
while guarding our ordinary knowledge. His ﬁrst paper (1979) failed because, like the
ordinary language philosopher, he succeeded in saving ordinary knowledge claims only to
the extent that he failed to separate two logically independent sets of conditions, those of
appropriate and true assertion. His later paper fails because he saves our knowledge there

only by presupposing, ignoring, and question-begging.

Section vi: DeRose and Invariantism

Keith DeRose, I noted earlier, believes that his solution to the problem of scepticism and
ordinary knowledge is superior to that of other contextualists. Is he more successﬁrl?
Let’s return brieﬂy to his Rule of Sensitivity. Does his Rule of Sensitivity avoid the
problems of the ordinary language philosopher of chapter II?

Whether DeRose does avoid those difﬁculties depends in part on the sense in which his
Rule of Sensitivity is a ‘rule’. But it also depends on the extent to which, if any, his Rule
is used to deﬁne the standards according to which we may be said, correctly, to know.

Here one must be a little cautious, because DeRose (unlike Lewis) does not use his Rule

 

6"Lewis, “Elusive Knowledge,” 562. (His emphasis.)

87

to argue our knowledge-claims are safe, at least not directly. Rather, he uses the Rule of
Sensitivity to explain the success of the sceptic. Only indirectly, by arguing the standards
put in place by this rule -- the standards, that is, that explain the sceptic’s success -- do not
apply in more ordinary contexts, does his Rule have consequences for our ordinary
standards of knowledge. So it is only indirectly that DeRose argues through this rule that
our ordinary knowledge claims are safe. Before I address the important question of the
origin of the standards according to which we (ordinarily) do know, and according to
which the standards brought about by the rule of sensitivity do not apply to our ordinary
knowledge claims, we must further examine of the Rule of Sensitivity.

DeRose used his Rule of Sensitivity to explain how the sceptic “drives up” the standards
for knowing to such a degree that we fail to know even those things we ordinarily take
ourselves to know. What is the justiﬁcation for this rule? What is its origin? Is DeRose’s
Rule of Sensitivity a “rule” in the prescriptive sense -- one that prescribes what ought to
happen -- or merely a rule in the descriptive sense, describing what happens when we
make and evaluate claims to know?

Like Lewis’ Rule of Accommodation, DeRose’s Rule of Sensitivity seems merely
descriptive: it offers a description of what takes place when the sceptic either asserts that
we fail to know, or brings forth one of her famous sceptical hypotheses which, she claims,
we cannot know to be false. According to DeRose, it is only when the sceptic makes
some such assertion, that the standards of knowing are raised to a level at which our
beliefs must be sensitive to count as knowing. Thus, DeRose says, the success of the

sceptic is restricted to the context created by her manipulation of the Rule of Sensitivity.

88'

Again that Rule states that:
When it is asserted that some subject knows (or does not know) that p, the
standards for knowledge (how good an epistemic position one must be in
to count as knowing) tend to be raised, if need be, to such a level as to
require S’s belief in that particular p to be sensitive in order to count as
knowledge. 6’

Noting that this Rule of Sensitivity operates in sceptical contexts is not enough,
however, to justify the claim that DeRose, like Lewis and the ordinary language
philosopher violates the distinction between the standards of appropriate assertion and the
standards of true assertion in deﬁning the ordinary standards of knowledge. For, as I have
noted, it is only in a negative sense that his Rule of Sensitivity has implications here for the
ordinary standards of knowledge. The Rule does not purport to offer a positive account
of the ordinary knowledge requirements. The Rule tells us only that at the “low” levels of
ordinary conversation, since the sceptical hypotheses are not at issue, sensitivity is not
required for knowledge, since it is only when the sceptic enters the discussion that the
Rule of Sensitivity is brought into play. Thus, whether we would believe that we have
hands, for example, even if we did not — that is, even if our belief that we have hands is
insensitive — does not affect the truth-value of our claim to know this in an ordinary

context.“5

The Rule of Sensitivity seems, therefore, to limit the sceptic’s success to the speciﬁc

 

“DeRose, “Solving,” 36.

“DeRose assumes throughout that such beliefs are sensitive. But the problem the sceptic brings in
manipulating the Rule of Sensitivity is that our ordinary beliefs become insensitive. For because at
these high standards we fail to know that we are not, for example, brains in vats, and because
moreover, we must be in at least as good an epistemic position to know that O as we are to know that
-H, we cannot, therefore, at these levels know what we ordinarily claim to know.

89

context of her sceptical queries. So one must look elsewhere for a justiﬁcation for
DeRose’s positive claim that we do in fact know those things we ordinarily take ourselves
to know, and for the standards according to which this judgement is made.

What, then, are the requirements of knowing at these low levels? And how does
DeRose arrive at a knowledge of them? By what standard does he know that these are the
ordinary rules? That is hard to say; for DeRose is not very forthcoming here. In fact,
DeRose nowhere speaks of their origin or their justiﬁcation. The most I am able to do,
therefore, is to infer the origin and justiﬁcation of these rules ﬁom what DeRose does say.
What DeRose does say leads me to conclude that the ordinary requirements of knowing,
for DeRose, correlate precisely with the standards which seem to govern what we let pass
as knowledge in everyday contexts. In other words, the conditions of ordinary knowing
are, once again, to be deduced from our ordinary practices. Whether this is in fact
DeRose’s position, I cannot say for certain since DeRose does not give much to go on: he
merely asserts, time and again, throughout his paper, that at the ordinary low levels of
ordinary conversation what we ordinarily take ourselves to know (that we have hands,
for example) is known.

Thus, it seems once again that the contextualist’s standards of knowing are derived from
mere facts of ordinary discourse: what we say, what we do, how we respond, and so
forth.” But again, as with the ordinary language philosopher of chapter H, this is, at best,

an incomplete argument for what the objective standards -- what is truly required -- of

 

67This I have argued is the method of derivation of the ordinary standards for knowledge in Lewis’
1979. And it is a method explicitly advocated in the contextualist David Annis’ paper “A Contextual
Theory of Justiﬁcation”. See, for example, 107.

90

knowledge are. For how can the mere facts of what we happen to say, what we happen
to accept others saying, and what they happen to accept us saying, and so on, establish
the objective conditions of knowing? It seems, as I argued in chapter H, they cannot.

For, it is only if the standards of what we do allow (what I have called the standards of
appropriate assertion) are logically related to the standards of what we ought to allow
(what I have called the standards of true assertion), in such a way that allows us to deduce
the latter from the former, that the mere descriptive facts can provide us with an objective
list of the criteria of knowledge.

But, as my many examples from chapter 11 (of appropriate yet false assertion, of true yet
inappropriate assertion, and so on) have shown, this is precisely what is not possible
because the standards of appropriate assertion are logically independent of the standards
of true assertion: it is entirely possible to speak in accordance with the former while
violating the latter. Hence the former cannot function as a deﬁnitive guide to the latter.
And, therefore, the mere fulﬁllment of the former cannot — in itself— support the claim
that we know what we ordinarily take ourselves to know.

Before concluding my discussion of DeRose, however, I should note that he does
consider a criticism to his position that is very much like my own. He argues that the
invariantist can offer an argument against DeRose’s contextualism, which although it
offers a radically different analysis of ordinary behaviour, essentially parallels his own.
According to the invariantist position DeRose describes, however, it is the standards of
true assertion that remain constant while it is the standards of useﬁrl or warranted

assertion that vary. And, this invariantist argues, in our common sense knowledge claims

91

we mistake the standards of true assertion — which we do not meet — for those of
warranted or useful assertion, which we do meet. DeRose attributes at least the
beginnings of such a view to both Peter Unger and Barry Stroud68 (to whom, as I
mentioned in chapter H, I am very much indebted.) And DeRose goes on to argue that
though anyone can offer such a parallel argument, if she has no other more weighty
considerations on her side, there is no reason to prefer her argument to DeRose’s own.

For, according to DeRose, the invariantist’s parallel argument involves the speakers of
English in widespread and systematic falsehood. And this, he thinks, counts “decisively”
against their position. I will address the extent to which my argument (that it is the
standards of appropriate assertion, not true assertion, which govern our knowledge
claims) is similar to and dissimilar from the argument he describes in chapter V.
Therefore, at this point I would like simply to consider DeRose’s objection to the
invariantist position, namely, that it involves us in systematic and widespread falsehood. I
should like to point out that DeRose’s objection is accurate only if the invariantist takes
precisely the position he describes: in particular, it is only if the invariantist assumes that
it is part of our common sense that we believe we know the things we say we know in
ordinary conversations, that the invariantist criticism of DeRose’s position does indeed
implicate us in systematic and widespread falsehood.

As I shall make clear in chapter V, however, this is not my view of common sense; for I

do not believe it is the correct view of common sense. Rather, as I will argue there, our

 

”The reference is to Peter Unger’s Ignorance: A Case For Scepticism (Oxford: Oxford University
Press, 1975), and Barry Stroud’s The Significance of Philosophical Scepticism.

92

common sense beliefs about our knowledge claims are quite different from what
philosophers ordinarily take them to be. In fact, I shall argue that we do not believe that
we know the things we say we know in ordinary conversations. And, if one’s sceptical
invariantism starts here, rather than from the assumption that we do believe we know such
things, then the invariantist’s argument that it is the standards of appropriate assertion, and
not the standards of true assertion, which vary, will not have us all believing falsely. For,
if we do not believe we know the things we loosely say we know, we cannot falsely
believe we know these firings. Thus, DeRose’s only and purportedly “decisive” objection
is, with respect to a different invariantist position, quite without foundation and
ineffective. Moreover, since this is the only reason he gives for preferring his
contextualism to invariantism, we have no reason to embrace his position.

Finally, let’s look at the contextualism of Stewart Cohen. He, like Lewis and DeRose,
argued that the sceptic’s success is explained by, and limited to, contextually-sensitive
standards of knowing. Moreover, like the other contextualists, he argued that we do
know those things we ordinarily take ourselves to know. And, though Cohen offered
neither “rules” for the making of ordinary knowledge claims nor for the explanation of the
continued hold the sceptic exercises on us, he did, like Lewis and DeRose, argue, that the
standards of knowing change from context to context: that the standards of knowing are
semantically contextual.

Recall that for Cohen, knowledge requires only “intersubjectively good reasons”. That
is, reasons for which there are no intersubjectively evident defeaters. And for Cohen,

which defeaters are evident — and therefore, what determines the standards of knowledge

93

-- is determined by the reasoning powers of the group which claims to know. Thus, if the
group to which I happen to belong happens to be particularly dull, what counts as an
“evident defeater” will differ greatly from the evident defeaters of a more accomplished
group of reasoners. For, while my group may lack the acuity to discern that some piece of
evidence is a defeater to our good reasons, another group may not.

What follows is that the standards of knowledge are only as high as the (logical)
intelligence of one’s group. As Cohen argued, “In general, the standards in effect in a
particular context are determined by the normal reasoning powers of the attributor’s
social group.”‘” It is for this reason that Cohen, like DeRose and Lewis, concluded that
both we and the sceptic can be correct in our respective claims about knowledge: The
ordinary person is (usually) correct in the knowledge-claims she makes, and the sceptic is
correct in claiming the ordinary person knows nothing at all. This can be so because, by
Cohen’s contextualist semantics, what it means to know for the ordinary person is
determined by the degree to which her group can penetrate the opacity of defeaters; while
for the sceptic and her group, being of a less dull sort, the evidentness of defeaters is
inescapable, and hence the requirements of knowing are unmeetable.

Here again, in seeking the foundation, the justiﬁcation, of those standards which Cohen
claims make both our ordinary claims to know true, and the claims of the sceptic
irrelevant, we ﬁnd only description. His initial promise to derive the standards of
knowledge from the concept deliquesces into a discussion of the relative intellectual acuity

of the groups and the standards which they dictate. There is no analysis according to

 

69Cohen, “Knowledge,” 579.

94

which ‘knowledge’ requires of us the fulﬁllment of certain necessary conditions. Rather,
to the extent that he offers an analysis at all, there is only talk of what would or would not
violate the “ordinary standards of reasoning”, which vary according to the ordinary IQ. of
the attributor’s social group.

But why should the standards of knowing thus vary? Why should the requirements of
myself and the group to which I belong be lower than the requirements of the sceptic or a
society of geniuses? The only explanation seems to be that this is in fact how we talk in
ordinary everyday contexts. But if this is the only explanation forthcoming, we once again
have a confusion of what passes as appropriate in ordinary (and extraordinary) discourse
with what is objectively required to make the claims made therein veracious. And, if we
take these standards to extremes there can be no doubt that the two sets of conditions are
logically distinct, and, therefore, no doubt that the one does not follow neatly from the
other.” Once again, therefore, the contextualist fails to save ordinary knowledge because,
in her failure to separate the logically distinct conditions of appropriate and true assertion,
he gives no objective foundation or criteria against which any such claim to know may be

judged.

 

”I have already offered a number of such illustrations in Chapter 11. But suppose, for example, that
I have a group of friends who are all rather dull. Suppose therefore that nearly any defeater which is
evident to you or me is opaque to them. Then, what passes for knowledge in their group is
propositions such as that the United States is the only country in the world, the moon is only just
above, and so on. Clearly, even though the assertion to know these things in their group is
appropriate, the things asserted are neither true nor, therefore, known. Precisely the same kind of
reasoning applies to the group of average persons, when compared, for example, with learned
scientists or philosophers. This, of course is not to say that either the scientists or the philosophers
know anything. It is just to point out that what defeaters are intersubjectively opaque to the average
person might be perfectly evident to the more learned.

95

Section vii: Conclusion

I have claimed that the contextualists make essentially the same critical error as the
ordinary language philosophers. The error of the ordinary language philosopher, as I
argued in chapter H, was that she rested her whole defense of the ordinary standards of
knowing on mere facts of linguistic behaviour. She rehearsed how we respond, when
asked, if we know, how, and why. She offered only these descriptions of how we
ordinarily behave in defense of her claim that the objective criteria are one and the same as
those standards seemingly embodied in our linguistic behaviour. In so doing, I argued, she
failed to separate the standards of appropriate assertion from those of true assertion. For
it was only on the basis of her conﬂation of these that she arrived at her claim to save the
knowledge of the common person. This failure is critical since it is possible, as I showed,
to behave verbally in accordance with the former while violating the latter. Without
further argument (to the effect that the rules of true assertion are derivable from the rules
of appropriate assertion), therefore, the ordinary language philosopher failed to save
ordinary knowledge ﬁ'om the sceptic.

I have shown, I believe, that the contextualist makes precisely the same kind of error.
Lewis, by deriving the Rule of Accommodation — and the corresponding standards
according to which we know —— from ordinary practices; DeRose by locating the
supposed fact that we know the things we ordinarily let pass for knowledge in just those
conversational facts which evidence this; and Cohen, by relativizing what is necessary for
knowledge to what the group perceives as opaque or evident. Each of the contextualists,

then, derives the standards of ordinary knowing from the mere facts of how we happen to

96

behave, reason, or speak. As a consequence, each fails to establish that our ordinary
knowledge claims meet the conditions of truth. Each fails, that is, to prove that we can
know what the sceptic argues we cannot.

Suppose, however, that we ignore the contextualist’s failure to distinguish the now
familiar categories of appropriate and true assertion. And suppose, therefore, that the
standards of knowledge are as low as they seem to be in ordinary conversation. Suppose,
then also that we admit they succeed in proving that we do know what we say we know.
How satisfactory is the contextualist’s conclusion that we know? Not very, I am afraid.
For we know, on the contextualists account, only to the extent that we presuppose (as in
Lewis’ 1979), we assume, ignore, or beg the question (as in Lewis’ 1996), or we rely on
our own lack of acuity (as in Cohen 1986 ). But how satisfactory is it to be able to say
that one knows only because one presupposes or assumes what one cannot argue for, or
ignores or begs the question with issues that are intractable, or hides behind the average
intelligence of her group? Not very satisfactory I think. Indeed, as a victory for common
sense, it is cheap indeed!

In knowing that I am not a mere brain in a vat, or a mere idle dreamer, I want to know
because I can rule out, because I can argue against, because I can provide deﬁnitive
evidence; not because I am sneaky and dull. In fact it gives me no satisfaction to know
that a contextualist would claim I know I am not a brain in a vat simply because I may
presuppose what those around me (no matter how ignorant) presuppose, or ignore
possibilities (no matter how compelling) simply because I cannot rule them out, or

overlook defeaters to my “good reasons” simply because my group ﬁnds them opaque.

97

Indeed, on this deﬁnition of knowledge, it seems the very foundation of what distinguishes
mere belief (even true belief) from knowledge is wholly lost. For a belief based on
ignoring, presupposing, and question-begging is not epistemically different from a belief
based on mere guess-work or prejudice. To call such belief knowledge, in fact, makes

little sense at all.

98

CHAPTER IV:
Relevant Altemativism
In the previous chapter I found that the problems which riddle the contextualist theory of
knowledge gave conclusive reasons for rejecting both the contextualist’s claim to reﬁrte
the sceptic and, therewith, her claim to save —— from the specter of scepticism —
common sense. In this chapter, I examine the relevant altemativist’s epistemological
theory. The relevant altemativist’s theory of knowledge is very similar to the
contextualist’s theory; so similar in fact that Keith DeRose, a prominent contextualist,

”1

calls relevant altemativism a “prominent form of contextualism. Of course, if DeRose is
correct in describing relevant altemativism as a form of Contextualism, then we should
expect relevant altemativism to fail in precisely the same way as its explicitly Contextualist
counterpart? In the present chapter, I examine a number of such “relevant altemativist”
theories of knowledge. I argue that properly conceived the relevant altemativists do in
fact fail in much the same way as the contextualists, though, properly conceived, the

relevant altemativists are more like distant cousins of the contextualists than immediate

family. For though there are in some of the relevant altemativists certain fundamental

 

‘Keith DeRose, “Contextualism and Knowledge Attributions,” (1992): 19. (My emphasis.)
2Hereafter, I shall sometimes refer to contextualist as CT 5, to the contextualist theory of knowledge

as CT, and so on. I shall likewise often refer to the relevant altemativists as RAs, to the relevant
alternatives theory of knowledge as RA, and so on.

99

similarities to the contextualists3 (they allow, for example, a kind of contextuality to the
conditions of knowing and, thereby, provide the foundation for a kind of contextualist
account of knowledge), the most prominent relevant altemativist theories of knowledge
are not contextualist theories in the sense that DeRose has in mind. The failures,
therefore, are not to be attributed to the kinds of similarities DeRose believes the relevant
altemativists have to the contextualists, but rather to a different set of similarities — a set
responsible also for the shortcomings of contextualism —— which he overlooks.

On DeRose’s account, a contextualist is one who stipulates (after chapter III, I dare not
say “argues”) that the conditions of knowing are altered by contextual features, features
which include mere conversational elements. The relevant altemativists, however, argue
that such conversational elements alone can never change the fundamental conditions of
knowing.’ In fact, the relevant altemativist argues that the fundamental conditions of
knowing never change. To the extent that the relevant altemativists differ from the
contextualists on such positive theoretical commitments, they have distinct advantages
over them. For example, though the relevant altemativists - like the contextualists — do
sanction a kind of ﬂexibility in the conditions of knowing, the contextualists allow for
much greater ﬂexibility and thereby (arguably) undermine the semantics of our ordinary
concept of knowledge that much more. Moreover, unlike the contextualists, the relevant

altemativists claim that there are certain conditions of knowing which are necessary,

 

3This is particularly true of Gail Stine. See: “Skepticism, Relevant Alternatives, and Deductive
Closure,” Philosophical Studies, 29, (1976).

‘Strictly speaking, I am not aware of any RA actually arguing this directly. Yet each does imply
this.

100

irrespective of context. To this extent, the relevant altemativist concept of knowledge
accords much better with our ordinary beliefs. Furthermore, and as a consequence of their
commitment to at least one immutable condition of knowing, the relevant altemativists are
better able to maintain the integrity of the belief/knowledge distinction, a distinction
wholly lost on their contextualist cousins. These differences must be kept in mind one is
to appreciate the uniqueness of the relevant altemativist position.

Despite these important differences between the contextualists and relevant
altemativists, however, I ﬁnd that there are number of crucial similarities, similarities not
surprisingly passed over by DeRose. These sirrrilarities lie not only in their respective
approaches to the problem of scepticism, but also in their consequent failures to deal
adequately with that problem. Both the contextualists and the relevant altemativists, for
example, share the common goal of establishing that the sceptic’s Cartesian doubts are
irrelevant to the truth of our ordinary claims to know. To that extent they are, of course,
similar, though not in a way which uniquely deﬁnes their anti-sceptical arguments. More
important is the similarity evidenced in their respective arguments (or lack thereof) for that
irrelevance. (A lack of argument sometimes touted as a virtue, or at least a right. See, for
example, Dretske’s “Entitlement: Epistemic Rights Without Epistemic Duties?”
Philosophy and Phenomenological Researach, Vol. 60, No. 3, May 2000; and DeRose’s
“Ought we to Follow the Evidence?”, Philosophy and Phenomenological Research, Vol.

60, No.3, May 2000). For it is here that each appeals to the same fallacious principle,

101

namely, the Principle of Epistemic Conventionality.s (In this respect, of course, the
relevant altemativists share a fundamental characteristic not only with the contextualists,
but also with the ordinary language philosopher.) Arguably, it is this fallacious mode of
reasoning, of refuting the sceptic, which is the most important characteristic shared by
each of these positions. For it is this fallacious mode of reasoning which proves fatal to
each of their anti-sceptical positions.

The relevant altemativist’s appeal to this fallacious principle, like the contextualist’s and
the ordinary language philosopher’s, evidences a failure to distinguish two logically
distinct sets of standards now familiar to my readers; namely, the set of standards
governing appropriate assertion and the set of standards governing true assertion. In the
ﬁnal analysis, therefore,qI ﬁnd that the RA’s err in precisely the same way as their
Contextualist cousins: in their zeal to defend ordinary knowledge, they too rest too much
on our ordinary linguistic habits; too much that is to refute the sceptic and rescue our

ordinary knowledge from her apparent threat.

Section 1': Relevant Alternatives and Epistemic Operators
There is one notable exception to this simple failure amongst the relevant altemativists,

and that is Dretske. Fred Dretske is probably the most prominent relevant altemativist,

 

”This is something of an oversimpliﬁcation, though it will be cleared up later in my discussion. In
brief, though the RAs do each evidence this mistake — in a way that is fundamental to the details of
their respective positions — Dretske does not make this mistake in at least his early paper, “Epistemic
Operators,” The Journal of Philosophy: 24 (December, 1970). Though, even there, there is reason to
believe that this mistake is behind his overall relevant altemativist strategy.

102

and arguably the most sophisticated.6 His seminal paper “Epistemic Operators”7 was the
ﬁrst to attack the problem of scepticism in terms of this notion of relevant alternatives, and
did so in such a way that the sceptic seemed (almost) to have been deﬁnitively reﬁtted.
I-Iis arguments there, in this early paper (though not the arguments of his later papers), go
beyond the simple-minded appeals of the other relevant altemativists to our linguistic
habits and focus, rather, on the logical behaviour of sentential operators. I devote,
therefore, a greater amount of space here to Dretske than the other relevant altemativists.
I also, therefore, examine Dretske’s anti-sceptical arguments ﬁrst, and only later look to
the arguments of two other prominent RA’s, Alvin Goldman and Gail Stine.8 In my
discussion of Dretske, I ﬁnd that despite his detailed and persuasive arguments, his attack
on scepticism ultimately fails. It fails because in the course of arguing for his anti-sceptical
conclusions, central premises prove false and ﬁrndamental analogies fall apart.

But ﬁrst, what is relevant altemativism? Its distinguishing feature as a theory of
knowledge is its doctrine that the capacity to rule out relevant alternatives (to what we
claim to know) is a necessary condition of knowing.9 It is the relevant altemativist’s
uniform emphasis on this singular necessary condition of knowing that distinguishes their

approach to scepticism from other anti-sceptical theories, and distinguishes it in particular,

 

6His sophistication, compared to the other Ras, will become quite clear by the end of the chapter.

7Dretske, “Epistenric Operators.”

I’Gail Stine, “Skepticism, Relevant Alternatives and Deductive Closure,” Philosophical Studies,
(1976); Alvin Goldman, “Discrimination and Perceptual Knowledge,” The Journal of Philosophy,
73.20 (1976).

S"This feature is unlike the sceptic’s insofar as ‘relevant’ here implies the existence of the non-relevant.

103

from the contextualist’s.lo Knowledge does not — as the philosophical sceptic has
argued — require ruling out every alternative which is incompatible with what we claim
to know. It does not, for example, require ruling out the possibility that we may be mere
brains in vats (or BIV’s), or unwitting victims of an evil and deceiving genius. Knowledge
also does not — contrary to what the contextualist has argued — require ruling out
mentioned, though inane, alternatives. Rather, the relevant altemativist argues, it requires
only the ability to rule out those alternatives which are both incompatible with and

1 What determines relevance, the relevant

relevant to the knowledge claim that is made.l
altemativist argues, is not mere logical possibility, nor is it conversational utterance or
assumption.12 Rather, it is the (environmental) context in which the knowledge claim is
made. ‘3

Thus, when S claims to know that P, for example, there may be any number of

alternatives which are both possible and incompatible with what she claims to know. In

claiming to know that she has hands, for example, the possibility that S is really a mere

 

loAs I will discuss in section v of this chapter, though the CT 5 also think some alternatives are
irrelevant, they do not make ruling out relevant alternatives a general or necessary condition of

knowing, and for that reason, are not, I think, properly characterized as relevant altemativists.

“See, for example, Dretske’s “Epistemic Operators,” Stine’s “Skepticism,” and Goldrnan’s
“Discrimination.”

l2I mean here to hold distinct what is conversational from what is linguistic habit.

Ii‘At this point, the RA position sounds very much like the contextualist’s. But, as will become clear
later on, the positions are quite distinct in virtue of which features of the context they allow to affect
the relevance of any incompatible alternative. Moreover, to say that the determinant of relevance for
the RA is context is something of an oversimpliﬁcation. As I will illustrate, there is actually a bit of
divergence on this issue, and even where there is divergence the actual criteria of relevance are at
times somewhat unclear. At this point, however, to say that the determinant of relevance is context is
an innocuous oversimpliﬁcation.

104

brain in a vat, or that S is really the unwitting victim of a deceiving and evil genius are
examples of such incompatible and possible alternatives. The sceptic, of course, has
argued that each of these alternatives must be ruled out, if S is to correctly claim to know.
The contextualist argued that such alternatives must be ruled out only if they are raised in
the context of a knowledge conversation or if they are part of what the group perceives as
relevant. The relevant altemativist, however, argues that the knowing subject need rule
out only those alternatives which are relevant to P, and such Cartesian alternatives as
these — she argues — are not.

In “Epistemic Operators,” Dretske’s effort to establish that the sceptic is wrong in her
claim that we must rule out such incompatible alternatives proceeds somewhat differently
than the other relevant altemativists. The argument Dretske offers proceeds indirectly and
by analogy. There is no simple-minded appeal to our ordinary linguistic practices which, I
will show, is a central feature of later relevant altemativist reasoning. There is no appeal
to the intersubjectively opaque, or the conversationally raised or assumed, as with the
contextualists. Instead, the focus of Dretske, in this paper, is on the logical behaviour of
operators. The singular purpose of his analysis is to establish that the epistemic operators,
in particular, are not ﬁrlly penetrating. For, ﬁ'om here, Dretske thinks, the refutation of
the sceptic is a fait accompli .

Before we can ﬁrlly appreciate or evaluate Dretske’s argument, one must then ﬁrst be
ﬂuent in the various ways operators are distinguished, one from another. One must, for
instance, be privy to the distinction between an epistemic and a non-epistemic operator,

and also to the distinction between ﬁrlly penetrating, semi-penetrating, and non-

105

penetrating operators. To this end, let us consider the following presentation of some of
the more pertinent examples Dretske offers.“

On Dretske’s view, some operators have the quality of being fully penetrating.15 If 0
stands for the operator, and P the proposition on which 0 operates, then if 0 is fully
penetrating, then the consequences which follow logically from P follow also from O(P).
The following are examples of such fully penetrating operators:

It is a fact that
It is necessary that
It is true that
It is possible that
Thus, If Q is a necessary consequence of P, it follows that: if it is a fact that P, then it is a
fact that Q; if it is necessary that P, then it is necessary that Q; if it is true that P, then it is
true that Q, and so on.“ In other words, no matter what the consequences of P, if we
apply a fully penetrating operator to P, the operator will penetrate or “carry over” to those
consequence of P.

That some operators are non-penetrating is obvious. (Dretske’s example is: though it is
strange that P, and Q is a necessary consequences of P, it is not necessarily strange that

Q") What about the epistemic operators? Are they ﬁrlly penetrating? Is it the case that

if I know that P, and I know that Q is a necessary consequence of P, I know that Q?

 

”To the reader already familiar with these distinctions, I apologize.
l’Dretske, “Epistemic Operators,” 1007.
l“Dretske, “Epistemic Operators,” 1007.
l7Dretske, “Epistemic Operators,” 1008.

106

Dretske argues it is not. He maintains that all of the epistemic operators are semi-
penetrating only: they are such that the operator does not penetrate to all of the necessary
consequences of the sentence it operates on, though they do penetrate to some. In
particular, he argues, it is the case with the operator ‘knows that’."’

Thus, for example, if S claims to Imow that P, and Q is a necessary consequence of P,
and moreover, S knows that Q is a necessary consequence of P, it does not follow that S
Imows that Q, even though she knows that P and knows that Q is a necessary consequence
of P.” ”

To illustrate: Suppose that S claims to know that she has hands (P). A necessary
consequence Q (admitted by most, and, in particular, by Dretske) of her having hands (of
P) is that she not be a mere BIV (since, obviously, mere BIVs have no limbs or
appendages at all). Suppose further that S knows that Q (she is not being a mere BIV) is a
necessary consequence of P (her having hands). It does not follow, according to Dretske,
that she knows that she is not a mere BIV. For according to Dretske, the operator knows
that does not penetrate to such consequences. Therefore, even if S knows that P, and
knows that Q is a consequence of P, she need not know that Q.

Clearly, if Dretske’s analysis of the operator knows that is correct, and, in particular, if
she is right that one may know that P, and know that P entails Q, without knowing that Q,

this has important implications for the relevance of the challenges posed by the sceptic.

 

l8Dretske, “Epistemic Operators,” 1009.

1”For the sake of argument and simplicity, for now I will assume that if knows that is only semi-
penetrating then it does not penetrate to any of the known consequences of P.

”Dretske, “Epistemic Operators,” 1010.

107

For the sceptic has argued, in effect, that if S is to know that P, and if P is known to entail
Q, then she must know that Q (where the Q’s in question, of course, include not being a
BIV, not being the victim of an evil genius, and so on). In fact, according to Dretske,

21

almost all sceptical objections trade on this pattern of argument He maintains that:

Whether it be a question of dreams or demons, illusions or fakes, the same

pattern of argument emerges. If you know this is a chair, you must know

you are not dreaming (or being deceived by a cunning demon), since its

being a (real) chair entails it is not simply a ﬁgment of your imagination.”

Whether Dretske is correct or not about how prevalent this pattern of reasoning is, the

pattern is certainly familiar enough that Dretske’s argument against it must be taken
seriously.” For Dretske, if successful, will have shown the Cartesian kinds of doubts and
alternatives that are thought to threaten our ordinary knowledge are actually irrelevant
thereto. As he states: Whether or not a certain objection is relevant depends on whether
the consequence is one to which the operators penetrate. 2" Moreover, he states:

What [he] wish[es] to argue . . . is that the traditional sceptical arguments

exploit precisely those consequences of a proposition to which the

epistemic operators do not penetrate, precisely those consequences which
distinguish the epistemic operators from the fully penetrating operators.”

 

2'Dretske, “Epistemic Operators,” 1011.
22Dretske, “Epistemic Operators,” 1011-1012.

23Clearly many sceptic’s have argued for the much stronger thesis that to know that P one must
know all the consequences of P, whether these are known to follow from P or not. But since Dretske
conﬁnes himself to refuting the weaker version of this thesis (knowing the known consequences of P)
I will do the same. See Dretske, “Epistemic Operators,” 1010.

2‘Dretske, “Epistemic Operators,” 1012.

2’Dretske, “Epistemic Operators,” 1012. (My emphases.)

108

Insofar as these kinds of consequences are the primary ones which have proved both
enigmatic and retractable to the history of epistemology, Dretske’s anti-sceptical
argument, if successful, will then have lain to rest a substantial source of philosophical
doubt

Dretske calls the thesis behind this pattern of reasoning the thesis of penetrability.26 27
For, in eﬁ‘ect, the pattern rests on the assumption that the epistemic operator ‘knows that’
penetrates to all the known consequences of the object of knowledge: Such that if Q is a
consequence of P, and S knows that P, and knows that Q is a consequence of P, then S
must know that Q, else she is only incorrectly described as knowing that P.

Up to this point in his reasoning, however, Dretske, has merely stated how one might
undermine the sceptical challenges to our knowledge claims. He has stated that if the
epistemic operators do not penetrate to the kinds of alternatives the sceptic appeals, then
those alternatives will prove both irrelevant to and, therefore, inefﬁcacious against our
ordinary claims to know. The question is: how does he proceed from here? How does
Dretske establish that the epistemic operators do not penetrate to such consequences?

Dretske’s argument is fairly complex (much more complex than that offered by any
other relevant altemativist). As I mentioned earlier, he proceeds indirectly and by analogy;

arguing ﬁrst that the epistemic operators have two characteristics in common with what he

 

26Dretske, “Epistemic Operators,” 1011.

27This same thesis is also, of course, discussed under the name principle of deductive closure, or
simply deductive closure by both Stine and Nozick, among others. See, for example, Stine’s
“Skepticism,” and Nozick’s Philosophical Explanations, Cambridge: Harvard University Press,
(1981).

109

calls the explanatory operators,28 and arguing then that because the explanatory operators
possess the third characteristic of not penetrating to the “contrast consequences” of
explained propositions, the epistemic operators probably (and by analogy) also possess
this characteristic.” Signiﬁcantly so. For if they do share this characteristic with the
explanatory operators, they will be quite ineﬁicacious in the sceptical arguments. For,
Dretske argues, it is precisely such contrast consequences which the sceptic argues we
must know if we are to know anything at all.

But what characteristics do the epistemic operators share with the explanatory
operators? The ﬁrst, and simplest to prove, is the characteristic of semi-penetrability. Or,
I should say, not non-penetrability. For, as Dretske argues the epistemic operators are at
least more than non-penetrating. This can be seem ﬁ'om elementary logic. If, for example,
S knows that P and Q, it follows, necessarily that S knows that Q, a logical consequence of
what she knows, namely P and Q. 3° More importantly, however, and less obviously,
Dretske argues that the epistemic operators share with the explanatory operators the
further characteristic of not penetrating to the “presuppositional consequences” of a
proposition.31 What precisely a “presuppositional consequence” is (or perhaps I should

say why a presupposition is considered to be a consequence), is not particularly clear.

 

28Explanatory operators include the operators is a reason (justiﬁcatory) for, is a reason
(explanatory) for, and would not be the case unless. Dretske, “Epistemic Operators,” 1017.

2‘The analogy is admittedly weak, though at times persuasive. There remains, however, a somewhat
stronger analogy hidden within this rather weaker more explicit one. This will be detailed forthwith.

3"Dretske, “Epistemic Operators,” 1009.
31Dretske, “Epistemic Operators,” 1012-1014.

110

Such a consequence, however, appears to be what most of us would call simply a
presupposition. Dretske, however, views presuppositions as a “class of consequences to
which the epistemic operators do not penetrate.”32 And because establishing that the
epistemic operators do not penetrate to such presuppositional consequences is a central
element in his analogy, one must examine with care how he claims to do this.

Dretske claims to establish that the epistemic operators do not penetrate to
presuppositional consequences with two examples. Dretske’s examples focus on the
epistemic operator realizes that and purport to establish that a given proposition can be
realized, even though a necessary presuppositional consequence of that proposition is not.
The examples, therefore, serve to undermine the general claim that the epistemic operators
penetrate to all the consequences of a proposition: If they do not penetrate to all of the
presuppositional consequences, they do not penetrate to all consequences. Moreover,
and more importantly, the examples serve to establish that the epistemic operators share
this second characteristic — of not penetrating to the presuppositional consequences - with
the explanatory ones.33

Dretske’s ﬁrst example: Dretske asks us to imagine that he is traveling on a train with
his brother. And, as is typical of rude siblings (of which, I must say, I have none), his
brother refuses to give up his seat to an older lady who boards the train. Dretske

proposes that the correct description of the older lady’s perception of his brother’s

 

32Dretske, “Epistemic Operators,” 1014. (My emphasis.)
33I do not address the issue of whether the explanatory operators penetrate to the presuppositional

consequences because inasmuch as Dretske fails to establish the epistemic operators do, this is
sufﬁcient to undermine his claim that the two sets of operators share this characteristic.

111

behaviour is the following: she realizes that" his (Dretske’s) brother will not move.35

Importantly, however, a (presuppositional) consequence of what the older lady realizes
is that the man on the train is Dretske’s brother. And though, according to Dretske, this is
a necessary consequence of what she knows (that his brother will not move), the lady need
not know this (that this is Dretske’s brother) to know what she does know (that his
brother will not move). Thus, Dretske concludes, the epistemic operator realizes that
does not penetrate to the presuppositional consequences, and therefore, does not
penetrate to all consequences.

Dretske’s next example takes the same form as the last and is meant to establish the
same thing. When I have a reason to believe that the Church is empty, he says, a
necessary consequence of what I have a reason to believe is that the building (which I
have a reason to believe is empty) is a church. But, he says, I need not have a reason to
believe this (that the building is a church) to have a reason to believe that the church is
empty. Therefore, again, the epistemic operator reason to believe that does not penetrate
to all the consequences of a proposition. In particular, it does not penetrate to its
presuppositions.

Dretske realizes, of course, that it is not the presuppositions of a proposition that the
sceptic is interested in. So, one must ask, what is the point of proving of epistemic
operators that they do not penetrate to the presuppositions of a proposition? Again, the

answer lies in Dretske’s desire to establish that at least some of the epistemic operators

 

34At times he says knows that. See, Dretske, “Epistemic Operators,” 1013.

3’Dretske, “Epistemic Operators,” 1013.

112

share some characteristics with the explanatory operators; for it is only from here that he
can argue analogically that they probably also share others. Establishing the latter is quite
important since it is only if he can establish a likeness of the epistemic operators to the
explanatory operators that he can proceed analogically to claim that the epistemic
operators probably also share the characteristic of not penetrating to the contrast
consequences of the proposition which they operate upon. It is this latter class of
consequences which Dretske believes the sceptic exploits, and to which he argues the
epistemic operators do not penetratef“6 Hence establishing that the explanatory operators
do not penetrate to the presuppositions is really only a secondary move; one which enables
him to get to the class of consequences he is really interested in, namely, the contrast
consequences.”

Why does Dretske think that such contrast consequences are exploited by the Sceptic?
And how might this exploitation be related to the thesis of penetrability; a thesis he claims
is also ﬁmdamental to the sceptic’s challenge? “The exploitation,” Dretske claims,
“proceeds as follows”:

Someone purports to know that X is A, that the wall is red, say.
The sceptic now ﬁnds a predicate ‘B’ that is incompatible with
‘A’. In this particular example, we may let ‘B’ stand for the
predicate ‘is white’. Since ‘x is red’ entails that ‘x is not white’ it
also entails that x is not-(white and Q) . . . . Therefore the sceptic

selects a ‘Q’ that gives expression to a condition . . . under which
a white wall would appear exactly the same as a red wall . . . let

 

3“Dretske, “Epistemic Operators,” 1015.
37A ‘contrast consequence’ is a consequence of what we claim to know which is incompatible with

(or ‘contrasts’ with) what is known. Therefore, for example, it is a consequence of my knowing that I
have hands that I am not a BIV (since, among other things, BIVs do not have hands).

113

‘Q’ stand for: cleverly illuminated to look red.38
The sceptic, of course, proceeds to argue that unless S knows that not (B and Q), she
cannot know that A. For a necessary consequence of A is that -(B and Q). This
reasoning, Dretske argues, rests on the thesis of penetrability: that if S knows that P, and
S knows that P entails Q, then S must know that Q.39 Or, stated differently, if P entails Q
and S does not know whether Q is true or false, it follows that S does not know that P.
The question, Dretske claims,40 is whether we do in fact ever know any such thing?“ Do
we, when we claim to know that the wall is white, know that it is not merely cleverly
illuminated to look white? A possibility the sceptic argues we must know the falsehood of
to know that the wall is red. Dretske argues that though we habitually take such things
for granted, in particular we habitually assume that the conditions under which we make
our knowledge claims are normal, we do not have sufﬁcient evidence to know any such
thing.

Importantly, he concedes this much to the sceptic. Moreover, Dretske goes on to
illustrate his agreement with the sceptic on our lack of such general knowledge with his
notorious zoo example. Suppose, he says, you take your son to the zoo. Suppose,
moreover, that while there you point to a group of zoo animals and profess to know that

they are zebras. One consequence of the zoo animal’s being zebras is that they are not

 

38Dretske, “Epistemic Operators,”1015. I think this description distorts the sceptic’s motivation,
though since this does not affect Dretske’s overall argument, I won’t pursue it here.

39Dretske, “Epistemic Operators,” 1011.
”Though, as will become evident, this is not the question all relevant altemativists would ask.
"Dretske, “Epistemic Operators,” 1015.

114

merely mules cleverly painted to look like zebras. Yet, Dretske concedes, even though
their not being cleverly painted mules is a consequence of what you purport to know, it is
not something you do know. You know, he argues, no such thing. At least, ex hypothesi,
you have not tested the animal’s hair for paint, and you know sufﬁciently little about
zoology that you could easily be fooled by a painted mule into believing it is a zebra. Yet,
in spite of your ignorance, Dretske claims, you do know it is a zebra.“2

The point of Dretske’s example is general: when we make a claim to know, normally we
do not know that the conditions under which we make our claim are normal. We do not
normally check, for example, whether we might be dreaming. We do not inspect for
phony lighting when we enter a room. We cannot (and do not try to) prove we are not
mere brains in vats. Dretske concedes all of this. Nonetheless, Dretske argues, we do
know the things we claim to know: that the animals in the zoo are zebras, that the wall is
white, and so on. Ifhe is to support these contentions, Dretske must, however, ﬁnd a way
to reject closure. For unless it is not the case that a condition of knowing that P is
knowing all the known logical consequences of P, he must, having conceded we fail to
know these, admit that we fail to know generally. (In this Dretske differs markedly from
the contextualist, whose strategy here is, as described by DeRose a “non-heroic Moorean”
one. See DeRose, “How Can we Know we are not Brains in Vats?”, The Southern
Journal of Philosophy, Vol. 38, 2000.) It is here, of course, that Dretske’s analogy is
supposed to do its important work.

Because of the analogical form his argument takes, Dretske must ﬁrst establish that the

 

”This, of course, is quite reminiscent of Austin’s goldﬁnch example. Cf chapter 11 of this work, 8-11.

115

explanatory Operators (explains that, is a reason that, would not be the case unless, and so
on) do not themselves penetrate to contrast consequences in order to establish the ﬁsrther
conclusion — by analogy - that the epistemic operatorss do not penetrate to those
consequences. He must, that is, establish that the explanatory operators do not penetrate
to the particular type of consequences which the sceptic “exploits.” For from here,
Dretske thinks, by analogy, we may conclude that the epistemic operators do not
penetrate to contrast consequences either and, therefore, that the thesis of penetrability —
upon which the ediﬁce of scepticism is built — is false.

Dretske, to this end, presents us with another example.‘3 Suppose, he says, that after
having dinner with Brenda, I notice that Brenda orders no dessert. There are a number of
contrast consequences (consequences, that is, that contrast with the fact that she did not
order desert) which follow necessarily from the fact that Brenda did not order dessert. It
is implied, for example, that Brenda did not order dessert and throw it at the waiter. It is
also implied that Brenda did not order dessert and take it home to a sick ﬁiend. When,
however, I explain why Brenda did not order dessert by remarking that she was not
hungry, I do not explain either of the aforementioned consequences, even though they are
both necessary (contrast) consequences of the fact that she did not order dessert; implied,
that is, by the fact I have explained. We have then the following set of entailment
relations:

Let P stand for the fact that Brenda did not order desert.

Let Q stand for the (contrast) consequence of P, that Brenda did not order desert and
throw it at the waiter.

 

"Dretske, “Epistemic Operators,” 1021.

116

Let R stand for any consequence at all.
Q is a (contrast) consequence of P, and
O(P) does not entail O(Q), nor O(Q&R).

That is to say, the explanatory operator explains why does not penetrate to the contrast
consequences of the proposition upon which it operates. For, even though it is implied
that she did not order desert and throw it at the waiter, explaining the fact that she did not
order desert does not explain why she did not order and throw it at the waiter; it does not
explain one of the consequences of what it does explain. Returning to his analogy,
Dretske therefore concludes that the epistemic operators also do not penetrate to the
contrast consequences of the proposition upon which they operate. As he reasons: the
epistemic operators share each of two characteristics with the explanatory operators (they
are not non-penetrating, and they do not penetrate to the presuppositional consequences,
of the propositions upon which they operate). Therefore, he reasons, it is likely that the
epistemic operators share also this third characteristic possessed by the explanatory
operators, viz., they also do not penetrate to the contrast consequences.“

At this point, I think, Dretske’s analogical argument appears so weak as to stand in need
of no refutation. For those, however, who ﬁnd Dretske’s analogy persuasive, I offer the
following objections. First, even if Dretske has established his claim that the epistemic
operators share the ﬁrst characteristic (not non-penetrability) with the explanatory

operators, the illustrations he uses to establish that they share the second characteristic

(not penetrating to the presuppositional consequences) fail.“5 Recall his proof took the

 

“Dretske, “Epistemic Operators,” 1022.

”In so far as I am right about this, it would be a waste of paper to discuss whether or not he
succeeds in establishing this for the explanatory operators. For, if he fails to establish that the

117

following form: S e ’s that P (where ‘e’ is any epistemic operator); a presupposition of P is
that Q; yet S need not ‘e’ that Q. Thus he argued that though the lady on the train
realized that my brother would not move, she did not (need to) realize that the person
who would not move was my brother.

But what grounds does Dretske offer for claiming that what she realized was that his
brother would not move? He offers no argument. Merely assertion. Yet most of us, I
think, would reject this description of what the lady realized. We would reject it because
it is false. By hypothesis, the lady on the train had no idea what relation Dretske’s brother
stood in to him, if any. It is, therefore, simply incorrect to describe her as knowing or
realizing that Dretske’s brother would not move. The correct description, given her
ignorance of Dretske’s familial relations, is this: she realized that the man sitting to
another man’s side would not give up his seat. Without further information, she could
not, I think, realize anything more.

Hence, even if the explanatory operators happen to possess the characteristic of not
penetrating to the contrast consequences of the proposition upon which they operate,
Dretske has not proven that the epistemic operators do. He has also, therefore, not
established that the two sets of operators share two characteristics, but one. This,
however, is surely not reason enough to accept his conclusion that they probably also
share the characteristic of not penetrating to the contrast consequences. Yet the minimal

likeness of the two sets of operators to one another is the only reason Dretske sufﬁciently

 

epistemic operators have this characteristic, whether the explanatory operators have it is moot: their
doing so would contribute nothing to the analogy.

118

defends for concluding that the epistemic operators do not penetrate to the all important
contrast consequences.

Though unnecessary, I think, in light of my ﬁrst objection, one might offer the further
objection that Dretske fails also to establish that the purported consequences to which the
epistemic operators do not penetrate are consequences, and therefore, fails not only to
show that they share this characteristic with the explanatory operators, but also to provide
a counter-example to the thesis of penetrability. For, even if we allowed his description of
what the lady realized on the train, his own description of such presuppositions as a class
of consequences is quite unconvincing. Presuppositions are not consequences. They are
presuppositions.

Finally, of course, there is the problem of the weakness of the form Dretske’s argument
takes. For even if he had established that the epistemic operators shared two
characteristics with the explanatory operators, such minimal likeness is a very weak
footing upon which to base his conclusion that they share the third characteristic, not
penetrating to the contrast consequences. It assumes, among other things, that the two
“sets of operators share the same logic in this respect?“ Indeed, his entire reﬁrtation rests
on this assumption. But what grounds are there for believing that these distinct sets of
operators — the epistemic and the explanatory — share the same logic here? Dretske
provides a partial, though somewhat opaque and unsatisfactory, answer. His answer is
articulated in the context of his discussion of the explanatory operators and their failure to

penetrate to the contrast consequences of the propositions upon which they operate.

 

“Dretske, “Epistemic Operators,” 1019.

119

In discussing the explanatory operator explains that, for example, Dretske argues that
when we give an explanation of a particular fact, what is being explained is a function of
two things:

1. The fact

2. The Range of relevant alternatives47

Thus, for example, when I explain why Brenda did not order dessert, I explain why she did
this (the fact) as opposed to ordering dessert and eating it (the relevant alternative).
Hence, my explanation ‘she wasn’t hungry’ succeeds relative to this fact and this relevant
alternative. However, if we change the relevant alternatives - to include ordering dessert
and throwing it at the waiter, for example — my former explanation is no longer
explanatory. (This, of course, is just another way of saying that the explanatory operators
do not penetrate to contrast consequences. For that she did not order dessert and throw it
at the waiter is a contrast consequence of what I explained originally; namely, that she did
not order dessert. Yet my having explained why she did not order dessert does not
explain why she did not order dessert and throw it at the waiter.) For the fact that she
was not hungry does not explain why she did not order desert and throw it at the waiter.
What I explain, therefore, is rightly argued by Dretske to be a ﬁinction of these two
factors. I cannot change the relevant alternatives (or, of course, the fact) without thereby
changing both what I explain and whether the explanation I have given is explanatory.

This much, then, is clear: with respect to explanation, when I change the range of

 

"Dretske, “Epistemic Operators,” 1021. As a point of historical interest, this is the ﬁrst occurrence
of the phrase ‘relevant alternatives.’

120

alternative facts, I change thereby the explanandum. In so doing, I change also what will
count as adequate explanans. Dretske goes on, however, to conclude, without further
argument, “So it is with our epistemic operators”:“8

Within the context of explanation and within the context of our other

operators, the proposition we operate on must be understood as embedded

within a matrix of relevant alternatives."9

To know that x is A is to know that x is A within a framework of relevant

alternatives . . . . This set of contrasts [the relevant alternatives], together

with the fact that x is A, serve to deﬁne what it is that is known when one

knows that x is A. One cannot change this set of contrasts without

changing what a person is said to know when he is said to know that x is

A‘so
Dretske — in the above cited passages - has argued only that when I change the range of
relevant alternatives, I change what I am purported to know. Thus, for example, while at
the zoo, if I claim to know that the animals are zebras as opposed to gazelles, giraffes, or
bison, I claim to know something different from what I claim to know when I change the
range of relevant alternatives to include cleverly painted mules. The change in alternatives
causes a corresponding change in what I am purported to know.

To complete the analogy, however, and establish that the sceptical alternatives are not

relevant to the truth of my knowledge claim — to what I am (intending) to claim to know
(less they change the very thing that I claim to know), Dretske needs to argue that

changing the range of relevant alternatives changes whether I know what I claim to know

(just as changing the relevant alternatives in the case of explanation changes whether my

 

“Dretske, “Epistemic Operators,” 1022.
”Dretske, “Epistemic Operators,” 1022. (My emphasis.)
”Dretske, Epistemic Operators,” 1022.

121

proposed explanation is explanatory); not just what I know. Without this ﬁirther
argument the analogy is incomplete. Dretske, however, does not provide ﬁirther
argument. He seems to think that it is obvious that this ﬁirther conclusion follows.

But how, exactly, might the (purported) fact that changes in the relevant alternatives
change what I claim to know establish that those alternatives are irrelevant to my
(original) knowledge claim and hence to whether I know that? I think Dretske’s reasoning
is as follows: Suppose I claim to know that ‘S knows Lefty killed Otto.’51 What I claim S
knows is a ﬁinction of which alternatives are relevant to the fact S is purported to know.52
Dretske illustrates this in terms of the variety of emphases we might place on the
proposition that S is purported to know. Thus, for example, I might be claiming of S that
he knows the identity of Otto’s killer (that Lefty killed Otto). The relevant alternative is
that a non-lefty killed Otto. Or, I might be claiming of S that he knows Otto was killed,
rather than merely injured (S knows Lefty killed Otto). What I claim of S to know here is
different from what I claimed S knew when I claimed he knew that Lefty killed Otto. And
so, Dretske thinks, are the conditions of knowing. For if what I claim to know is changed
by the changing relevant alternatives then so must be, he thinks, the conditions according
to which my claim to know should be evaluated. If I don’t (intend to) claim to know that
Otto was not killed by someone other than Lefty, then I do not have to rule this alternative

out. Relative to what (I intend to) claim to know, this alternative is not relevant, even

though it is a necessary consequence of what I claim to know. At least this must be the

 

"Dretske’s example, “Epistemic Operators,” 1022.

”I think what Dretske means here is this: what I intend to claim to know is a function of which
alternatives I intend to rule out/view as relevant.

122

case If the epistemic operator knows that ﬁmctions in the same way as the explanatory
operator explains that.

When I claim to know that the animals in the zoo are zebras, then, as opposed to
gazelles, I need not rule out that they are merely painted mules cleverly disguised to look
like zebras. Not being part of my (intended) set of relevant alternatives, such an
alternative, Dretske thinks, is irrelevant to what I claim to know, and, therefore, to
whether I know what I claim to know. For, Dretske maintains, in claiming that the
animals are zebras as opposed to gazelles, I am not (and normally neither are you)
claiming to know that they are not merely cleverly painted mules. Hence even though
their not being cleverly painted mules is a necessary consequence of what I claim to know,
it is not a relevant alternative and hence not something I must know in order to know that
they are zebras. Hence the fact that I cannot know that the zoo conditions are normal
does not undermine my claim to know that the zoo animals are zebras. So Dretske
argues.

What I am suggesting is that we simply admit that we do not know some of
these contrasting ‘sceptical alternatives’ are not the case, but reﬁise to
admit that we do not know what we originally said we knew.’3

But is Dretske right about this? Is it the case that knowing ﬁinctions like explaining
when it comes to relevant alternatives? Is it the case that if a particular alternative to what
I claim to know is not part of what I (intend to) claim to know, I need not know it to

know what I do (intend to) claim to know? Is it possible for me to selectively intend

which alternatives are relevant to my claim to know and which are not? Is it the case that

 

”Dretske, “Epistemic Operators,” 1016.

123

the thesis of penetrability is false, and the alternatives which are (on all accounts)
necessary consequences of what I claim to know need not be known for me to know what
I claim to know? Do alternatives to what I (intend to) claim to know and changes therein
function like alternatives to what I claim to explain? I think not. Let me explain.

Suppose that I claim to know that I have hands. And suppose further that I do not
(intend thereby to) claim to know that I am not a brain in a vat, but only that I am not an
amputee (a necessary consequence of what I claim to know). Is the possibility that I
might be a mere brain in a vat irrelevant to what I claim to know? I think it is arguable
that it is irrelevant to what I intend to claim to know, but not to whether I do know what I
claim to know.“ For unlike explanations, where what I claim to explain is a function both
of the fact and the intended relevant alternatives, in the case of knowing, while what I
claim to know is a ﬁrnction of the fact and the limited alternatives I have in mind, whether
I know is not, at least not exclusively. Whether I know is admittedly a function of what I
claim to know, but it is also a function of the objective criteria of knowledge which must
be ﬁrlﬁlled no matter what particular I may (intend to) claim to have knowledge about.
Unlike the explanatory operator, therefore, changes in the set of criteria for what I know

do not affect changes in the criteria for whether I know.55

 

“For, if we allow it is possible for me to intend to claim only this much, then arguing that I do not
know more than I intend, does not undermine what I claimed to know. As I will argue, however, it is
not at all clear that ‘know’ functions in such a way as to allow such selective intending.

”Actually, although I do not have time to pursue this here, ‘explaining’ is arguably just like
‘knowing’ insofar as there are certain conditions which must be met if something is to count as an act
of explanation (as opposed, for example, to exclamation, imitation, or denunciation) at all. And these
remain so as to enable us to distinguish the kind of act that is explaining from other kinds of acts (like
exclairning, imitating, and denouncing).

124

To illustrate: when I claim to know that I have hands, no matter what alternatives 1
have in mind which govern what I claim to know, the set of criteria which must be met are
constant, if I am to know anything at all. And, what the sceptic argues, of course, is that
these criteria of knowing in general include the following: I must know that I am not a
handless brain in a vat. I must know I am not constantly deceived by an evil genius. And,
the sceptic argues, I must know these whether my claim to lmow is about hands, ships,
sealing wax, cabbages or kings. For in each case, she argues, a necessary condition of its
being so is that none of these Cartesian alternatives obtain. One does not refute the
sceptic, therefore, unless one also shows that the general conditions of knowing anything
at all (i.e., of correctly describing myself as knowing, as opposed to merely believing) do
not include the kind of Cartesian alternatives that the sceptic argues are ineliminable.

Because the necessary conditions of knowing anything at all are always a constant and
proper subset of the conditions of what I know, (just as the necessary condition of
‘doing’, ‘singing’ and ‘thinking’ anything at all are always a constant and proper subset of
the conditions of what I do, sing, or think, ) shifting what I claim to know never affects a
shift in this (constant and proper) subset of conditions governing whether I know. And
whether I claim to know that x is a tree, a hand, or a wall, I must, the sceptic argues, in
each case know that the object is not phony in any respect. I must know it is not the result
of a mere dream, a manipulative evil genius, or clever lighting. For these simply are the
conditions of knowing. Thus, even with changes in fact (with what I claim to know) the
range of relevant alternatives (with respect to knowing) remains constant and does not

change whether I know, for these alternatives represent general conditions of knowing

125

anything at all. ’6

That this is so comes out readily when, rather than focusing on the intentions of the
knower, we look at the following kinds of assertions the knowing subject might make, or
at least be coaxed into making when the implications of his (and Dretske’s) position are
drawn out. ’7 Take, for example, the father at the zoo. Ifqueried whether he knows the
zoo animals are zebras, he would, of course, say “Yes.” If we go on to inquire whether he
knows they are not painted mules, he will (at least Dretske will) say “No,” he does not.
We have, then, the following conjunctive statement “I know they are zebras, though they
might, for all I know, be painted mules.”

Or, to take another of Dretske’s examples, suppose I place what appears to be an
ordinary tomato before you. Suppose I ask you if you know it is a tomato. You assert
that you do. But, when queried, you also assert that it might be a fake tomato, or merely
an image of a tomato produced in your envatted brain by a deceiving and powerful super
computer, for all you know. We then have the conjunctive statement that “I know it is a
(real) tomato, though (for all I know) it might be a fake.”

The sceptic argues (and I am inclined to agree) that each of these statements is absurd,

and the reason they are absurd is because they are contradictory.58 They are contradictory

 

”This is not to say, of course, that when I claim to know x is a tree I must know it is not a cleverly
painted mule, or that when I claim to know that x is a zebra I must know it is not a cleverly painted
tree. Rather, again, it is to say that no matter what I claim to know, if the sceptic is correct, then I
must always know the object of my knowledge is not phony, an imposter. Whereas when I claim to
have explained something, I need not explain every possible alternative, but only the one I mean to
include in my set of relevant alternatives.

”I owe this insight to Peter Unger, Ignorance, Clarendon Press: Oxford, (1975), 140.
581 owe this also to Unger, Ignorance, 139-40.

126

because a condition of knowing anything at all is knowing that the object of knowledge is
not fake in any way. And a condition of knowing the object is not a fake is knowing that
you are not an envatted brain. A further condition is knowing you are not merely
dreaming. For, by deﬁnition, the objects of your perceptions when your brain is envatted
or dreaming are paradigrnatically fake. The Moral? Dretske’s analogy fails. Why does it
fail? It fails because it does not take into account the fact that certain conditions of
knowing are constant, and do not themselves change with the changing intentions of what
I claim to know.

At this point, Dretske has at least two possible alternatives: ﬁrst, he might try to fall
back on the main analogy; second, he might return to the (purported) fact that I can intend
to know something more limited. Let’s consider the second possibility ﬁrst. Can I, simply
by intending to know only that the zebras are not gazelles, for example, rightly claim to
know that they are zebras, as opposed to gazelles. I think that the sceptical arguments
above make a strong case that I cannot. For, there are certain general conditions on
knowing anything, and if these include knowing that the object of your purported
knowledge is not phony in any way, then one cannot ignore these possibilities. Once you
make a knowledge claim, therefore, no matter how limited you might intend it to be, you
are bound by the general conditions of knowing. And, as I argued above, the alternatives
you do not claim to know might nonetheless a necessary component of the conditions of
what you do (intend to) claim to know, since they represent the conditions of ‘knowing’
anything at all. Thus, to reﬁrte the sceptic, one must show that these general conditions

do not include those which the sceptic has argued they do.

127

Therefore, even if you intend only to claim to know that the 200 animals are not
gazelles, your knowledge claim has a much wider logical implication: in claiming to know
the zoo animals are not gazelles, you thereby — simply in virtue of making a knowledge-
claim — claim to fulﬁll the general conditions of knowing. And, unless you establishe
that these do not include knowing they are not phony in any way, you cannot correctly
claim to know merely that the zoo animals are zebras (as opposed to gazelles) simply by
selective intending. For, just as claiming to have done something implies claiming to have
fulﬁlled the conditions of doing, and claiming to have made something implies claiming to
have ﬁrlﬁlled the conditions of making, so also, claiming to know something implies
claiming to have ﬁrlﬁlled the conditions of knowing, and these, the sceptic argues, go far
beyond the mere ruling out that zoo animals are not gazelles.

What about Dretske’s potential to fall back on his main analogy? Recall Dretske argued
that the epistemic operators share two characteristics with the explanatory operators and,
therefore, probably also share a third characteristic, namely, not penetrating to the
contrast consequences. This analogy was meant to establish that the thesis of penetrability
was false. For if the contrast consequences were not part of what was operated on when
the claim to know was made, then it would be false to assert that if Q is a logical
(contrast) consequence of P, and you know that Q is a consequence of P, you must know
that Q (in order to know that P). This response, however, has been already shown to fail.
I argued above that not only is the general form of the argument very weak, it was not
established that the epistemic operators share this characteristic — not penetrating to

presuppositional consequences — with the explanatory operators. Hence, even if the

128

weakness of the argument form were not a factor, the analogy would still fail, since

Dretske does not establish the elements of the analogy.

Section ii: Does Henry Know? Stine ’s Reason to Believe and the Simple Approach to
Relevant Alternatives

In the preceding pages, I have argued that Dretske’s arguments in “Epistemic
Operators,” though somewhat beguiling and more complex than those offered either by his
contextualist cousins or (as we shall see) his relevant altemativist colleagues, fail. They
fail to substantiate the conclusion that the sceptical alternatives to what we claim to know
need not be known to know what we claim to know, and they fail, therefore, to support
the claim that the thesis of penetrability is false. I now turn, therefore, to the relevant
altemativist positions of Dretske’s early contemporaries, Gail Stine and Alvin Goldman.
The arguments of Stine and Goldman, though ultimately relevant altemativist in nature,
differ signiﬁcantly from the arguments of Dretske’s early paper. Stine, for example, is a
relevant altemativist but maintains that the thesis of penetrability need not be shown to be
false in order to establish that we know what we claim to know. And Goldman argues for
the irrelevance of at least some alternatives without any reference to the logical behaviour
of operators, epistemic or otherwise.

Insofar as these later relevant altemativist arguments differ quite markedly from
Dretske’s early ones, they hold out hope for the RA’s to reach their anti-sceptical position.
To the rue of many an anti-sceptic, however, I ﬁnd that they too are unsuccessful in their
efforts to refute the sceptic. Not only are they — like Dretske — unsuccessﬁrl in arguing for
the major theses of relevant altemativism, but because they fall back on the simple-minded

129

kind of anti-sceptical argument offered by both the ordinary language philosophers as well
as the contextualists,59 their failures are only too familiar. Moreover, because of the
greater simplicity of their arguments, their failures are both more obvious than Dretske’s
and, therefore, easier to establish.

Like Dretske, both Stine and Goldman argue that knowing is not a matter of ruling out
all alternatives to what we claim to know, but rather, is a matter of ruling out only those
alternatives which are relevant to what we claim to know. And, like Dretske, they argue
that the relevant alternatives are not identical with the set of all possible alternatives, but
instead are a proper subset thereof. Stine and Goldman’s arguments for the irrelevance of
some alternatives are, as I have indicated, much simpler. They proceed not from the
perspective of epistemic logic, but rather, from certain assumptions about ordinary
language and its ﬁrnction in deﬁning what it means to know. It is from this springboard of
ordinary language that both Stine and Goldman argue that at least some alternatives to
what we claim to know are not relevant. Each fails, however, because each rests her/his
conclusions about what it means to know on our ordinary linguistic practices. Each fails
also, therefore, to defeat the sceptic and so also to vindicate (what she/he takes to be)
ordinary knowledge.

Stine’s paper, “Scepticism, Relevant Alternatives, and Deductive Closure,” like
Dretske’s “Epistemic Operators,” is a defense of the relevant altemativist view. She — like
he — defends the view that knowing is a matter of ruling out not all alternatives, but all

relevant alternatives. Unlike Dretske, however, Stine devotes no attention or space to

 

”The arguments I have in mind, of course, are those discussed in chapter 11 of the present work.

130

establishing the view that some alternatives to what we claim to know are irrelevant.
Instead, she simply assumes that the relevant altemativist view is correct, and proceeds
from this assumption to argue that Dretske’s defense of our ordinary claims to know is
wholly unnecessary.

In Stine’s opinion, there is no need to argue that the principle of deductive closure (what
Dretske calls the thesis of penetrability) is false. Because Dretske believed both that we
must concede we do not know the falsity of the contrast consequences implied by what we
claim to know, and that the principle of deductive closure — according to which we must
know these — underlay all (or at least most) sceptical arguments, he had argued that if
scepticism could be defeated, closure must be proven false. Stine, contrariwise, argues
that the thesis of penetrability and relevant altemativism are wholly compatible. Stine
believes we do know the falsity of the contrast consequences, whose falsity the sceptic
says we cannot know. Hence, the question whether we need know their falsity is set
aside. Hence her defense of relevant altemativism takes a much different turn.

In the course of arguing for the anti-sceptical conclusions of relevant altemativism,
however, Stine makes a number of critical mistakes. Her argument for the criterion of
relevance, for example, rests explicitly (and merely) on an appeal to our ordinary linguistic
practices, and her claim that we know the falsity of the Cartesian alternatives is question
begging at best.

What grounds does Stine give for claiming that the thesis of penetrability, the same
thesis which proved so imposing and problematic to Dretske, poses no threat to relevant

altemativism? To understand Stine’s view, we must remind ourselves ﬁrst of a number of

131

features of Dretske’s. Dretske argued that the kinds of alternatives which the sceptic
claims we must know do follow necessarily from the facts we claim to know. Dretske
conceded, moreover, that we do not know those sceptical alternatives to be false. If,
therefore (he thought), we are to know the things we claim to know, we must either show
that we need not know the consequences which follow necessarily from what we do
know, or we must show that we do know them to be false. Conceding, however, that we
do not know those (Cartesian) consequences to be false, Dretske was compelled to argue
that we need not know them to be false, in order to know what we do (take ourselves to)
know. In other words, he argued that though Q (a Cartesian contrast consequence) is a
necessary consequence of P, and S knows that Q is a necessary consequence of P, she
need not know that Q in order to know that P. This, of course, is just to say that he
argued that the thesis of penetrability was false, and hence that our ignorance of the falsity
of the Cartesian alternatives was immaterial.

Stine, however, does not concede that we fail to know that the necessary contrast
consequences of our knowledge claims are false. In fact, she argues, we do know those
alternatives to be false. Moreover, because we know their falsity, she argues, the fact that
they follow necessarily from what we claim to know can pose no threat to our knowledge
claims. Therefore, she argues, insofar as we know these alternatives do not obtain (i.e.,
we know conditions are normal), we need not show the principle of closure is false. For if
we know the falsity of the Cartesian alternatives whose falsity follows necessarily from
what we claim to know, then there is no incompatibility between the principle of epistemic

closure and our common sense knowledge claims. The incompatibility exists only so long

132

as we fail to know the known consequences of what we claim to know. Therefore, the
veracity of the principle of closure — on her view — poses no threat to ordinary
knowing.

If Stine is correct in her assertion that we do know that conditions are normal — that we
know, contrary to what Dretske has argued, that the animals in the zoo are not merely
cleverly painted mules, that we are not mere brains in vats, and so on — this truly is a
revolutionary and far-reaching achievement. For insofar as scepticism about knowledge
rests fundamentally on the claim that we cannot know that conditions are normal, but
must, if we are to know anything at all, Stine will have (as Dretske intended) laid to rest
most, if not all, sources of philosophical scepticism. Though intriguing, Stine’s
arguments for these philosophical conclusions are inherently disappointing. Indeed, they
are not arguments at all.

Stine’s “proof” that we know the Cartesian alternatives are false, for example, consists
simply of two assumptions: 1) Evidence is not always needed for knowledge. In
particular, it is not needed (as Dretske has argued) to know that the conditions under
which we make our knowledge claims are normal; and 2) We know what we take for
granted. Stine gives no argument for either of these claims. She simply asserts them.
That is it: Stine’s proof that we know the falsity of the Cartesian alternatives, and so need
not know the falsity of the principle of deductive closure, in an unreduced nutshell.

When it comes to establishing that certain alternatives to what we know are irrelevant,
Stine’s paper is weaker still. To begin with, Stine again assumes without argument that

the relevant altemativist position is correct. She offers no proof or evidence for the thesis

133

that some alternatives to what we claim to know are not relevant.60 Her delineation of the
relevant ﬁom the non-relevant alternatives is equally without merit. She distinguishes the
relevant ﬁom the non-relevant alternatives in two ways: ﬁrst, she appeals to ordinary
language, and second, she rejects the only (ﬁ'om her standpoint) alternative relevant
altemativist position of the determinants of relevance.

Stine’s ﬁrst approach is ﬁamed in the context of a discussion of how we might know if
an alternative is relevant or not. With respect to the simple cases, Stine merely asserts the
following:

in ordinary life, we do exhibit rather strong agreement about what is
relevant and what is not.61

Clearly, in the context of her discussion, this singular observation is considered sufﬁcient
to establish which alternatives are relevant and which not. But like every argument which
draws conclusions about what is objectively the case from what happens as a matter of
fact to be the case, Stine’s ‘argument’ is subject to the familiar criticism that it conﬂates
the standards of appropriate with the standards of true assertion. Stated differently, she,
like the ordinary language philosophers and contextualists before her, appeals to the
principle of epistemic conventionality, arguing (by implication) that the mere facts of

linguistic habit alone are revelatory of the objective criteria of knowing. In so far, then, as

 

”The absence of any argument here strikes me as rather astounding given the relative newness of
the RA position at the time Stine wrote this, and also given her refutation of Dretske’s arguments, one
of the only other papers on the subject at the time. Stine does, in a sense, appeal to both Austin and
Armstrong here, but she does not argue that the position contained in the passages she quotes is

correct. Rather, she merely asserts her agreement with it. See 254 and 258. In this respect, Stine’s
response to scepticism is like the self-described (by DeRose) “non-heroic Moorean” responses of the
contextualists.

6|Stine, “Skepticism,” 252.

134

her only support for this criterion of relevance — our agreement in ordinary life — is her
reliance on common linguistic practices, her criterion remains unsatisfactory and
unconvincing.

The second way Stine gives for distinguishing the relevant ﬁom non-relevant
alternatives takes as its subject the grey areas where there is less obvious agreement on
which alternatives are relevant. Here she appeals to the example attributed by Goldman to
Carl Ginet.‘52 She argues that when there are barn facades in the immediate vicinity of
Henry, we count the alternative that what Henry might be seeing is a mere facade as
clearly relevant. When, by contrast, there are no barn facades in Henry’s immediate
vicinity, we reject the alternative that he may be viewing a mere facade as irrelevant.
When, however, there are barn facades neither in the immediate vicinity of Henry nor yet
too far off, she states that the issue of relevance is less clear.‘53

Stine argues that there are two possible relevant altemativist ways for deciding the
question of relevance in such cases: the ﬁrst, a view she defends as her own, states that an
alternative is relevant only if there is a reason to think it is true The second, a view she
mistakenly attributes to Dretske, states that an alternative is relevant if it could be true.
Stine, having mistakenly attributed the second view to Dretske, goes on to argue against
it. She offers two reasons for rejecting the view that relevance is a function merely of
what could be true: First, this criterion of relevance, she argues, would deprive the

relevant altemativist view of its distinguishing feature, namely, the claim that some

 

62See Stine, “Skepticism,” 252.
“Stine, “Skepticism,” 252.

135

possible alternatives -- in particular, the Cartesian alternatives —— are not relevant.
Second, she claims (without argument) that determining relevance by mere possibility is
not as easy as determining relevance by when there is a reason to think an alternative is
true. In light of these two objections, Stine concludes the correct criterion of relevance is
the ﬁrst: an alternative is relevant only if there is a reason to think it is true.‘54

Again, however, Stine’s defense of the relevant altemativist view is disappointing. For,
to begin with, although Stine characterizes the second view of relevance as a possible
relevant altemativist view (in fact, an actual one, insofar as she attributes it to Dretske), it
is quite evident that the view she describes is paradigmatically sceptical, and not,
therefore, one to which any relevant altemativist would subscribe. For the making
relevant of any alternative so long as it is merely possible makes irrelevant only those
alternatives which are logical falsehoods (or at least physical impossibilities).

Hence, because all of the Cartesian alternatives are possible (in the sense that none are
logically contradictory), and we have no reason to think any are physical impossibilities,
the criterion of relevance Stine attributes to Dretske would make relevant all those
alternatives the relevant altemativist view is designed speciﬁcally to reject. Hence, even
though Stine correctly points out that this criterion of relevance would deprive relevant
altemativism of its distinguishing feature, it is something of a straw man to argue it is a
possible relevant altemativist view. Attacking this view then, as a means of supporting her

own, is quite unpersuasive. Moreover, in arguing against this paradigmatically sceptical

 

“One, of course, is reminded of Austin’s conditions of appropriate assertion. See chapter I] of this
work.

136

view of relevance, her reason why it should be rejected — because it would allow sceptical
alternatives to be relevant — is clearly question begging.

Stine’s second reason for rejecting the logical possibility criterion of relevance
(according to which an alternative is relevant so long as it is possible) is quite
unmistakably false. For it is not, as she claims, difficult to decide if an alternative is
logically possible: we simply assess whether it is logically false.65 On the other hand,
whether there is some reason to think that an alternative is true is a much more
complicated matter. Moreover, even if it were more diﬂicult to assess logical (or physical)
possibility, this is hardly an objective basis for ruling out this — or any other — criterion of
relevance. Clearly the fact that something is difficult is not proof that it is false or

incorrect.

Section iii: Goldman and Perceptual Discrimination

In my discussion of Stine, I noted that it is odd that she leaves the relevant altemativist
view essentially unargued for, especially odd in light of the dearth of extant arguments for
the relevant altemativist position. It is possible, however, that Stine believes Goldman to
have argued adequately for the relevant altemativist position, and relies therefore,
implicitly, on his relevant altemativist achievements. I turn now, therefore, to Goldman’s
defense of relevant altemativism, as presented in “Discrimination and Perceptual

Knowledge.” Like Dretske and Stine, Goldman defends the view that to know that P one

 

"Perhaps proving physical possibility would be more diﬂicult, but I don’t see why this should be so,
except insofar as we can’t know the laws of nature, but this is a difficulty only if we are sceptics, and
Stine clearly is not.

137

need rule out only those alternatives relevant to P. Also like Dretske and Stine, he argues
that at least some alternatives are not relevant to our ordinary claims to know. In a vein
similar to Dretske, Goldman argues that it is a necessary condition of knowing anything at
all that we be able to rule out those alternatives which are relevant to what we claim to
know.“5 Goldman, however, argues that this condition of knowing follows from the more
general requirement that the processes or mechanisms which produce belief be reliable.
A reliable belief forming process, according to Goldman, is one which produces true

beliefs rather than false ones.‘S7 Moreover,

to be reliable, a cognitive mechanism must enable a person to discriminate

or diﬂerentiate between incompatible states of affairs. It must operate in

such a way that incompatible states of the world would generate different

cognitive responses.68
With respect to a perceptual belief forming mechanism,

[it] is reliable to the extent that contrary features of the environment (e.g.,

an object’s being red, verses its being yellow) would produce contrary

perceptual states of the organism, which would, in turn, produce suitably

different beliefs about the environment.”
It is, therefore, an essential and necessary condition of knowing that we be able to

discriminate the state of aﬁ’airs we believe to obtain from other possible states of affairs

that we do not believe to, but might, obtain. As Goldman points out, this view of

 

66I don’t include Stine here because although she argues that ruling out relevant alternatives is a
sufﬁcient condition of knowing, she does not argue it is as necessary condition. In this respect, her
position is much closer to a contextualist one. I will discuss this shortly.

“Goldman, “Discrimination,” 163.

“Goldman, “Discrimination,” 162.

“Goldman, “Discrimination,” 162.

138

knowledge accords with a sense of the verb to know listed in the OED; a sense oﬁ ignored
by philosophers:
“to distinguish (one thing) from (another,)” as in “I know a hawk from a
handsaw” (Hamlet) and “We’ll teach him to know Turtles from Jays”
(Merry Wives of Windsor).70
At this point, Goldman’s views are quite congenial to those of the sceptic.7l For the
view that a necessary condition of knowing is the capacity to distinguish the state of
affairs we believe to obtain from other contrary states of affairs which might obtain is
foundational to sceptical epistemologies. Goldman goes on, however, to distinguish his
view from that of the sceptic, and in the course of so doing, his relevant altemativism is
made clear:
A knowledge attribution imputes to someone the discrimination of a given state of
affairs from possible alternatives, but not necessarily all logically possible
alternatives.72
In the span of a single sentence, then, Goldman commits himself to the central tenet of
relevant altemativism: the view that we need know to be false only those alternatives
which are relevant to our knowledge claim, not all those which are merely logical possible.
But what is Goldman’s justiﬁcation for so qualifying his original contention that

knowledge requires the ability to discriminate between incompatible states of affairs?

What justiﬁes his claim that only some of the class of possible alternatives are relevant?

 

7"Goldman, “Discrimination,” 163.

71In fact, to some extent, his views — unlike those of any other RA — remain congenial to the
sceptic. This be explicitly recognizes, although he states that he remains “oﬂicially neutral” on the
question of scepticism. See, “Discrimination,” 166.

72Goldman, “Discrimination,” 163.

139

Like Stine, Goldman appeals to ordinary language. He justiﬁes the circumscription on
ruling out alternatives by appealing to what we “normally do” and “ordinarily require”:
In forming beliefs about the world we do not normally consider all logical
possibilities."
And,
In deciding whether someone knows that P, we do not ordinarily require
him to discriminate P from all logically possible alternatives."
Clearly what Goldman here asserts is correct: we do not, in fact, ordinarily do or require
either of these things. When, for example, in the company of ﬁiends, someone claims to
know that P, we do not ask of her that she prove that she knows she is not dreaming, nor
do we insist that she establish she knows she is not a mere brain in a vat. Equally, we do
not impose such requirements on ourselves. But does it follow straightaway — as
Goldman and Stine suggest - that these alternatives are not therefore relevant to the
objective truth of what is claimed to be known? Does it follow that knowing the falsity of
these alternatives is not — objectively — a condition of knowing? A condition of being
correct in professing to know? Clearly it does not. The lesson of both Chapter’s II and
III is that the irrelevance of the sceptic’s alternatives cannot follow — without further
argument — from the mere facts of linguistic habit alone. It is entirely possible for an
individual to meet the standards of ordinary behaviour while failing to meet the standards

objectively required to make that behaviour veracious.

Let us return for a moment to my ﬁiend Rodger’s party, a situation discussed in chapter

 

73Goldman, “Discrimination,” 163-4.

"Goldman, “Discrimination,” 164. Once again, the appeal to what we ordinarily do and require is
reminiscent of Austin.

140

II.75 Recall that upon arriving at the party, Rodger asked of me if John would be coming.
I asserted that I knew he would. Rodger, of course, did not ask me to prove that I had not
merely been dreaming that John was on his way, nor that I was not an envatted brain,
incapable of real contact with John; to do so would have been inappropriate, requiring as
it would more than we ordinarily require when someone makes a claim to know. Thus
Rodger accepted my claim to know and went about his way.

Nonetheless, what I claimed to know proved false. John did not arrive at the party
because, having been struck by a meteorite, he lapsed into a coma and died. Thus,
though my claim to know was perfectly ordinary, meeting the standards of appropriate
assertion, it was nonetheless objectively false. Obviously, though I met the conditions of
appropriate assertion, I failed, nonetheless to meet the conditions of true assertion.
Clearly, therefore, the conditions of appropriate (habitual) assertion and the conditions of
true assertion are not one and the same, and meeting one does not amount to meeting the
other.

The moral, of course, is this: one cannot infer from the mere fact that the conditions of
ordinary assertion are met that the conditions of true assertion are met also. The
illustration establishes the obvious truth that the conditions we ordinarily impose (what I
have called the conditions of appropriate assertion) are logically separable ﬁ'om the
conditions objectively required by the truth conditions of knowing. To infer, therefore,
from the fact that we ordinarily do not require that certain alternatives be ruled out, the

conclusion that we need not rule these alternatives out to have objective knowledge is to

 

75See Chapter II of this work, 16-17.

141

argue fallaciously. This, however, is precisely what both Stine and Goldman do.

That we do not normally consider all logically possible alternatives when we claim to
know something, and that we do not normally require of others that they rule out such
alternatives are the primary justiﬁcations Goldman offers for thus limiting the alternatives
we must rule out so as to exclude at least some of the logically possible alternatives to
what we claim to know. Goldman’s defense, therefore, of his claim that some alternatives
to our claims to know are irrelevant suffers ﬁom the same kinds of difﬁculties as those
which riddled both the ordinary language approach to scepticism and the contextualist.
To the extent that his argument does so suffer, it fails. It fails to establish that some
alternatives are not relevant and, therefore, fails to establish that the sceptical alternatives
in particular are irrelevant. 7"

Goldman has, however, an additional argument to support his relevant altemativist
position, namely its explanatory power. He illustrates the explanatory power of relevant
alernativism by reference to Carl Ginet’s example of Henry and the issue of relevance
concerning barn facades.77 Like Stine, Goldman asserts that when there are no facades
present, we acknowledge that Henry knows that what he sees is a barn. Also like Stine,
he argues that if there are such facades present — unbeknownst to Henry — then Henry

does not know that what he sees is a barn. But how, Goldman asks, can we explain the

 

76This latter, of course, is not strictly a problem for Goldman since he wishes to remain neutral on
the issue of scepticism, but it is a problem for the general RA program which — apart from Goldman
— is fundamentally anti-sceptical.

77Goldman attributes the example to Ginet, though he does not state where Ginet discusses the
example. See Goldman, “Discrimination,” 164.

142

difference in our reactions to the two alternative situations?78 How do we explain that in
the ﬁrst case we say of Henry that he knows, while in the second we say he knows not?
Goldman considers a number of explanations which, he argues, will not do. Among the
explanations he considers is the justiﬁed true belief analysis of knowledge. According to
this analysis, of course, one knows only when she has a justiﬁed, true, belief. Such an
analysis will not explain our different reactions to the alternative situations, however, F
because, Goldman argues, in each case Henry’s belief that what he sees is a barn is both
justiﬁed and true.79 It is with respect to explaining our attribution of knowledge on the

one hand, and our denial of the same on the other, that Goldman believes the relevant

 

altemativist analysis of knowledge both succeeds and thereby gains further support. For
the relevant altemativist analysis can, he claims, unlike other analyses of knowledge,
explain both why we say of Henry that he knows (when there are no barn facades in his
vicinity), and why we say of Henry that he fails to know (when there are, unbeknownst to
him, barn facades in his vicinity).

The correct explanation of our different reactions, Goldman thinks, lies, therefore, in the
respective irrelevance and relevance of the alternative that what Henry may be seeing is a
mere facade. That is, it lies in the principles of relevant altemativism. In the ﬁrst case,
where there are no barn facades in the vicinity, the reason we say that Henry knows, even
though it is of course logically possible that what he sees is a mere facade, is that the

alternative that what he sees is a mere facade is simply not relevant and hence needn’t be

 

7"Goldman, “Discrimination,” 164.
79Goldman, “Discrimination,” 164.

143

ruled out by him in order to correctly claim to know. When, on the other hand, the
environment changes such that there are facades in Henry’s vicinity, the alternative that
what he may be seeing is a mere facade becomes relevant, and thereby, insofar as Henry
cannot rule it out, undermines his claim to know
Goldman’s relevant altemativism, then, does indeed provide an explanation for our

differing responses to the alternative instances where Henry is, in the one case, said to F.
know, and, in the other, is not. Goldman, however, by my lights, still owes us an account
of what affects the change in relevance from one case to the next. Goldman remains

ofﬁcially neutral on the question of whether there is a unique set of alternatives which

 

must be ruled out for any particular claim to know.80 But, on my count, there are three
potential explanations of the change in relevance to be found in Goldman’s paper.

First, there is the fact (agreed upon by both Goldman and Stine) that we do not
ordinarily require such alternatives to be ruled out. Second, there is Goldman’s
discussion of perceptual equivalents, and the requirement that an alternative be a
perceptual equivalent in order to be relevant.”1 Finally, there is the fact that the only
apparent change from one case to the next is the objective presence or absence of an
environmental feature.82

The ﬁrst explanation — that changes in relevance are simply a result of our ordinary

 

”Goldman, “Discrimination,” 166. This leads him also to remain ofﬁcially neutral on the
question of scepticism; a neutrality absent in all other RA accounts.

8|Goldman, “Discrimination,” 166.

82This, we shall see, is Dretske’s criterion of relevance in his paper “The Pragmatic Dimension of
Knowledge,” Philosophical Studies, 40, (1981).

144

attitudes toward relevance — has already been shown to fail. In both Stine and Goldman,
there is no argument beyond the fallacious inference from facts of ordinary practice to
objective requirements. Hence to explain the apparent fact that the facades are not
relevant in the one case but are in the other by reference to the mere fact that ordinarily we
would not require their presence to be ruled out in the ﬁrst case but would in the second,
is not to give an explanation at all.

What about the second possibility? Explaining changes in relevance by reference to the
notion of perceptual equivalents? Goldman argues that a necessary condition of an
alternative’s being relevant is that the alternative represent a perceptual equivalent of what
we claim to know.83 One might think, therefore, that Goldman could explain the
purported difference in relevance by arguing that in the ﬁrst case where no facades are
present, the alternative that Henry is seeing a facade is not a perceptual equivalent,
whereas in the case where facades are present the alternative is a perceptual equivalent.
This explanation, however, makes no sense when we recognize that on Goldman’s ﬁnal
analysis, the alternative in each case does count as a perceptual equivalent and, therefore,
cannot be ruled out as irrelevant on these grounds.

What distinguishes the respective relevance and irrelevance of the facades, must,
therefore, be something beyond the meeting of the condition of perceptual equivalents,
and also beyond the mere facts of linguistic habit. Might the determinant of relevance then

be that in the one case there is no objective feature in the environment corresponding to

 

83Goldman, “Discrimination,” 167.

145

the alternative, but in the other there is? If so, how do we explain why that would affect a
change in relevance? Working with the basic assumptions of the relevant altemativist
view, and what Goldman has given us to go on, I think the only plausible explanation is
that in ordinary life this is pretty much how we would decide whether or not Henry need
know that what he is looking at is not a facade in order to know that what he sees is a
barn. If there were actually circumstances (a practical joke, say) in which a number of
facades had been erected in an area of Henry’s travels, we would require him to rule out
that what he was seeing was a mere facade before we would allow that he knows that
what he sees is a barn. Likewise, normally, where no such facades are present, the mere
logical possibility alone that there could be facades would not sufﬁce to demand of Henry
that he rule such an alternative out.

Yet, this last explanation also fails to make clear what are the objective determinants of
relevance. For clearly to explain the change in relevance in terms of how we would
ordinarily judge the relevance of alternatives is simply another version of the fallacious
appeal to the principle of epistemic conventionality. Hence, the last potential explanation
can fare no better than the ﬁrst. For, if the presence or absence of environmental features
affects relevance merely because this is how we ordinarily make judgements about
relevant alternatives, we have once again an argument from mere facts of how we happen
to behave to conclusions about the necessary conditions of truth. This, of course, is an
argument form whose fallaciousness I have hitherto established.

In spite of the differences in Goldman’s approach to relevant altemativism, I ﬁnd no

146

greater reason for accepting the relevant altemativist picture of knowledge than I found in
Stine. At this point, therefore there is no given reason (at least no good reason) to adopt
the relevant altemativist view of knowledge, nor therefore to ﬁnd refuge therein ﬁom the

threat of scepticism.

IV: Dretske Revisited
Before closing my discussion of relevant altemativism, however, I owe Dretske one ﬁnal

look. “Epistemic Operators” does not represent his ﬁnal word.84 In “The Pragmatic ;

 

Dimensions of Knowledge,” Dretske offers another argument for relevant altemativism, an
argument which makes no mention of epistemic operators, and which is quite independent
of any arguments offered in that seminal paper. In this later paper, Dretske approaches the
relevant altemativist position from a model of concepts which he borrows, in part, from
Peter Unger.85 Dretske argues that ‘knowledge’ is, as Unger argues, an ‘absolute’

t.“ An absolute term, Dretske notes, “depicts a situation as being completely

concep
devoid of a certain thing . . . . The fact that there can be nothing of this sort is what

makes it an absolute concept.”87 As an absolute concept therefore, ‘knowledge’ cannot

 

84Dretske has at least one further paper relevant to the discussion here, but which for consideration
of time I leave out. That is his “Conclusive Reasons,” Australasian Journal of Philosophy, 49, May,
( 197 l).
”Unger, “A Defense of Skepticism,” Philosophical Review 80, (1971).
8”Dretske, “Pragmatic Dimensions,” 364.
87Dretske, “Pragmatic Dimensions,” 366.

147

admit of degrees: one either knows or one fails to know.

Unlike Unger, however, Dretske argues that knowledge is not just an absolute concept,
but a relationally absolute one; as such, it can be sensitive to the changing needs and
interests of the individuals who employ it. It is, Dretske argues, from the relationally
absolute character of the concept of knowledge that the irrelevance of the Cartesian
alternatives follows. For insofar as knowledge has this relationally absolute characteristic,
the alternatives which must be ruled out differ with changes in context. The purpose of
“The Pragmatic Dimensions of Knowledge” is to explore how, and the extent to which,
knowledge exhibits such contextual relativity.88 Dretske states:

In short, I want to explore the way, and extent to which, this absolute
notion exhibits a degree of contextual relativity in its ordinary use.89

A paradigm example of an absolute term (perhaps made paradigmatic by Peter Unger) is
the term ‘ﬂat.’ Flat is an absolute term in the sense that if a particular object or thing is
ﬂat, it can have absolutely no bumps. Hence if a particular surface is ﬂat, it follows that
nothing can be ﬂatter than it. ‘Flat’ is, as Unger says, an absolute limit term; deﬁning, as it
does, an absolute limit which must be met if a thing is to be ﬂat at all. If the limit is not
met completely, then the thing in question is not (really) ﬂat. If a thing really is ﬂat, then
nothing can be ﬂatter. Thus, Unger argues, when we say of one thing that it is ﬂatter than

another, we are really assessing the degree to which the respective surfaces approximate

 

88In the ﬁnal section of this chapter, I will explore the relationship between Dretske’s position here
and that of the contextualists.

”Dretske, “Pragmatic Dimensions,” 365. (My emphasis.)

148

ﬂatness. Because ﬂat is an absolute term, however, Unger argues it has no true
application. Nothing, he says, is really such that it is (absolutely) ﬂat. For to be
absolutely ﬂat is to have no bumps at all, while everything has at least some surface
irregularities, even if only microscopically.

Knowledge, Unger argues, is also an absolute term. Like ﬂat, therefore, it too deﬁnes
an absolute limit. If I know, I must be completely justiﬁed. There must be nothing that
undermines my justiﬁcation. Because of the absolute nature of ‘knowledge,’ anything less
than complete justiﬁcation will not do. So, like ﬂat, given the “epistemic bumpiness of our
world,”90 nothing is truly known, just as nothing is truly ﬂat. Thus, on Unger’s view,
when we claim to know, we are either speaking loosely, incorrectly, or indirectly about the
degree to which we approximate knowing. I cannot truly know, Unger argues, because
there will always be competing possibilities to what I claim to know which, since I cannot
rule them out, undermine my justiﬁcation."

Dretske agrees that ‘knowledge’ is an absolute term. But, unlike Unger, he does not
draw ﬁom this any sceptical conclusions.” In fact, he argues, terms like ﬂat and
knowledge are relationally absolute and hence admit of application according to

contextually relevant standards. Thus he argues that “Although nothing can be ﬂat if it

 

9"The phrase is Dretske’s, “Pragmatic Dimensions,” 367.

9|Unger, of course, later (and regrettably, I think) gives up this view. See, for example, his “The
Cone Model of Knowledge,” and his Philosophical Relativity.

92Dretske, “Pragmatic Dimensions,” 366.

149

 

has any bumps, what counts as a bump . . . depends on the surface being described.”93
Similarly, he argues: “Something is empty if it has nothing in it, but this does not mean
that an abandoned warehouse isn’t empty because it has light bulbs and molecules in it.”9"
What counts as a ‘thing,’ Dretske maintains, depends upon the context and interests of
those who wish to know if the warehouse is empty.

The contextual relativity, then, of these absolute terms, is in part a function of the terms
we use to deﬁne them. It is also, however, a ﬁrnction of the interests of the speakers
involved in the conversation. Whether or not the warehouse has things in it, for example,
depends upon the interest we have in establishing whether or not it is empty. If we are
scientists interested in creating a vacuum, then certain things will count as things in the
warehouse which, if we were merely apartment dwellers looking for a place to store our
goods, would not. Similarly, if we are expert golfers, looking for a ﬂat green on which to
practice our putting, certain minute bumps may disqualify a green from being or counting
as ﬂat which, for the amateur golfer, would go wholly unnoticed and not undermine the
purported ﬂatness of the surface.

What, then, has all of this to do with knowledge, scepticism, and relevant alternatives?

Dretske wants to take this model of relationally absolute terms as a model for knowledge.

In accordance with this approach . . . I propose to think of knowledge as
an evidential state in which all relevant alternatives are eliminated.

 

”Dretske, “Pragmatic Dimensions,” 366.
“Dretske, “Pragmatic Dimensions,” 366.

150

Thus, he argues, just as what counts as a bump in deﬁning what counts as a ﬂat surface
may vary with context, so also what counts as an ‘alternative’ to what we know — a
relevant alternative — varies with context. An alternative may be relevant in one context,
and hence in need of being ruled out, but not relevant in another context and hence not in
need of being ruled out. Of course, it is precisely those Cartesian alternatives appealed to
by the sceptic which Dretske wishes to argue are not relevant, do not count as relevant, in
any ordinary contexts and hence cannot undermine our ordinary claims to know.

This [requiring that we rule out all relevant alternatives] makes knowledge

an absolute concept, but the restriction to relevant alternatives makes it,
like empty and ﬂat, applicable to this epistenrically bumpy world we live

in.95

What, then, does the relationally absolute model of concepts tell us about our ordinary
claims to know? It does not tell us that they are justiﬁed. But it does, if Dretske is
correct, tell us that to be justiﬁed it is unnecessary to ﬁrst rule out the Cartesian kinds of
alternatives appealed to by the sceptic. I need not, for example, know that I am not a
brain in a vat in order to know that I have hands. I need not know that I am not the victim
of an all-powerful, evil and deceiving genius. In fact, I need not know the falsity of any
Cartesian alternatives to know what I ordinarily take myself to know. For, ordinarily,
what counts as a relevant alternative are the less outrageous alternatives we are familiar

with in ordinary life. And it is what counts as known, ordinarily, which determines

relevance.

 

95Dretske, “Pragmatic Dimensions,” 367.

151

On Dretske’s view, then, the fact that I cannot know I am not a mere brain in a vat does
not undermine my (ordinary) claim to know. Moreover, once again, the pattern of
sceptical argumentation seems to be undermined. For the sceptic, on Dretske’s view
reasons thus: If Q (that I am not a brain in a vat, for example) follows from P (that I have
hands, for example), and I know that Q follows from P, it follows that I cannot know that
P is true unless I know that Q. (That is, I cannot know that I have hands unless I can
know that I am not a handless brain in a vat.) Yet Dretske’s argument appears to establish
the contrary: I need not know that Q in order to know that P, even though Q may be a
necessary consequence of P. For Q (and its kind) is an alternative not normally counted
as relevant. In relation to most contexts and interests, not knowing the falsity of -Q does
not normally count as a blow to certainty. Hence not knowing the falsity of the Cartesian
alternatives cannot, normally, undermine our knowledge claims.

But what has Dretske actually established? Has he, for example, established that the
Cartesian alternatives are irrelevant to the truth my knowledge claim? Has he, unlike
Stine and Goldman, given us an objective basis from which we may dismiss the sceptical
alternatives? Dismiss them as objectively irrelevant to ordinary claims to (objective)
knowledge? Has he gone at all beyond the reasons offered by Austin, Stine, Goldman, or
any of the contextualists (Lewis, DeRose, or Cohen, for example) in arguing that the

alternatives raised by the sceptic are objectively irrelevant to the question of knowledge? I

152

think not.”

Consider his Proof. Dretske’s “proof” thus far for the relational or relative nature of the
correct application of the concept of knowledge has consisted merely in the fact that we
often let certain things count as known or certain. He has argued that ordinarily we do
not count the Cartesian type of alternatives to what we claim to know as relevant to
whether we know. We do not — ordinarily — require that any such alternatives be
ruled out for the purported knower to count as being completely certain. Such facts as
these are, however, the only considerations Dretske offers for his conclusion that
knowledge is only relationally absolute.

By my lights, Dretske still owes us an argument. He owes us an argument for the
conclusion that such practices are not merely widespread, but correct. Merely pointing
out what we happen to let count as known or certain, and concluding thereﬁom that
knowledge does not require the Cartesian alternatives to be ruled out, is but one more
fallacious appeal to the principle of epistemic conventionality: the principle that any doubts
the introduction of which violates the conventions of appropriate assertion are irrelevant.
As we saw with the other relevant altemativists, arguing from the mere fact that in
ordinary conversations and contexts certain things count as known or as irrelevant to

knowledge, to the conclusion that those things manifest the objective nature of the

 

9”Actually, the absence of reasons in defense of his position that we need not rule out such
alternatives is consistent with Dretske’s defense of his extemalism, according to which one need not
be justiﬁed in order to have not only ‘the epistemic right to believe,’ but also, in order to know. See
for example, Dretske’s “Entitlement: Epistemic Rights Without Epistemic Duties.”

153

concept of knowledge itself is paradigmatically fallacious. It leaves out the crucial step
and hard work of establishing that what we let count as irrelevant is irrelevant. As I have
illustrated time and again, the mere fact that we happen to behave verbally in a certain
way, that we happen to have certain ordinary standards for appropriate assertion, does not
justify any inference to the nature of the objective standards for true assertion.

In this case, it is of course true (and the sceptic happily concedes) that we allow many a
knowledge claim to go unchallenged, to “count” as certain or known, even though the
sceptical alternatives to it have not been ruled out. When, for example, my brother claims
to [mow which stocks I should buy to double my money in the very near term, I do not
challenge his knowledge claim. He was, after all, trained as a broker, he has studied
market ﬁrndamentals, P/E ratios, and so on. Nonetheless, even though what he claims to
know may count as knowledge, it does not follow that he knows. It does not follow that
the objective conditions of knowing have been met. It does not follow that the knowledge
claim is true. And as I have already argued, though there may (or may not) be a relation
between meeting the standards of appropriate assertion (of making claims which we let
count as known) and the meeting the standards of true assertion, the two sets of
conditions are logically distinct. The one set having been ﬁrlﬁlled does not —
straightaway — justify the inference that the other set has been ﬁlled as well.

In fact, the very examples which Dretske uses to prove the relationally absolute nature
of absolute concepts themselves evidence clearly, I think, the distinction or difference

between having ﬁrlﬁlled the conditions of appropriate assertion and having fulﬁlled the

154

conditions of true assertion. Take, for example, the admittedly absolute term ﬂat. Recall
that for something to be ﬂat it must be not at all bumpy. When, therefore, the expert
golfer is searching for a truly ﬂat surface, though microscopic bumps in the surface may
not “count” as bumps relative to her interests, they do undermine the truth of her claim
that the surface is ﬂat. For to be ﬂat is just to be not at all bumpy: there is no allowance
for degrees. Ordinarily, of course, we take her to be speaking somewhat loosely. We
assume, for example, that he means that the surface is very close to being ﬂat or ﬂat

enough to play well on. No one, I think, believes her to mean that the surface really is

 

completely devoid of even minute bumps; and so Dretske’s claim, that the term ‘ﬂat’ is
relationally absolute because what counts as a bump varies with context, is false. A bump
is a bump, no matter how small. What changes with context is not whether a given thing
is a bump, but whether or not we care if the bumps we encounter are big or small. If we
are only concerned that there be no perceivable bumps we call a surface ﬂat, though only
loosely. To the extent that there must be no bump at all, no matter how small, we say of
even the seemingly ﬂattest of surfaces either that they are not ﬂat or that they are very
close to being ﬂat.

The mere fact that we can speak differently of different surfaces in different contexts, or
even speak differently of the same surface in different contexts, does not establish that our
terms are ﬂexible, but rather that we are. It establishes that we use our terms ﬂexibly, not
that the objective conditions of their true application are ﬂexible. And, the fact that our

ﬂexible application of our terms is ordinarily — in ordinary conversation —

155

unobjectionable (or, at least, not objected to), establishes only that the conditions of
appropriate assertion that we hold ourselves and others to are themselves somewhat
ﬂexible.

To the extent, therefore, that Dretske’s proof here of the irrelevance of the Cartesian
alternatives rests merely on what we let “count” as knowledge, it is inefﬁcacious in
establishing that knowledge itself is relationally absolute. So also, therefore, is it
ineﬁicacious in establishing that the sceptical alternatives to what we know are irrelevant
to the objective truth (or falsity) of what we claim to know.

Admittedly, however, Dretske does not rest his entire case for the relevant altemativist
view on the mere facts of what we happen to let count as knowledge. Instead, he goes on
to oﬁ‘er additional reasons for viewing knowledge as relationally absolute. The reasons he
offers are not, however, conclusive. Not even close. Rather, they are indirect and
contingent, contingent upon the presumed correctness of factors for which he does not
argue. He argues, for example, that his view, can accommodate the position of those ——
like the contextualists — who believe that knowledge is a ﬁrnction of pragmatic factors,
factors such as the changing interests of the conversational participants.” His view also —
unlike that of the contextualists — avoids “the problem of proliferating different meanings

of knowledge.”98 Additionally, he thinks, his view enables us to handle “otherwise

 

97Dretske, “Pragmatic Dimensions,” 368.
98Dretske, “Pragmatic Dimensions,” 368.

156

 

puzzling examples.”
None of these reasons, however, for viewing knowledge as thus relationally absolute,
are compelling. For, though it may be an advantage over other analyses of knowledge that

he can accommodate pragmatic views of knowing, why should we want a theory that can

 

accommodate such a view? Clearly the ability to accommodate a view like the
contextualist’s is not a material advantage unless the pragmatic view is objectively correct.

Without an argument for their objective correctness, there is no reason to accommodate

 

them, and hence no (proven) advantage in so doing. Moreover, the only arguments I have

-r\.-a.3m. u]- .-' I ' .

found in defense of such pragmatic views of knowing have all been based on the erroneous
principle of epistemic conventionality.100 The fact, therefore, that his model of knowledge
can accommodate such views is not a reason to prefer it to any other.

What about Dretske’s claim that his view of knowledge avoids the proliferation of
different meanings of knowledge? To some extent this is true. For, unlike the
contextualists, for example, who claim that the necessary or truth conditions of knowing
change with context, Dretske’s view is that the necessary conditions remain constant ——
we must , for example, always rule out all relevant alternatives. Yet, like the
contextualists, this is a reason to prefer his view, I think, only if there are independently

good reasons for thinking his relevant altemativism is correct. Yet the reasons he offers

 

99Dretske, “Pragmatic Dimensions,” 368.

l°°In particular, this is so of the contextualist views, and the RA views which have a contextualist
component.

157

in “The Pragmatic Dimension of Knowledge” are fundamentally the same as the reasons
offered by the contextualists for preferring their own. Each rests their case ultimately on
the fallacious and direct inference from our linguistic habits and the standards of
appropriate assertion manifested therein to the objective conditions of knowing. I have
already argued, however, that this inference fails. Hence, the mere fact that his view
avoids proliferating multiple senses of ‘knowledge’ is not reason enough to prefer it.

Finally, what about Dretske’s claim that his view enables us to handle otherwise
puzzling examples? Dretske defends this claim by reference to the following: Suppose an
amateur bird watcher is searching for an exclusively North American bird —— the Godwall
— in a North American pond. He discovers a bird with all of the markings of a Godwall
ﬂoating restfully on the pond. In this case Dretske argues we would say that the amateur
bird-watcher knows that the bird he sees is a Godwall.

Suppose however that an expert ornithologist is visiting in the area. Suppose,
moreover, that he has good reasons to believe that the Siberian Grebe —— not ordinarily
found in North America — has been migrating to North America. Suppose, moreover,
that the Siberian Grebe is indistinguishable from the American Godwall, except for
markings on the stomach which, however, would not have been visible to the amateur
bird-watcher, since the underside of the bird was hidden beneath the pond waters. Does
the amateur bird-watcher know that what he identiﬁed. as an American Godwall is not a
Siberian Grebe? He does not, Dretske says, and therefore cannot know that it is an

American Godwall. For its being a Siberian Grebe is — in this context — a relevant

158

alternative, and hence must be ruled out if the amateur bird watcher is to correctly claim to
know.

Consider one ﬁnal possibility: suppose that the expert ornithologist does not have good
reasons for suspecting that the Siberian Grebes have been migrating to North America,
and hence, does not have good reasons for thinking that the bird identiﬁed by the amateur
bird-watcher may be a Siberian Grebe. Does the amateur bird-watcher still need to rule
out that the bird he sees is a Siberian Grebe, in order to know that it is an American
Godwall? It is, after all, possible that the bird he sees is a Siberian Grebe. Dretske argues
that the amateur bird-watcher need not rule this possibility out because, in this instance,
such an alternative is not relevant. Hence the amateur bird-watcher knows he sees a
Godwall, even though it is possible that the bird he sees is a Siberian Grebe.

Dretske’s explanation here seems to rest on Stine’s criterion for the relevance or
irrelevant of alternative possibilities. After all, one factor which changes from case to case
is simply the presence or absence of good reasons for thinking the alternative might
obtain. Clearly Dretske does not base his assessment of relevance on the mere possibility
(as Stine had suggested he does) that the purported Godwall might be a Siberian Grebe.
Yet, although Stine claimed there to be only two alternative relevant altemativist views,
Dretske here in fact rejects both of these in favour of a third. For it is not, he argues the
mere presence or absence of good reasons which determines the relevance of the
alternative. Rather, he argues, such “good reasons” are actually irrelevant to the question

of relevance. As he states:

159

 

h

 

The Wisconsin lakes could be overloaded with migrant Siberian Grebes

without the bird watcher having any reason to think that such look-alike

birds actually existed. His lack of any reason to doubt, his ignorance of the

possibility that what he sees is a Grebe and not a Godwall, is irrelevant.101
Hence, Dretske concludes that having a reason to doubt is not -- as Stine had claimed --
a necessary condition of an altemative’s being relevant, though it may be sufficient. He
argues that the real criterion for distinguishing relevant from irrelevant alternatives resides
not “in what we happen to regard as a real possibility (whether reasonably or not), but in
the kind of possibilities that actually exist in the objective situation.” "’2

Dretske’s relevant aternativist explanation of the above “otherwise puzzling” example is

meant to serve as ﬁrrther support for his relevant altemativist view. But again, the fact
that his view can explain the example is reason to prefer Dretske’s view to that of others
only if either his is the only view capable of explaining the example, or, of the alternatives
available, his is the best. There are, however, other views at least equally capable of
explaining the example.103 Moreover, the fact that he can explain at least one (kind) of
puzzling example is not reason enough to accept his view in the absence of other reasons.
Yet, as we have seen, his other reasons (e.g., ordinary linguistic practices) are really non-

productive in the absence of further, and better, arguments.

Finally, I should note that Dretske’s explanation of the bird-watcher example amounts,

 

‘°'Dretske, “Pragmatic Dimensions,” 376.
l”Dretske, “Pragmatic Dimensions,” 337.
l”I will defend this claim at length in chapter V.

160

 

in the end, to either a quite circular argument or a virtual capitulation to scepticism. Take
the circularity ﬁrst. The issue in question is what we need or need not know in order to
know what we claim to know. Dretske has argued that what determines relevance is the
objective features of the environment —— whether there are or are not migrant Grebes in
North America, for example. Yet, if the determinant of what it takes to know is the
objective nature of the external environment, then either we know what the world is like
and, therefore, which alternatives are relevant, or we know nothing at all. But to claim

that we know what the external environment is like is to beg the very question at issue:

 

whether we can know anything at all. While, to concede that we do not know what the E
environment is like is to concede that we can never know which alternatives we must rule
out in order to know, since what determines relevance just is what is objectively possible
in that environment. Thus, to concede that we do not know what the objective world is
like is to concede we know nothing at all, or at least that if we do know anything, we
cannot know that we know it.

In light of this latter objection, one might expect Dretske to fall back upon what we
ordinarily suppose the objective environment to be like in determining which alternatives
are relevant; since, after all, he defended his analysis of knowledge by appealing to what
we let count as relevant. But, to do so, would — by his own argument — be arbitrary.
For, he has already stated that what we happen to think, even what we happen to have
reason to believe, is irrelevant to the question of what might objectively be the case, and

therefore irrelevant to which alternatives might be relevant to our claims to know. Of

161

course, such a move would also suffer the further -- and familiar -- difﬁculty of appealing
to the principle of epistemic conventionality, a principle already proven fallacious and

incapable of establishing the objective conditions of knowledge.

Section v: Relevant Altemativism and Contextualism: relevant differences
In my discussion of relevant altemativism I argued that the relevant altemativists err —

in a number of critical instances — in precisely the same way as their contextualist
cousins. I would like now, therefore, to ask to what extent these positions not only are
fundamentally the same, but also to what extent they differ. At the beginning of this
chapter I quoted the noted contextualist Keith DeRose who views the relevant altemativist
position as but a prominent form of contextualism. I suggested there that though the two
views are quite importantly different, and, therefore, that the relevant altemativists should
really be viewed as a distinct group of anti-sceptical epistemologists, they also share
certain fundamental similarities, though not of the kind DeRose contends. It is time now
for the details.

In DeRose’s paper, “Contextualism and Knowledge Attributions,”104 DeRose argues that

there are essentially two types of factors which may affect the relevance of an alternative.

 

'“DeRose, “Contextualism,” 918.

162

 

 

‘F-n-nﬂ’hhh': 1s 'A‘ I
t

DeRose calls these subject and attributor factors.” ’06 DeRose argues that insofar as a
relevant altemativist allows that the attributor factors affect the relevance of an
alternative, he is a Contextualist. He states that nearly all relevant altemativists allow such
attributor factors to affect relevance, though he gives no textual evidence of this, and goes
on to conclude that most relevant altemativists are, therefore, Contextualists.107
In what follows, I shall argue that DeRose’s categories of the different factors which 1. I

may affect relevance are construed somewhat too narrowly; too narrowly to capture either

what is essential to the contextualist position or to the relevant altemativist position. In

 

If” "" '

fact, once the narrowness of his categories is augmented, I ﬁnd that what distinguishes a
relevant altemativist from a contextualist is not the extent to which he allows attributor
factors to affect relevance (though this may be part of the story), but rather, the extent to
which he allows conversational factors to do so. I then argue that because none of the
relevant altemativists allow conversational factors alone to affect relevance —— as do the
contextualists — they are not properly viewed as contextualists. For it is the allowance

of mere conversational factors which deﬁnes the contextualist’s identity, and it precisely

 

such factors” which none of the relevant altemativists allow to affect the relevance of an

alternative to a claim to know.

 

lo5DeRose, “Contextualism,” 918.

l”DeRose develops this distinction also in his “Relevant Alternatives and the content of knowledge
Attributions,” Philosophy and Phenomenological Research Vol. LVI (1996).

l”DeRose, “Contextualism,” 918.
10"Though ,of course, there are others; e.g., mere logical possibility.

163

Finally, I argue that though the RA’s and contextualists are really, therefore,
fundamentally different positions —- in precisely that respect in which DeRose thinks they
are the same, the manner in which they argue for the irrelevance of the Cartesian
alternatives to our knowledge claims is ﬁrndarnentally the same. Hence, the relevant
altemativists and contextualists are alike, not — as DeRose contends —— in positive
theory, but rather, in the practical failure of their respective attempts to support their
positive theoretical commitments.

DeRose deﬁnes ‘subject’ and ‘attributor’ factors as follows: Subject factors are factors
that have to do with the subject of the purported knowledge; that is, the knower.“
Attributor factors, on the other hand, are factors which have to do with the speaker or one
who makes the knowledge attribution. “° (Such factors might, of course, overlap if the
purported knower and the attributor are one and the same, as when I say of myself that I
know that P.) Examples of subject factors given by DeRose include there being a shortage
of zebras, leading zoos to use painted mules, and there being a change in the hours of local
banks. Examples of attributor factors, on the other hand, have always to do with the
“speaker’s linguistic and psychological context.”lll From contextualist literature,
including DeRose’s own, linguistic and psychological factors include what alternatives to

the knowledge claim are raised in a conversation, the importance of being right about

 

“”DeRose, “Contextualism,” 918.
”DeRose, “Contextualism,” 919.

”‘DeRose, “Contextualism,” 919.

164

 

 

l
E

what one claims, and whether or not the alternative possibility is considered by the parties
to the conversation.112

DeRose claims that it is the extent to which the relevant altemativist allows the latter ——
the linguistic and psychological factors — to be a determinant of relevance that the
relevant altemativist is a contextualist. And, as noted above, he asserts that most relevant
altemativists are contextualists. Yet, in my reading of the more prominent relevant
altemativists I ﬁnd none who allows things such as the mere mentioning of a possibility or
the importance of being right to affect the relevance of an alternative. In Stine, for
example, a necessary condition of a factor’s being relevant is that we have some reason to
think it is true. Hence the mere mentioning of a possibility does not, for Stine, make an
alternative relevant. For Goldman, a necessary condition of a factor’s being relevant is
that it be a perceptual equivalent of that which is claimed to be known. Again, therefore,
whether it is important to be right or whether a possibility is entertained is not sufﬁcient to
make an alternative relevant, and certainly is not necessary. In Dretske (as well as in
Stine), one ﬁnds the explicit rejection of the view that mere mentioned possibilities can
affect or determine relevance. Dretske argued, for example, that if the expert
ornithologist has no basis for his hypothesis that there may be migrating Siberian Grebes in
the area, then that possibility is not relevant, whether mentioned or not (unless of course
the objective environment is such that there are Siberian Grebes in the area. But this has

nothing to do with whether the possibility is mentioned, considered, important, or

 

“’See chapter 111 of this work, and also DeRose, “Contextualism,” 915.

165

 

~4‘ 'Lm

 

l:-

discussed). In fact, then, contrary to DeRose’s claim, not one of the prominent relevant
altemativists does consider ‘attributor’ factors alone to be a criterion -- necessary or
suﬂicient -- of relevance.113

Earlier I noted that I believe DeRose’s construal of the factors which make one a
contextualist is too narrow. This is evident I think because though the attributor proper is,
of course, a part of the linguistic context, she is a part only, unless, of course, she is
conversing with herself alone. Usually, as the examples of the contextualists normally
evidence (including DeRose’s own), the factors a contextualist considers to be relevance
determining are those factors which contribute to the linguistic context as a whole. These
factors include any statements made by the attributor, as well as any group assumptions,
individual queries, interests, and so forth. They are not, then, limited to the contributions
of the attributor alone. Thus, I would suggest that the relevance determinant category
which best identiﬁes the Contextualist position ought to be named “conversational
factors,” not attributor factors. Thus revised, I would agree that to the extent that a
relevant altemativist allows such conversational factors to determine relevance, she a
contextualist. Yet, as suggested above, each of the relevant altemativists rejects such
factors as a determinant of relevance, and does so unequivocally. Why, therefore, DeRose

asserts that most relevant altemativists are contextualists, or that the relevant altemativist

 

”Indeed, it is precisely this diﬂ‘erence which leads to another difference in the respective views of
the RAs and the CTs, namely the fact that the RAs do not (ordinarily) allow any room for the
correctness of scepticism or the compatibility of scepticism with common sense, whereas this is
another hallmark of the contextualist view. (Stine is something of an exception here, though the
extent to which she is, she is a contextualist, not an RA.)

166

 

39"!"

view is a prominent Contextualist one escapes me.

On the other hand, DeRose’s categorization of the factors the non-contextualist-
relevant- altemativist allows to affect relevance as ‘subject factors’ also seems
inappropriate. For examples of such factors are invariably not something uniquely tied to

the subject, but rather to the objective state of his environment” Such factors as whether

 

there are barn facades nearby, whether there is a shortage of zebras, whether there are any E
migrant Siberian Grebes are the kinds of factors the relevant altemativist allows to

determine relevance. Thus I suggest we change DeRose’s ‘subject factors’ to

‘environmental factors.’ This label, I think, captures better the range of factors allowed L

and disallowed by the relevant altemativist to be determinants of relevance.

We have then two categories, the ‘conversational’ and the ‘environmental’ which, I
think, enable us to distinguish a relevant altemativist from a contextualist, much in the way
that an expert ornithologist might distinguish a Godwall from a Siberian Grebe: if the
philosopher in question claims that conversational factors can affect which alternatives to
a knowledge claim must be ruled out in order for S to know that P, this is sufficient reason
for identifying her as a contextualist. (It is also, as I argued above, a necessary condition.
For a philosopher cannot be a contextualist unless she does consider such factors to be
relevant.) If, on the other hand, she allows that only environmental factors can affect

which factors must be ruled out in order for S to know that P, she is a non-contextualist

 

“"Stine may be something of an exception here, but insofar as she is, she is again closer to being a
CT than an RA.

167

 

 

relevant altemativist. If she allows both, she is not a relevant altemativist, but a
contextualist.

Interestingly, DeRose in this (his 1992) paper argues also that it is contextual changes in
the attributor factors (what I am calling conversational factors) which change the meaning
of the term ‘knowledge.’ Equally interesting is his claim that mere subject factors alone
are not capable of affecting such a change in meaning. This interests me because , though
I do not agree that either factor can change the meaning of ‘knowledge,’ I do believe that
another factor which distinguishes the relevant altemativist view from that of the
contextualist is the RA’s steadfast adherence to the view that there are necessary
conditions of knowing anything at all, which as such do not change in accordance with
changes in context. This is supportive of, or at least consistent with, my view that strictly
speaking the relevant altemativists do not allow the mere conversational factors to affect
the relevance of alternatives, or the meaning of ‘knowledge.’115 It is also supportive of, or
again at least consistent with, my view that the contextualists do allow for such changes in
meaning. Indeed as I asserted at the outset, this is one area where the RA’s position is
clearly preferable to the contextualist’s. For, unlike the contextualists (and as noted by
Dretske) the relevant altemativists do not allow for a proliferation of meanings with
respect to the term ‘knowledge,’ and hence do not undermine the semantics of

‘knowledge’ to the extent to which the contextualists do.

 

“5 With respect to the question of meaning, Stine again seems to be something of a noteworthy
exception here, but to the extent that she is, I would argue she is a contextualist, not an RA. This
issue must be left for another time, however.

168

With respect to positive theoretical commitments, then, the relevant altemativists and
contextualists are — contrary to DeRose’s assessment — quite distinct. Though each
tries to refute the sceptic by arguing that the Cartesian alternatives appealed to by the
sceptic are irrelevant to knowledge, the theoretical basis on which they do so differs: For
the Contextualists, the sceptical alternatives are irrelevant so long as they are not part of
the conversational context; which, since ordinarily they are not, saves our ordinary
knowledge claims. For the relevant altemativist, the sceptical alternatives are not relevant
— whether part of the conversational context or not — unless there are objective
environmental features which correspond to them; which since (it is presumed) ordinarily
there are not, also saves our ordinary knowledge.“6

Yet there are fundamental similarities too. These similarities have primarily to do with
the extent to which representatives of both groups attempt to establish their respective
positive theoretical commitments in the manner of Austin.117 Each, I have argued, argues
for her position by appealing either implicitly or explicitly to our ordinary linguistic
practices. But, as I have also argued, this equally dooms each of their respective proofs
to failure. For the inference from mere facts of linguistic behaviour to the objective

conditions of knowing is fallacious. It involves an appeal to the fallacious principle of

 

ll“This may be something of an oversimpliﬁcation with respect to Stine, since she requires only that
there be a reason to think an alternative is true if it is to be relevant. But I do not think this
oversimpliﬁcation affects my main point which is simply that the RA’s require more than mere
mentioning and conversational attention for an alternative to be relevant.

”’As detailed in chapter 11 of this work.

169

 

‘31:]: ﬁ'ﬁ still

I "m “at“ EI‘JH.

t 1
‘V

epistemic conventionality, a principle which supports such an inference only to the extent
that it conﬂates two logically distinct sets of standards, the conditions of appropriate and
the conditions of true assertion. Therefore, though DeRose is correct in attributing a
fundamental similarity to the relevant altemativists and the contextualists, the source of
that similarity lies not in positive theoretical commitments, but rather in the manner in
which each tries to establish what are the objective conditions of knowing, and thereby to

refute the sceptic.

170

 

Jr‘s.

 

 

we

 

CHAPTER V:
The Epistemic Relevance of Common Sense
Section i: Prefatorial remarks:

What follows is a matter not of the way in which things most assuredly are, but
rather, of how things seem to me now, after reﬂecting on the positions of the ordinary
language philosopher, the contextualist, and the relevant altemativist. I do not intend
what is said be taken in the spirit of the dogrnatist (nor in the tradition which follows
therefrom), a tradition which asserts positively that things are thus and so, without

appreciation for the potential equipollence of the arguments.‘ The following, rather, is

intended in the spirit of the Pyrrhonian, who claims only that things appear thus and so.

Regarding the phrase "I determine nothing" this is what we say. We
hold that to determine is not simply to state a thing but to put forward
something non-evident combined with assent. For in this sense it will be
found that the sceptic determines nothing, not even the very proposition
"1 determine nothing, " what he means is "I am now in such a state of
mind as neither to afﬁrm dogmatically nor deny any of the matters now
in question." And this he says undogmatically by way of announcing
undogmatically what appears to himself regarding the matters presented,
not making any conﬁdent declaration, but just explaining his own state
of mind.2

In this chapter, I examine the extent to which the contextualists, the ordinary

 

'I allude to the Pyrrhonian observation that for every argument there is a counter argument, and for
every belief that P there are reasons and evidence also for not-P. See my discussion of this in chapter
I.

" Outlines, 1 17.

171

U E

 

w—wv.

 

 

‘7'“?

 

language philosopher, and the relevant altemativist met their goals. Did they, for
example, explain the force of scepticism? Did they explain why common sense is
compatible with the truth of scepticism? Did they save common sense? And,
moreover, to the extent that they did achieve these goals, did they do so without those
dogmatic assumptions the sceptic argues undo even the cleverest of arguments? On the
assumption that arguments and reasons still matter, I argue, they failed establish their
anti-sceptical conclusions. For, not only do each of the theories we have looked at
suffer ﬁ'om the difﬁculties already discussed, they also suffer from the making of many
dogmatic assumptions, assumptions which by their very nature as dogmatic are both
inherently objectionable and fallible. In fact, ﬁrrther exploration into the foundations of
this genre of anti-scepticism (the genre embraced, that is, by the ordinary language
philosopher, the contextualist, and the relevant altemativist) reveals that their
multiplication of contextualized theories of knowing (together with their numerous ad
hoc assumptions, and relativistic exceptions") rests essentially on an insidious (though
not uncommon) misunderstanding of both the nature of scepticism and the nature and
contents of common sense. This misunderstanding rests on a number of not only
uncritical and dogmatic assumptions, but assumptions that are erroneous. The
erroneous nature of these dogmatic assumptions is clear once they are subjected to
critical reﬂection.

One aspect of the ordinary language philosopher, the contextualist, and the relevant

altemativist’s misunderstanding of the nature and contents of common sense is the

 

3Exceptions and assumptions I have described in the foregoing chapters.

172

 

 

 

if"

 

unstated, but crucial, assumption that common sense consists not just in ﬁrst-order
beliefs about the world, but also in a set of second-order beliefs about those beliefs.
(By ﬁrst-order, I mean beliefs about the world; by second-order I mean a kind of meta-
belief about those beliefs.) This false and uncritical assumption leads to the ﬁrrther —

and equally erroneous — conviction that common sense is threatened by scepticism. It

 

is this latter conclusion that leads to the further (and also erroneous) conclusion that I” '
scepticism must be reﬁtted in order to save common sense.

What follows upon these assumptions, therefore, is the proliferation of attempts to

 

establish that the sceptic's doubts are irrelevant. Moreover, having made the

1r“ «my .' !.

assumption that the ordinary is revelatory of the true, each of these attempts rests on
proving that by the ordinary standards of knowledge revealed in ordinary behavior the
sceptic's standards are false in almost all (that is, excepting the extraordinary) contexts.
For it is context which determines the standards one must meet in order to correctly
claim knowledge. Hence both the sceptic's failure, as well as her mitigated success, are
explained in terms of a contextualized semantics of knowing.

In this chapter I argue that because this anti-sceptical tradition rests on such uncritical
assumptions about the nature and contents of common sense, rejecting these
assumptions (as well, therefore, as the inferences made therefrom) leads to the end, not
only of the proliferation of this particular kind of anti-sceptical refutation, but also, to
the very need or foundation upon which many anti-sceptical theories are built; the
perceived need to defend common sense. Moreover, I argue, though common sense

and scepticism are indeed compatible (as the contextualist maintains), one can establish

173

 

this compatibility without making any of these mistaken assumptions, and furthermore,
one can do so without relativizing, contextualizing, or otherwise arguing that the
standards which govern knowledge are lower than the sceptic has claimed.

The assumption that common sense includes second-order knowledge beliefs is the
assumption ultimately responsible for the impression that scepticism threatens the

contents of common sense. The further assumption that the ordinary reveals the

 

objective standards of knowledge (an assumption evidenced by the direct inference i

 

ﬁom ordinary linguistic behaviour to the standards of knowing) is the belief ultimately

responsible for reﬁrting the sceptic by proving her standards are not ordinary. Once,

however, one rejects the assumption that common sense includes a corresponding set
of meta-beliefs to the effect that each of the beliefs we hold can also be known,
refutations, proofs, and the like, each prove unnecessary. For what immediately
follows is the conclusion that scepticism cannot threaten common sense, even if

scepticism is objectively correct in its most radical form.

 

In fact, I think, scepticism has not historically been directed at common sense.
Rather, scepticism's object is second-order beliefs about those common sense (and
other) beliefs.

Adhering then to appearances, we live in accordance with the normal
rules of life, undogmatically, seeing that we cannot remain wholly
inactive, and it would seem that this regulation of life is fourfold, and
that one part of it lies in the guidance of nature, another in the constraint
of the passions, another in the tradition of laws and customs, another in
the instruction of the arts. . . . But we make all of these statements
undogmatically.4

 

‘Outlines, 17.

174

The sceptic's argument is neither directed at nor threatening to common sense.
Moreover, once the ﬁrrther assumption that the ordinary reveals the objective is
subjected to critical reﬂection, I ﬁnd that merely noting the sceptic's concerns are
extraordinary cannot reﬁrte her.

The argument of this chapter requires, therefore, a careful re-examination of our
(mostly unexamined) conception of the relation between scepticism and common sense.
It requires careful reconsideration of our assumptions about both what is and what is
not implied by our ordinary behavior. It requires an attention to the potential (and, I
argue, actual) distinctness in kind of the standards which govern knowing behavior and
those which govern knowledge. Finally, it requires that each of these reconsiderations
be carried through, as much as possible, without the dogma of traditional assumption.’
Truly re—examining both what common sense is and how it and scepticism stand related
necessitates the removal of dogmatic (especially philosophical dogmatic) assumption.

The result of both becoming aware of and removing these dogmatic assumptions is
an ordinary conception of the contents of common sense. According to this resulting
conception of common sense, second-order beliefs about knowledge are not part of the

contents of common sense. That is, though we may indeed believe that P, it is not the

 

’Of course, as all of the great epistemologists have pointed out for the last several rrrillennia, the
instinct to believe and to accept belief without evidence is everywhere. So ubiquitous, in fact, that it
is impossible to be certain one is not under its inﬂuence without continuous and systematic attention.
Even then, of course, we cannot transcend certain fundamental principles of thought. Nonetheless,
within such limitations, one can still assess whether or not a position rests only on what are arguably
dogmatic assumptions and nothing else. That is the task of this last chapter.

175

 

 

case that for each of our beliefs that P — even of our most intractable beliefs that P —
we also ordinarily believe that we know that P. Rather, according to the resulting
conception of common sense, though we believe many things, we do not ordinarily
also believe — in addition to each of those beliefs — that we know them. That, of
course, is not to say that it is part of common sense to believe that we do not know
them. Rather, the point is that common sense is concerned with existence, not
semantics, not epistemology. Its business is guiding action, not the conscious
fulﬁllment of the objective conditions of knowledge or the analysis of knowledge.
Neither has the least bit to do with common sense. And, insofar as common sense
makes no such pretensions to ﬁrlﬁll the objective conditions of knowing, it cannot be
undermined by arguing its pretensions to know are mistaken.

Supposing that common sense and ordinary belief are not governed by semantical or
epistemological standards, it follows then that common sense is not, and cannot be,
threatened by scepticism. Moreover, because the content of common sense does not
include second-order beliefs about knowing, to the extent that our ordinary behavior is
governed by common sense, that behavior is neither revelatory of the standards of
knowing, nor is it governed by that kind of standard. To the extent that it is not
governed by epistemic standards, common sense cannot be threatened by scepticism,
even when scepticism’s objective truth is most radically understood. For common
sense consists not of second-order beliefs about knowing, but of ﬁrst-order beliefs
about the world.

I, therefore, urge the reader to take a suspensive, dubiative, and non-dogmatic

176

 

 

approach to the traditional assumptions herein discussed. I urge, that is, a kind of
Pyrrhonian approach, an approach committed to rejecting assumptions for which there
is not unequivocal evidence or reason. In particular, I urge the reader to reconsider the
evidence for the position that common sense includes not only ﬁrst but second-order
beliefs. This approach reveals an important set of consequences for our understanding
of the relation between epistemology, scepticism, and common sense. This set of
consequences is uncovered as one leaves behind one’s philosophical sophistication
(until it becomes once again necessary, that is, and appropriate).

When one begins with a proclivity and commitment not to what is schooled and
philosophical, but to what is ordinary and common, our non-dogmatic study of
common sense leads to conclusions so contrary to those thought evident by the great
anti-sceptics of this century (indeed, of others also) that one marvels at the divergence
in conclusion, given the seeming equivalence in beginning; the ordinary, that is. The
rub, of course, is that the beginnings are not really equivalent, nor is the kind of
reasoning which follows therefrom. Their beginnings are ﬁ'aught with unargued for and
unnecessary assumptions, leaving the inferences drawn therefrom in a quite vulnerable,
because ill-founded, position. It is these latter — their assumptions, and ill-founded
inferences — which account for how the end can differ so greatly from beginnings
which seem synonymous.6

One’s instinct to judge without evidence is everywhere, and must -- along the way —

 

‘Both I and they appeal to the ordinary, though I begin also with a commitment to resisting the
dogmatic, while they begin with a proclivity to engage implicitly in it.

177

- i. .nnu!‘

 

war

 

be intentionally resisted. To resist such instinct by speaking the language of the
Pyrrhonian, however, can also not be done, at least without a great deal of unnatural
attention to language. Since the language of the Pyrrhonian - relative to our current
linguistic habits - is at once both unnatural and convoluted, it is distracting, and would
serve, I think, only to undermine what we are after, namely, a clean sense of what is
ordinary; a pre—philosophical understanding of common sense.

If then while the dogmatizer posits the matter of his dogma as

substantial truth, the sceptic enunciates his formulae so that they are

virtually cancelled by themselves, he should not be said to dogrnatize in

his enunciation of them. And, most important of all, in his enunciation

of these formulae he states what appears to himself and announces his

own impression in an undogrnatic way, without making any positive

assertion regarding the external realities.7
My central aim in this chapter is to challenge the dogmatic assumptions of the clearest
representatives of the kind of anti-sceptical approach I have considered throughout.
My aim is not to offer an alternative theory of knowledge. Rather, it is to look
critically at the theories I have discussed, and delineate the consequences of rejecting
certain principles which they uncritically assume. I attempt to use only those principles
which must be assumed (e.g., -(A & -A), (A v -A), etc)),by all parties, lest I be dealt
the fate of Craytalus, unable to explore human discourse and reason, capable only of
wiggling my little ﬁnger. To the extent, therefore, that I appear to make positive

assertions that things are thus and so, let it be understood that I mean only that it

appears to be the case that things are thus and so.

 

70utlines, 11.

178

 

 

 

Again, the direct result of removing uncritical assumption is the discovery that the
ordinary language philosopher, the relevant altemativist, and the contextualist's
conception of common sense is erroneous. The assumption that knowledge beliefs can
be derived directly from knowledge behavior is false, and, therefore, to defend the latter

against the sceptic represents a ﬁrndamental error; an error regarding the kind of beliefs

 

and the kind of standards at stake. When the sceptic claims that we cannot know that

" IJIJ I”

P, her assertion is not governed by the same kind of standard as that which governs our
saying that we know that P. The argument from ordinary language (embraced by not

only the ordinary language philosopher, but also, the contextualist, and the relevant

 

iff

altemativist) conﬁrses the distinctness in kind of the standards governing what we say
and the objective conditions which govern knowledge. The two are separable in kind:
the one a question of pragmatics, manners, and habit; the other a question of truth.
The sceptic’s criteria of knowing pose, therefore, no threat to ordinary common
sense. In that sense, scepticism is, as the contextualists maintain, ﬁrlly compatible with
common sense. It must be emphasized, however, that scepticism’s consequent

compatibility with common sense and ordinary knowing is not (as the OLs, CTs, and

 

RAs have supposed) owed to the sceptic’s raising or manipulating ‘ordinary’
standards (she does not). It is, ﬁrrthermore, not owed to the criteriological signiﬁcance
of her doubts lying outside (as they have oft been accused“) of the conditions of
knowing (there is no reason to suppose they do). Equally, this compatibility between

common sense and scepticism is not owed to the sometime or special relevance of the

 

8An argmnent variously made by each of the OLs, CT 5, and RAs.

179

alternatives to P.

The irrelevance of scepticism and Cartesian doubt to the contents of common
sense is owed, therefore, not to any error in the sceptic's analysis, but rather, to the
separability in kind of the standards which govern the objective conditions of knowing
from those which govern knowing-behavior, the ﬁrst being epistemic and semantic, and
the second not. It is owed to the fact that common sense consists not in the abstract
and unordinary beliefs about those beliefs which concern philosophers, but rather, in
ordinary ﬁrst-order beliefs about the world.

In what follows, therefore, I offer a compatibilist defense of common sense which
does not rest on ad hoc assumptions, contextualized standards of knowing, or
relativized alternatives to P. In fact, my defense of common sense rests not on
defeating the sceptic, nor on a particular theory of knowledge, but rather, on the
recognition that the subject matter of ordinary common sense is distinct in kind from

the subject matter of philosophical scepticism.

Section ii: The Dogmatic Foundations of Epistemic Conventionality

In chapters 11, HI, and IV, I argued that there are critical threads which bind the
pages of these three major anti-sceptical epistemologies: the ordinary language, the
relevant altemativist, and the contextualist. Salient among these is their common
perception that in order to save common sense ﬁom ruin, the alternatives the sceptic
claims we must rule out to know must be proven irrelevant to the accuracy of our

knowledge claims. These anti-sceptics believed that unless the sceptic’s alternatives

180

 

 

could be proven irrelevant to knowledge, our common sense beliefs would be suspect
of systematic falsehood. Why would the sceptic implicate us in systematic falsehood?
Because ordinarily we claim to know many things. Proving that we need not rule out
the sceptical alternatives to our knowledge claims, they thought, was the only way to
make safe our ordinary knowledge claims ﬁom the peril of scepticism.

In what follows I too argue that the Cartesian alternatives to P are irrelevant to our
ordinary knowledge claims. I too argue that these alternatives need not be ruled out to
save ordinary belief. Nor need they be ruled out to vindicate our ordinary claims to
know. However, this irrelevance derives not from any error in the sceptic's reasoning.
It derives not from the irrelevance of the sceptic’s alternatives to actual conditions of
knowledge. (In fact, I largely leave the question of the relevance the sceptic’s
alternatives to knowledge undetermined.) For the irrelevance of those alternatives
derives ﬁom the fact that the object of scepticism, what the sceptic argues cannot be
justiﬁed, is not the kind of belief evidenced by common sense (a belief the sceptic
acknowledges is both necessary and natural), but instead, the kind of belief that is
philosophical. In my argument for the compatibility of scepticism and common sense I
ﬁnd fault, then, not with scepticism, but with the assumption that scepticism’s truth
threatens the ordinary. An assumption which creates the need for such contextualized
semantics.

Again, the method the OLs, CTs, and RAs chose to reﬁrte the sceptic and prove her
irrelevance to knowledge was to appeal to the facts of ordinary language , what we

might call 'ordinary speak.’ This latter -- ordinary speak -- proved ,they thought, the

181

 

irrelevance of sceptical doubt. They argued that from the conditions which (appear to)
govern ordinary language we can infer — without further argument - the criteria which
govern the truth conditions of knowledge. Moreover, they believed that the conditions
which appear to govern ordinary speak could be discerned from our ordinary behavior.

Of course, were it the case that our ordinary behaviour had this kind of signiﬁcance, the

weaker standards evidenced there would insulate our knowledge claims ﬁom the F

sceptic's more demanding criteria. Yet, as I have already shown, the inference from the 7

ordinary to the normative is fallacious. .1
What I have not, as yet, fully explained is why philosophers with such ability would j: '

 

chose this means, rather than another, to save common sense. I have not explained
why this kind of approach to scepticism might seem correct. It is here, however, that
the uncritical assumptions I spoke of above come into play. Consider the role played
by the assumption that what we ordinarily do reveals the standards of what we ought to
do. It is true, of course, (and obviously so) that ordinarily we neither observe nor
require what the sceptic asserts is necessary for knowledge. We do not, for example,
require of S that she rule out that she may now be dreaming; we do not require that
she rule out that she may be vatting; nor do we require that she rule out that she may be
a victim of evil and cunning genius. Such conditions are neither ordinarily observed,
nor ordinarily are they expected to be ﬁrlﬁlled. Indeed, to the extent that we evaluate
the claim of S to know that P, one set of considerations we ordinarily never consider
are these.

But without ﬁrst assuming that the standards we follow in ordinary contexts reveal

182

 

 

the standards we ought to follow in claiming to know, what can we actually infer from
these practices? What follows ﬁom these facts about how we behave? Of what
philosophical signiﬁcance is the fact, for example, that ordinarily we do not rule out,
nor expect others to rule out, these Cartesian alternatives before we claim to know?
Does it follow that we need not — according to the objective conditions of knowing —
rule out these Cartesian alternatives to P? Or, might there be an alternative and better
explanation of what we ordinarily do?

The ordinary language philosopher, the contextualist, and the relevant altemativists
assume that the only explanation is that the condition which govern our ordinary
linguistic behaviour are the conditions of knowledge. They assume also, therefore, that
we can save the ordinary by establishing that the sceptic of changes, manipulates, or
otherwise distorts the true standards of knowledge. For, they also assume, the
standards governing ordinary behavior are the same kind of standard as those
governing knowledge. They conclude, therefore, that ﬁ'om our ordinary behavior the
irrelevance of sceptical doubt is clearly implied. They argue, if we do not ordinarily
rule out those alternatives to P, we need not rule them out either. What happens,
however, when we deduct ﬁ'om their arguments the uncritical assumptions on which
they rest?

As I argued earlier”, without such assumptions in place, the matter of what we
happen to do, is of little philosophical signiﬁcance with respect to scepticism. At least,

it does not have the kind of signiﬁcance these philosophers have taken it to have. For

 

9In each of the previous chapters.

183

 

without the assumption, for example, that the ordinary reveals the objective we cannot

infer ﬁ'om the fact that we do not ordinarily rule out Cartesian alternatives that we

ought not to. We can not infer that these alternatives are not objectively relevant to

what we claim to know. No such conclusion follows. Without these assumptions,

what happens to be true of how we speak (or how we respond to the spoken word of

others) does not imply the necessary conditions of the truth of what is so spoken. The a?“

relevance, of course, in the case of knowledge in particular, is that from mere facts it

detailing how we habitually or ordinarily use the term 'knowledge’ we cannot derive the j

conditions of its true application.10 5
These anti-sceptical philosophers, however, do not pause to consider whether the 1

standards we (seemingly) happen, ordinarily, to use in ordinary discourse and the

standards which concern the sceptic might differ not in strength, but in kind. They do

not take pause to consider the possibility that what we say in ordinary conversation

might be governed by a different kind of standard altogether than the kind which

governs our concept of knowledge and, therefore, the objective conditions of its

correct application. They do not consider whether it might differ in kind from the kind

of standard which concerns the sceptic. Each, therefore, is led to conclude that the

standards the sceptic appeals to are much higher than those necessary for knowledge.

They are lead also to conclude that it is only by her artiﬁcially high standards that we

can fail to know.

But where is the proof? Where is the proof that the sceptic’s standards are

 

l°1 established this in chapter 11.

184

 

 

artiﬁcially high? Where is the proof that our common sense behavior represents the
standards of knowing? What is the evidence that the standards which govern our
ordinary knowledge assertions play any epistemic ﬁrnction? Beyond their uniform
chorus of homage to what is ordinary, little of rational defense can be discerned,
excepting perhaps the oft rehearsed — though inadequate - 'explanation' that we must
avoid scepticism.

How important is this avoidance of scepticism? Important enough that Lewis allows
assumptions to “spring into existence” when needed, "lest we be made false."ll And
Cohen requires only that we rule out only those defeaters which are inter-subjectively
evident. This, of course, at the same time entitles us to ignore all those defeaters which
are “inter-subjectively opaque.”12 Knowledge is, after all, only relationally absolute. ‘3
Stine too states directly that avoiding scepticism is in-itself a reason to embrace a
contextualistic theory of knowing.” Lewis concedes he has no reason for contextualism
other than the desire to avoid capitulation to scepticism, deﬁning the exceptions
necessary to avoid scepticism as decidedly "ad hoc."15 Indeed, DeRose goes so far as

to argue that the mere avoidance of being implicated in systematic error is the central

 

”David Lewis, “Scorekeeping in a Language Game,” Journal of Philosophical Logic 8, (1979).
12Stewart Cohen, “Knowledge and Context,” Synthese 73, (1987) .

l3Cf Dretske’s “The Pragmatic Dimension of Knowledge,” Philosophical Studies, 40, ( 1981).

1"Gail Stine, “Skepticism, Relevant Alternatives and Deductive Closure,” Philosophical Studies, (1976).
l5David Lewis, “Elusive Knowledge,” Australian Journal of Philosophy 74 (1996).

185

 

 

reason for preferring contextualism.l6 And in a later paper, DeRose confesses that this
is the motivation behind not only his, but also, Stine, Lewis, and Cohen’s contextualism
as well.17 As DeRose states in his “How can we know we are not brains in vats?”, “ . . .
the motivation — or at least a main motivation — for thinking we know we are not BIVs
is to avoid scepticism . . . . ”’8 Even Goldman and Dretske lay ultimately their
conclusions before the tribunal of the ordinary (the latter, if not the former”) in the
hope of avoiding scepticism.

But these reasons are quite incomplete for claiming that the standards of ordinary
language refute the sceptic, or prove her doubts irrelevant to ordinary claims to know;
incomplete, that is, unless you are naturally inclined to make the additional correlative
assumptions. In light of both the role and the uncritical nature of these assumptions,

I believe that not only do these anti-scepticisms fail, but that ultimately, the very need
to reﬁrte the sceptic rests on an ill-gotten fear, a fear which rests on an erroneous
understanding.

The drive to reﬁrte the sceptic is premised upon an erroneous understanding of
common sense; an understanding of common sense bred by too much philosophy and a

lack of attention to what we really, ordinarily, believe. Properly understood common

 

l“I discuss this at some length in section iv of the present chapter.

l’Keith DeRose, “How do we know we are not Brains in Vats‘l”, The Southern Journal of Philosophy,
38, (2000).

”DeRose, “Brains in Vats,” 122.
19Since Goldman maintains neutrality on the question of scepticism. I discussed this in Chapter
IV of this work. See also Goldman’s “Discrimination and Perceptual Knowledge,” The Journal of
Philosophy, 73.20, (1976).

186

' "(le

 

 

Qaﬁnmmﬂ1n-‘ur.’ .4744
{I

sense is not threatened by the truth of scepticism, and the fear that scepticism might
land us in systematic error is unwarranted. So too, therefore, is the need for the OL,
CT, and RA ‘reﬁrtation’ of scepticism. Their fear that the sceptic poses a threat to
ordinary belief is illusory; common sense is not in danger. That the sceptic's doubts and
conclusions pose no threat to common sense does not, however, imply that the sceptic's
philosophical conclusions are false or ill-derived. (In fact, though I do not have the
space to argue it here, I believe they are neither.) Yet there is a marked assumption not
uncommon amongst philosophers, embraced also by each of the philosophers I have
addressed, that the vindication of common sense amounts to the refutation of
scepticism, or, that the reﬁrtation of the sceptic is an event sufﬁcient to vindicate
common sense.

Yet this tendency to embrace the vindication of common sense as a positive
refutation of the sceptic rests also on the singular, yet false, assumption that the
standards which govern ordinary common sense discourse differ only in strength from
the sceptic's standards of knowing. It is this singular assumption which leads not only
to conﬂating the vindication of common sense with the refutation of the sceptic, but
also, to a distorted conception of common sense now taken on as a virtual ﬁrst

principle, and perhaps most disappointingly, at the expense of objectivity.

Section iii: Common Sense: the separability of ﬁrst and second-order belief
To establish that the fear behind the anti-sceptical philosopher's epistemic theories is

ill- founded, we must then not only reacquaint ourselves with our (ordinary) concept of

187

 

m a. A L; ~ 35a

:1—

31,. 7‘

 

 

 

knowledge, but also, with our (ordinary) conception of common sense. For so deeply
entrenched is the philosopher's (and in particular, the anti-sceptical philos0pher's)
understanding of common sense, that to return us to our native beliefs, we must undo
the confusion she has laden that conception with. We must retrieve that sense of what
we actually, ordinarily, believe from the hands of the philosopher and what she claims
we believe. I, therefore, again invite the reader to approach the following comments on
common sense anew, without the preconceived (or perhaps post-philosophically
conceived) notion that common sense is threatened by scepticism and, must, therefore,
be defended thereﬁom. In particular, I invite her to re-examine the long held (and un-
ordinary) association between the vindication common sense and the refutation of the
sceptic. It is, after all, what the ordinary person believes which constitutes common
sense, not the beliefs of the non-ordinary person whose brain is nourished by a diet rich
in philosophy. For the association between the vindication of what is common and the
refutation of what is not represents not the ordinary person’s beliefs, but the
philosopher’s. In particular, the inclusion of second-order beliefs in common sense — in
addition to the ﬁrst-order — is an act of the philosopher, not a fact of the ordinary
person.

That the truth of scepticism can pose no threat to our common sense follows not
only from the fact that the common beliefs which comprise our ‘common sense’ do not
include such second-order beliefs, but also from the falsity of the singular assumption
that the standards which govern both ordinary (common sense) discourse and

scepticism are of the same kind, differing merely in strength, but not in kind. The

188

 

 

resulting conclusion that these are but two alternative sets of standards for the
applicability of one concept must be rejected forthwith. The assumption that they are
otherwise is but one more instance of the conﬂation of what is logically distinct, the
standards of appropriate and true assertion.

How then do I establish what our common sense behavior does or does not entail?

How, in particular, do I establish that it is not part of common sense to believe that we

.7

know what we believe or what we say we know? I suggest we take a closer look at t

that behavior. And, as we consider that behavior, we ask not merely what can explain

 

it, but also what minimal (non-extraneous) suppositions about that behavior we must
assent to in order that it be adequately explained. Once we reﬂect on our ordinary
behaviour, we ﬁnd that none of the above assumptions is necessary to its adequate
explanation.

In my efforts to explain our ordinary behaviour, and in particular, those aspects of
our ordinary behaviour the OLs, CTs, and RAs believed could be adequately explained
only by assuming it revealed second-order knowledge beliefs, I return to an example of
my earlier chapter H. Now, recalling my earlier discussion of Beau, my (apparent) dog.
What is it that I actually believe about him? What are my common sense beliefs about
Beau? I, of course, say of him many things. I say that he is my dog, that he is a perfect
specimen of the Great Pyrenees, that he is in the room with me, and alas, that he drools.
These are things I most deﬁnitely say and believe about Beau. No sceptic, however

talented, can convince me that I do not. (Equally, I think, no talented sceptic would

189

try”)

The sceptic asserts only that if I cannot know that I am not now dreaming, if I
cannot know that I am not now vatting, if I cannot know that I am not now the victim of
evil and cunning genius, then I cannot know these things that I believe. I do not (and
cannot) [mow that Beau is a dog. I do not (and cannot) know that he is a Pyrenees. I
do not (perhaps a good thing, for which I ought to be thankful!) know that he now
drools, nor that his drool is real, nor that it is about to be shaken from his gentle face
onto your newly cleaned clothes.

Such assertions from the sceptic tear at the heart of the anti-sceptic. For the anti-
sceptic hears her beliefs being challenged. She senses a threat to her most fundamental
beliefs. Beliefs which are dear not only to her but to all; the kind of beliefs which she
thinks make up common sense. Thus does DeRose assert that the truth of scepticism
lands us in systematic falsehood.”

The sceptic’s challenge is not, however, whether what I (and you) believe is true
(though, of course, one implication of what she says is that it may be false). Rather, she
challenges most speciﬁcally the belief that one can know; know, that is the propositions
that you and I believe.22 Know, that is, what one ordinarily believes (a belief the

unschooled in philosophy do not have). The sceptic’s challenge is directed to knowledge

 

20Anyone who drinks she has should review, as she has clearly misunderstood her.
21I discuss this argument in section iv of this chapter.

22The point, of course, is to distinguish the ﬁrst order-belief that P, from the second-order belief that
I know that P.

190

 

beliefs.

Wherein, then, lies her threat? From the beliefs which I surely do have about Beau --
that he is a dog, of Pyrenean origin, with a penchant to drool, especially in the company
of ﬁiends — it does, of course, seem as if I take myself to know them; and, moreover, it
seems (to some) that these kinds of beliefs are at the very heart of the ordinary individual’ 5
common sense. I think, however, that the attribution of the second-order belief that I
know what I believe is both unnecessary, and mistaken.

Suppose, for the sake of discussion, that you come for a visit. I do, after all, invite you
in, introduce you to Beau, call to Beau when it is time for his treat, feed him and wipe
away his drool. My behaviour seems to reveal that I believe many things, e.g., that you
are real, that Beau is real, that Beau needs to be fed, and so on. But is it necessary to the
explanation of my (or of your) behavior to suppose that I not only believe these things but
also that I believe that I know them? I think it is not. The explanation that I believe these
things — perhaps with great conviction — explains, by itself, why it is that I pet Beau and
wipe away his drool. Equally, it explains, by itself, why I invite you in, engage you in
discourse, and wave goodbye as you leave.

Supposing simply that I believe these things is quite sufﬁcient for explaining why I act
as I do. I believe that there is an external world. I believe that Beau is real. I believe that
you have a mind. I, therefore, treat Beau, you, and the word as if you are real. But
ordinarily I give no consideration to whether I have met a set of knowledge conditions
which justify my saying that I know them. Why do I, at times, say that I know them? I

do so to evidence the strong conviction with which I hold these beliefs, and my

191

 

iv
I ‘7

unwillingness to give them up. I do not say that I know to thereby evidence that I believe
I have met the objective conditions of knowledge. In fact, although it certainly seems to
be the case that I do believe these things, I do not (ordinarily) believe that I know them.
Ordinarily, I do not give any thought to whether I have met the conditions of knowing.
In fact, ordinarily, I give no thought to what those conditions are, not to any distinction
they might point to between mere strong conviction and knowledge.

Moreover, though you also believe these things, you do not (ordinarily) believe that you
know them; at least, I have no reason to suppose that you do. You evidence your belief
that there is a world, for example, that you are in it, that you are not now dreaming. But
nowhere, ordinarily do you evidence your belief that you know them, in addition to
strongly believing them. You, ordinarily, do not recite how you have defeated uncertainty,
that you are conscious that knowing has a speciﬁc set of conditions, nor that you have
scrutinized your evidence to make sure you meet them.

Though the respective beliefs we do have are, in fact, part of our 'common sense,‘
believing that we know them is not. Adding to the explanation of my, or anyone else’s,
ordinary behavior an entire additional set of beliefs does not add anything to the
explanation of that behavior. There is, in fact, nothing that this supposed additional set
of beliefs can explain that the original beliefs themselves cannot.

Putting one’s psychological and philosophical associations aside, then, I think we can see
that though it is the case that I believe these things, it is not the case (at least there is no
reason to suppose it is the case) that I believe that I know them. It is neither necessary nor

is it productive to suppose that, in addition to my ﬁrst-order beliefs about Beau and his

192

 

 

world, I have also the second-order beliefs that I know each of these things.” That in
addition to believing of Beau that he is a soft-eyed, gentle, Pyrrhonian drooler that I also
believe — as part of my share of our common sense -— that I know he is thus and so. In
fact, I see no reason to suppose that the second-order beliefs - that I know my common
sense beliefs - are really part of the content of my share of common-sense. In the absence
of any reason for supposing I do hold such beliefs, the supposition that I do is an artiﬁcial
addition to common sense, added on by a certain temperament, inattention, and conﬁrsion
of philosophers.

Suppose that upon visiting, you and I (and Beau) proceed to the parlour. I sit across
from you, and Beau, protective and cautious, near me. Suppose, moreover, that we begin
to discuss Descartes, Sextus Empiricus, or Hume. We laugh. We exclaim how difﬁcult it
is to teach students (ﬁrst year, especially) to take seriously the Cartesian challenge,
especially when the sceptic’s victory for the absurd seems to win hands down. I confess
that even I, at times, have trouble with the seeming absurdity of her alternatives: I may
now be dreaming, a mere envatted brain, or a victim of evil and cunning genius.24 I
concede the utter impossibility of suspending my ordinary beliefs about the world. Indeed,
I confess, that in one of my classes I burst into laughter (Perhaps the long winter. Perhaps

my common sense beliefs conﬂicting with the philosophical.) whilst trying to persuade my

 

23By ﬁrst-order beliefs I mean the beliefs I hold about the world and the things in it. By second-
order beliefs I mean the beliefs I may or may not hold about my ﬁrst order beliefs.

2‘Note: our trouble here is correct since we do not believe we know this things, and the doubts apply
only to knowing. It is therefore inappropriate to worry about removing these doubts from mere belief
(of the ﬁrst-order, that is).

193

 

‘1'“

 

students of the importance and centrality of the Cartesian Meditations on doubt; of the
possibility that Descartes — philosophically — was right (before his immediate fall back
into dogmatism); of the possibility that we do not really know anything at all.

We, (you and I, that is. Beau, though highly intelligent does not take part in the
conversation) go on to discuss what reasons Descartes' had entertained which might have
led him to conclude that such alternatives (to what we believe) are indeed relevant to
knowing. Why such troubling doubts ought, to anyone, to be troubling at all. At once,
however, my thoughts are interrupted: Perhaps I really do not know I am not now
dreaming. I recall a dream I once had in which you (quite unexpectedly) came to my door
and I, pleased to see you, invited you in. I remember that in my dream I introduced you
to Beau, we discussed Descartes, we sat in the parlour, you across from me, and Beau --
cautious and protective -- near me. I recall also that in my dream I told you the story of
my laughter in class, and of my students astonishment when my laughter continued. Of
course, I believe that I am not now dreaming, but I also, after considering my dream,
realize that I have no criterion within my present experience which would enable me to
know that I am not now dreaming. I believe that I am not, but I have never conducted a
study into making sure. How can I [mow that you are not now part of a dream in which
I tell you of a dream I once had? This is something I most assuredly believe, but I realize
I have never subjected that belief to anything like philosophical scrutiny. Initially, I search
ﬁendishly for some sign, and you with me, to prove I am now awake, that you really are
here with me, and that I am here with you.

It occurs to me, however, that whether Descartes' reasons to doubt are good or not,

194

 

whether his alternatives are relevant to the conditions of knowledge, is irrelevant to the
fact of my belief. I realize the sceptic threatens me only to the extent that I believe that I
know these things. Yet I do not believe that I know them. The fact that I have never
subjected these and the like beliefs to any set of standards implies clearly that I do not
believe that I know these things, no matter how tenaciously I cling to them. Hence, I
realize, the sceptic can pose no threat to my common sense. Again I relax.

You assert that surely we both at least believe that the dog is real, that the drool he has
now shaken onto your pants is also real, and that the conversation we've just had was not
a mere dream. There is simply no mistake about it. Indeed you insist that these beliefs so
are so strong and so ﬁrndamental that you cannot really bring yourself to doubt them. You
cannot engage in real doubt about them, the kind that leaves you in a state of uncertainty,
the kind that convinces you not to belief them until you have some ﬁrrther evidence.

But from the mere strength of our convictions need it follow that we believe that we
know? Or, nrightn’t it perhaps be that these beliefs are so fundamental, so non-negotiable,
to our common sense that we feel they are challenged by the sceptic when really they are
not? Mightn't it be that we confuse the strength of these convictions for the conviction
that we know them? After all, where in the context of common sense and ordinary
discourse do the objective conditions of knowing come up? They do not. The objective
conditions of knowledge are neither a part of the content of common sense nor of ordinary
discourse. Thus it is quite evident that it is not necessary to suppose that we have such
second-order beliefs; not necessary, that is, to the explanation of what we say or to the

explanation of what we do. It is not necessary to suppose we believe that we know what

195

 

we believe because we do not. We do not hold those beliefs which are the object of
sceptical doubt. At least, there seems to be no good reason for supposing that we do.

To the extent that you and I discuss knowledge at all, its objective conditions, the
relevance of certain alternatives to P, our discussion is not ordinary, nor is it about what
any of us would characterize as common sense. Rather, the discussion is philosophical,
everything what is common is not. Its concern is epistemic and justiﬁcatory; again, what
the common is not.

Moreover we can now see how the association between the vindication of common
sense and the refutation of scepticism might get started. It rests on a misunderstanding of
these relationships. It rests, in particular, on confusing the beliefs we do have for beliefs
they are not. It rests on confusing the strength and degree of conviction with which we
hold our beliefs for the strength of evidence required for knowledge. It rests on conﬁrsing
what we have a right to believe and assert, with the ﬁrlﬁllment of those conditions
necessary for true knowledge. It rests, that is, on conﬁrsing the fact of our fulﬁlling the
conditions of appropriate with those of true assertion.

One might object that our linguistic habits reveal that we do have such knowledge
beliefs. We do say, after all, that we know many things. Perhaps, however, our critic
infers with too much haste. Perhaps she is misled into believing that we believe we know
what we believes by too much philosophy. Perhaps she is again misled by the strength of
our (or her own) convictions, the strength of her ordinary beliefs. After all, it is not on the
basis of excluding uncertainty that I greet you, invite you in, or introduce you to Beau.

Equally, it is not on the basis of your exclusion of uncertainty that you get thru the day.

196

 

Yet, the critic does seem right about something: our claims to know must have some
explanation. Surely our behaviour which evidences our belief in an external world stands
also in need of explanation.

The question, however, is one of the right explanation. The right explanation, I think,
is that our conventions allow us to claim that we know, even when in fact we may not.
After all, one violates no convention when she does not normally attempt to prove she is
not now dreaming before brushing his teeth, dressing for the day, or walking his dog, any
more than I violate linguistic (or other) convention before I stroke Beau's white ﬁrr, give
him his daily treat, or wipe away his drool. Equally, I violate no convention when I greet
you, invite you in and introduce you to Beau. But what kind of convention governs the
acceptability of our behaviour? What kind of standards are the conventions which must
be met for our behaviour to count as acceptable.

I think the correct answer is this: the kind of conventions governing ordinary behaviour
are pragmatic, not semantic, not epistemological. As I argued in chapter 11, it is because
what governs ordinary discourse is standards of appropriate (not true) assertion that we
can say with impunity that we know, when in fact we may not know (witness my false
assertion that I know John will be at the party. A perfectly appropriate assertion, though
false). When I say I know that P, in ordinary conversation, it is not on the basis of
conscientiously applying a set of criteria I consider the objective criteria of knowledge.
Indeed, most ordinary people, when asked, concede they do not know the meaning of this
phrase ‘objective criteria of knowledge,’ nor what I have in mind when I ask them if they

think they can meet them. Importantly, in fact, if I ask of the individual on the street if she

197

 

is certain that P, it is not at all diﬂicult to get her to concede that she probably does not
know that P, even though she is pretty darn sure that P. If I ask, “Are you absolutely
certain that you know that P?” she concedes she is not. This concession, however, is not
part of her ordinary or common sense beliefs. It is arrived at upon reﬂexion. What she
does insist upon, however, is that she will continue to believe that the world is thus and
so no matter what I say. She will continue to believe that my dog is real, that she has
hands, and that she is not dreaming. These beliefs, she insists, are part of common sense.
I agree.

Believing that these things are so, however, is quite different than believing, upon
reﬂection and with intention, that one knows them absolutely, or that one has —- in her
common sense -- any conception of the objective criteria ofknowledge (whatever they may
be) which must be applied before she can correctly (as opposed to merely appropriately)
claim to know. It seems much more likely that what one has is a sense of what the
appropriate use of the term is, and it is that — that sense of appropriate assertion - which
explains her behavior, for it is that which gives her her linguistic direction.

What then is behind the ordinary claim to know, and the behavior ordinarily taken to
imply that one believes oneself to know? I propose ﬁrst of all, that it is one’s sense of
what it is appropriate to say, and secondly, that it is the intention of establishing the
differing degrees with which one is subjectively certain. I say that I know when I am pretty

darn sure, or when nothing could conceivably bring me to relinquish my belief.25 This

 

25In fact, on Dretske’s account of things, it is just this kind of irrelinquishability that gives me the
‘epistemic right’ to claim I know. See his 2000, “Ought we to follow the evidence?”.

198

behavior (as well as the standards that govern it) is learned as we learn the way the term
knowledge is appropriately used in our culture. If, for example, you ask me if I know that
the sun warms the earth. Without any reﬂection I would, ordinarily, most assuredly say
yes. If, by contrast, you ask if I know the building I see off in the distance is round, I
would, ordinarily, say I do not. My intention, however, is not to indicate that in the former
case I have mounds of evidence which I have scrutinized according to that set of
conditions I believe constitute the objective conditions of knowing, while in the latter case
I have not. It does not indicate that I have scrutinized the concept of knowledge, analyzed
its objective conditions, and carefully adjusted my beliefs accordingly. Not, anyway, if I
am an ordinary individual. Rather, I intend, I think, to convey to you that in the former
case I have a kind of belief about which no one can convince me I am wrong. The kind
of belief that I absolutely refuse to give up. In the latter case, by contrast, I indicate my
willingness to give up the belief in light of other considerations.

To the extent that I’m not that sure (in an appropriate sense of ‘ sure’) I will most likely,

,, 6‘

therefore, say I “think so,” as opposed to, “absolutely, without a doubt,” and so on. I
intend by my words to indicate to you that I am more subjectively certain of one than of
the other. The fact that I say I know the one and not the other is but my loose way of
letting you know that the one belief I will not relinquish whereas the other I (perhaps) will,
that the one belief is more certain than the other.

Why then are so many claims to know made (or implied) by so many ordinary people?

For precisely those reasons I have for claiming to know that the sun warms the earth, that

the moon is above my head, and the earth under my feet. One indicates thereby one's

199

 

1 ”mum-Kﬂ‘ M‘Iﬂ t

 

degree of conviction, one’s unwillingness to go without, one’s instinct. One does not, as
it has been widely assumed, indicate her epistemology, her semantics, her solution to
philosophical scepticism. The ordinary person habitually claims to know, but not because
she habitually believes that she knows. I accept her claim, as an ordinary person, not
because it is absolutely true that she knows. Rather, she claims to know, and I accept her
claim, because it is acceptable, given the conventions governing our linguistic culture. In
her case and in mine, the standards to which we appeal are the standards of appropriate,

not true, assertion.

Section iv: Objections and replies
First objection: One might, given too cursory a reading perhaps of this (and the previous)
chapters, conclude that ultimately there is no real reason to prefer my means of reconciling
common sense and scepticism to that of the contextualist, and perhaps even no reason to
prefer it to the position of the relevant altemativist. For, one might think, I have offered
but one more form of contextualism or relevant altemativism. One might conclude that my
compatibilism saves knowledge only by claiming that the objective standards of knowing
are diﬁ'erent at different levels of discourse. On this reading of my argument our ordinary
knowledge claims are true, and - in all but the most extraordinary contexts — the sceptic's
claim that I fail to know is false.

Such an objection, however, misses what is at the heart of my defense of common sense;
namely the separation in kind of two very different kind of standards. As I have shown,

once we distinguish the appropriate ﬁ'om the epistemic, we ﬁnd that common sense is

200

- LﬂuAM-‘I_.‘

 

 

 

made up not of complex second order beliefs about knowing, but rather, is made up of
simple ordinary beliefs about the world: I have a dog, his name is Beau, he is a drooler, and
the like. And these are not judgements about the truth of my claims; they are beliefs about
the world. On my understanding of common sense the question of whether my knowledge
claims are true represents a confusion about the kinds of standards that my knowledge
claim purports to meet with the kinds of claims it does not. Moreover, since on my
understanding the claims of ordinary speak are governed by the conditions of appropriate
assertion, they are really claims that it is appropriate to say that P; not that it is true —
according to the conditions of knowledge — to say that I know that P.

What follows, as I have argued, is that our ordinary knowledge claims are not
objectionable ﬁom the vantage point of one kind of standard; that which governs what is
appropriate. It does not follow, however, that these claims are unobjectionable from the
vantage point of another kind of standard; that which governs what is true. The
importance of the separation of the second-order ﬁom the ﬁrst (as well as the correlative
separation in kind of the standards governing each), and the consequent separation of the
subject matter of epistemology from the contents of common sense cannot be overstated.
With these distinctions, the very fear which serves as the foundation of ordinary language,
relevant alternative, and contextualist anti-scepticism proves irrational. Scepticism can
pose no threat where there is no claim to ﬁrlﬁll the conditions of knowledge

So too, therefore, does the multiplication of contextualized theories of knowing come
to an end. For it was only because common sense was thought to be threatened by

scepticism that these theories of knowing were developed. Each was designed speciﬁcally

201

 

to show how our common sense knowledge claims could be compatible with the truth of
scepticism. Each of the proponents of these contextualized theories of knowledge,
moreover, because of the dogmatic and false assumptions she made, could establish this
compatibility and save knowledge only by contextualizing the victory of the sceptic,
arguing that her standards are never applicable to ordinary knowing.

What I have suggested here, however, is that it is possible to save common sense
without contextualizing the sceptic’s victory. Moreover, one can argue for the
compatibility of common sense and scepticism without contextualizing the conditions of
knowledge, and without accusing the sceptic of distorting those same conditions. One can
achieve these results by establishing that the subject of common sense is distinct in kind
from that of the sceptic’s, by establishing that the kind of standards which govern ordinary
assertion and discourse are different in kind than those which govern knowledge, and by
establishing that common sense is not concerned with epistemology or meta-theory about
belief, but rather, direct belief and action.

Moreover, by ﬁnding that second-order knowledge beliefs are not a part of the content
of common sense, one can avoid a number of serious problems faced by the other theories
I have discussed. One avoids, for example, the proliferating meanings for the singular term
knowledge, solely for the reason of avoiding scepticism. Furthermore, one can respect the
time-honored distinction between mere belief and knowledge. And, perhaps most
importantly, arguing for the compatibility of scepticism and common sense on these
grounds allows us to acknowledge the possible truth of even radical scepticism without

thereby threatening common sense, and it does so without arbitrary or contextualized or

202

 

 

relativized standards of knowledge.

Second Objection:

There is, however, a further objection that I must consider. Keith DeRose constructs
this objection in “Solving the Sceptical Problem.”26 Because his objection is directed at
a position which appears very much like my own, I must not pass it by. Both the position
DeRose considers and the position I defend agree that if our ordinary claims to know are
interpreted as having an epistemic intent, they are no doubt false. (Importantly, the
position he constructs and then refutes does not consider whether our knowledge-claims
might be devoid of such intent. In this, the position described is signiﬁcantly different ﬁ'om
my own.) Additionally, the position which DeRose considers is an invariantist position.
(Where, by invariantism I mean a position according to which the conditions of knowing
are unchanging, illiquid, absolute.) Both the position I argue for and the invariantist
position which DeRose addresses represent alternative descriptions of the nature of
ordinary discourse. Each, in certain ways, parallels DeRose's contextualist description, but

from an invariantist, not a contextualist, perspective.27

 

Z‘The Philosophical Review, (1995) 104:1-52.

27I, of course, have not argued for such an invariantist position. What I have argued, however,
leaves open the possibility that an invariantist position is correct.

203

 

Suppose, DeRose suggests, that the invariantist argues the following: when one claims
to know in the context of ordinary discourse, she confuses the standards of warranted
assertability with those of knowledge. In fact, the invariantist goes on, though she meets
the conditions of warranted assertability, she does not meet the conditions of knowledge.
It follows that when one claims to know, her claim is false.28 Moreover, because any claim
of any S to know that (any) P is false, all of S's claims to know are systematically false.
This invariantist theory, of course, is precisely contrary to the theory DeRose would like
us to embrace, a theory which is contextualistic and relativistic; a theory according to
which we do meet the standards of knowing because those standards take into account our
contextually determined interests.

Of particular interest in DeRose’s response to invariantism is his contention that this
singular fact — that invariantism implicates us in systematic falsehood — is the reason to
prefer his own contextualist theory to the absolutist theory of the invariantist. One might
point out scepticism’s implicating us in systematic falsehood is hardly an objection to it.
Yet in response to a theory, or genre of theories, whose self-proclaimed raison d ’etre is
the avoidance of such a threat, no such response will do. A much better response is to
consider the form DeRose gives the invariantist position and ask if belief s perforrrring an
epistemic ﬁrnction is a necessary corollary of that position. We must ask, that is: Does

invariantism necessarily implicate us in systematic falsehood? By my lights, it does not.

 

28(Now this, of course, sounds very much like my claim that certain individuals confuse the
conditions of assertability with those of truth. As I will argue shortly, however, because I argue it is
the philosopher who makes this nristake, not the ordinary person, my and this variety of invariantism
have radically different consequences.)

204

 

At least, it need not.

DeRose, of course, is correct that the invariantist position he describes does implicate
us in systematic falsehood. But the invariantist theory he describes implicates us in
systematic falsehood only because it too takes for granted that common sense is concerned
with the same kind of standard as the epistemology. (The confusion of the two kinds of
standards is evidenced in the claim that we meet the standards of warranted assertibility.
It is also evidenced in the claim that our common sense knowledge claims are false. )
This, however, is a view which is mistaken. Moreover, the view that our common sense
knowledge claims should be judged according to the standards of knowledge is certainly
not a necessary accompaniment to an invariantist position. The invariantist need only
maintain that the objective standards ofknowledge are unchanging, irrespective of context.
He is not committed to any particular view about the nature or contents of common sense.

In fact the correct position, I suspect, is a combination of invariantism and the explicit
rejection of the supposition that common sense includes second-order judgements about
belief as a piece of unsupported dogma.” Once the dogmatic assumption that ordinary
common sense includes second-order epistemological belief is rejected, what becomes of
DeRose’s objection to invariantism? What becomes of his (apparently singular) reason
for preferring a contextualist semantics to an invariantist one?

Immediately, both DeRose’s objection and his singular reason for embracing a

contextualist semantics over an invariantist semantics loses all force. For it is only given

 

29Indwd I suspect (though I have not the space to defend my view here) that invariantism is the only
theory which makes any sense of either the power of scepticism or the concept of knowledge.

205

 

the supposition that our ordinary knowledge claims are epistemic in nature that
scepticism’s truth can imply their falsehood. It is only if second-order beliefs are included
in the contents of common sense that invariantism can (invariably) implicate our common
sense beliefs in systematic falsehood. Once we remove these uncritical assumptions, the
raison d ’etre of such defenses of common sense against philosophical scepticism
evaporates. And because, as I mentioned, there is no reason an invariantist cannot reject
this erroneous view of common sense whilst still maintaining her invariantism, equally,
therefore, there is no reason (by DeRose’s own account) to prefer contextualism to
invariantism. Invariantism need not implicate us in sytematic falsehood, then, any more
than need scepticism, and, of course, niether does once we understand the erroneousness
of that conception of common sense which places second-order epistemological beliefs
in its contents.

Once we recognize that second-order beliefs are not part of the contents of common
sense, then, we immediately see also that no systematic falsehood can result from the truth
of even the most radical invariantistic scepticism. Hence, DeRose’s objection to (at least
the variety I suggest) of invariantism is quite without power. Invariantism need not
implicate us in systematic falsehood, and does not, once we release it from the dogma of

traditional assumptions.

Section v: Concluding Remarks
Let us suppose, however, for the sake of argument, that none of my conclusions thus far

are correct. Second-order beliefs really are an ordinary component of common sense, and,

206

 

therefore, if the sceptic’s conditions of knowledge are unobjectionable, we fail to know,
systematically. Let us suppose, moreover, that we can -- as the ordinary language
philosopher, the relevant altemativist, and the contextualist suppose -- discriminate the
relevant ﬁ'om the non-relevant alternatives by appealing to ordinary linguistic practices.
Let us suppose also, therefore, that the conditions of knowledge the sceptic proposes are
not relevant, and that the real conditions of knoweledge change in accordance with
changes in the contextually sensitive determinants of relevance. Finally, let us suppose,
that because the conditions of knowing are thus contextually sensitive, we can correctly
claim to know, at least most of the time.

I then, in effect (and for the sake of argument), concede that the ordinary language
philosopher, the contextualist, and the relevant altemativist are correct: correct in
supposing that in certain contexts the sceptic’s criteria apply, but in most (because
ordinary) contexts they do not. I concede also, therefore, they are correct in supposing
that both we and the sceptic can be right (i.e., that common sense and radical scepticism
are compatible). I concede also, therefore, they are correct in supposing that we know
those things we ordinarily take for granted: that we have hands, that there is a ﬁnch in the
garden, that there is an external world. For, if the ordinary conditions of knowledge (as
opposed to the ordinary conditions of appropriate assertion) are determined by ordinary
context, it follows that we do not need to rule out any of the Cartesian alternatives to our
knowledge claims in order to correctly claim to know. We need not know we are not
merely dreaming. We need not know we are not brains-in-vats. We need not know that

we are not victims of evil and cunning genius. Why need we not know the falsity of these

207

 

 

alternatives to correctly claim to know? Because, ordinarily, no one rules them out.
Ordinarily, no one ever thinks of such alternatives. Ordinarily such alternatives as these
are, as Cohen says, "intersubjectively opaque. "

The following assertions are, then, not only unobjectionable from an epistemological
standpoint, they are true: "I know I am not merely dreaming, even though for all I know
I might be dreaming." "I know I have hands, even though for all I know I may be a
handless brain in a vat. ” "I know that two plus two is four, even though for all I know two
plus two is not four, but merely seems to be four because I am systematically deceived by
an evil and cunning genius." In fact, not only are these assertions true, but for the
contextualist at least, so are their negations.30 Equally, the follwing set of assertions are
also both unobjectionable and true: "I know my dog is real, even though for all I know
he is a fake." "I know there are other minds, even though for all I know there are not."
"I know there is an external world, even though for all I know there is not."

Under these conditions, what then does it mean to know? What does our knowledge
actually amount to? What criteria, if any, must we meet? How does knowledge differ
from mere (even true) belief? First, because on each account, there is no one singular set
of conditions which must be met for S to correctly claim to know that P, knowing does not
require of you the same things it may require of me. Moreover, since the conditions I must

meet to know may differ ﬁom the conditions you must meet, I can know that P while you

 

30See Cohen's ”Knowledge," and his ”Contextualism, Skepticism, and the Structure of Reasons,"
Philosophical Perspectives 13, (1999). See also DeRose, ”Solving the Sceptical Probelm" Both
DeRose and Cohen explicitly commit to the possiblity that both the ordinary claim to know and the
(extraordinary) claim of the sceptic that I fail to know are true.

208

 

 

J

 

know that-P. Hence, truth, it seems, cannot be a condition of knowledge. Furthermore,
because we need only rule out those alternatives to P which the context brings relevance,
the justiﬁcation of P becomes such a subjective matter that it is hardly a requirement at all.
This leads to the further problem that mere belief seems to become indistinguishable ﬁom
knoweldge.

How satisfactory, in light of its meaning and consequences, is this victory. How
satisfactory, under these conditions, is it to know what we ordinarily believe? By my
lights, this kind of victory -- in the face of scepticism -- is empty. This kind of victory
yields no satisfaction. Why is it empty? Why does it yield no satisfaction? Whereﬁ'om
comes the honest philosopher’s feeling of deep and unavoidable dissatisfaction with her
true knowledge? It comes from ﬁrst our intuitive understanding that knowing is different
from mere belief. It comes from the fact that knoweldge is not of the false, but of the true.
Moreover, it comes from the fact that the success our knowledge claims, on this theory,
is contingent upon our willingness to make knowledge a function of inter-subj ectivity: the
intersubj ectively evident, the inter-subj ectively ignored, the inter-subj ectively presupposed.
In fact, insofar as saving knowledge from scepticism was one of the unifying goals of not
only the contextualist, but the relevant altemativist and the ordinary language philosopher
as well, the anti-sceptic's I have considered seem to fail. For surely this is not the
'knowledge' we want rescued from the threat of scepticism. In fact it does not look like
knowledge at all.

In the last analysis, we are not obligated to fulﬁll any conditions not ordinarily imposed

by the average speaker in an average conversation. As Lewis states in “Elusive

209

knowledge,” “we are entitled to ignore whatever alternatives to P those around us
ignore.”31 This same conviction, as we discovered in earlier chapters, is found also in
Austin, DeRose, Cohen, Stine, and others. And, since ordinarily in ordinary conversation
we ignore the Cartesian alternatives to our knowledge claims, it follows that we need not
rule them out to correctly claim to know. Should the conversation change direction, away
from the ordinary and toward the extra-ordinary (ala the sceptic, for example) then, either
the pressupositions necessary to make our knowledge claims true “spring into existence,”32
or the higher level of standards brought into play—no matter how compelling -- because
they are not ordinary, need not be met.33 All of this thanks to a contextualistic semantics.

Among the consequences, therefore, of embracing this genre of anti-scepticism, are the
following: truth, certainty, and justiﬁcation are not (necessarily) necessary conditions of
knowledge.“ Mere belief, even true belief, cannot be distinguished ﬁom knowledge. We
know, but only to the extent that we ignore presuppose, assume, make ad hoc exceptions,
or beg the question. That, after all, is how people ordinarily behave. As Lewis concedes

in “Elusive Knowledge,”

I have to grant, in general, that Knowledge just by presupposing and

 

3'Lewis, “Elusive Knowledge,” 559.

32An argument of Lewis from his earlier paper “Scorekeeping and the language Game,” 2.

33I allude to DeRose’s “Solving The Sceptical Problem,” 36.

3“Of course none of the philosopher’s we’ve discussed states this explicitly. To do so would bring
instant scrutiny, something a position without supporting argument cannot afford. Each, however,
with the possible exception of Austin, does imply that P’s being true is not a necessary condition of
knowing that P.

210

 

 

‘t’ld

 

ignoring is knowledge.”

This is a theory which, by Lewis’ admission, avoids scepticism only by making ad hoc
exceptions,”S a theory the only ‘evidence’ for which (by the admission of Lewis, Stine,
DeRose, and others)37 is its capacity to both embrace what is ordinary and ignore what is
not. A theory, that is, which is victorious over scepticism, but only to the extent that it
throws out the heart of the issue, namely, justiﬁcation, a theory in which our ‘knowledge’
is ‘elusive’; vanishing (as Lewis remarks) the moment we pay attention to a higher
standard of knowing. Surely this is a theory whose victory is cheap indeed!

To leave out of epistemology what is arguably its most fundamental concept,
justiﬁcation, is to leave out epistemology altogether. By my estimation, therefore, what
each of the philosophers I have considered is doing is not epistemology, but something
much different, psychology I suppose.38 And, of course, changing the subject matter is not
a means of defeating the sceptic, but only of avoiding her. Mere avoidance, however, is
not victory.

Equally, as I have argued throughout, the ordinary language philosopher, the
contextualist, and the relevant altemativist have given no philosophical grounds for

rejecting any of these conditions on knowledge. Inasmuch, moreover, as the OLs, CTs,

 

3’Lewis, “Elusive,” 562.

 

 

3”As I quoted in chapter 111, Lewis concedes the ad hoc nature of a central assumption he makes to avoid

scepticism. See my discussion of this in chapter 111 of this work, 23.
37See my discussion of these in chapters III and IV of this work.

38In this last point, I am very much in agreement with, as well as indebted to, the opinions of
Jaegwon Kim, as argued for in his "What Is 'Naturalized Epistemology?" (1988)

211

and RAs are responding directly to the threat of the philosophical scepticism, they clearly
have not met (nor, in fact, showed much interest in meeting) their burden of proof.39
Rather than telling us what about knowledge makes possible its co-existence with
question-begging, assuming, and ignoring, they have simply appealed to convention.
Moreover, in so appealing to convention they have overlooked the distinction between two
fundamentally different kinds of standards: the pragmatic and the justiﬁcatory. They have
confused ﬁrst for second-order belief, and only on the basis of the confusions, they have
laid the objective conditions of knowing before the subjective conditions of group
behavior.

What is wrong with their kind of knowledge? What makes the proclamation of it and
of the consequent victory over philosophical scepticism so vacuous? As I have indicated,
winning an argument still requires the giving of reasons. It requires also that we not beg
the very question at issue. It requires that we have a better reason than mere dislike for
rejecting our opponent's position. Moreover, any winning argument ought not to violate
fundamental laws of thought ( the law of non-contradiction, for example, not to mention
excluded middle.)40 Equally, it ought to defend (not merely assume)the foundations from

which it derives.

 

39'To my near astonishment, DeRose seems to consider this a virtue,or at least, nothing to be ashamed
of. He states, for example, in his "BIVs," "1 don't make this knowledge that I claim we have
dependent on any argument I give - or any argument whatsoever." And, ”As a Moorean, I face the
pointed question: How do we know that we are not BIV's? I join my contextualist Moorean cohorts
[Stine , Cohen, and Lewis] in answering negatively: not by means of any argument, nor even by
means of out having any eﬂective evidence to that effect." 135.

”I allude to the consequence of this kind of contextualistic semantics that the same proposition can at
the same time be both true and false.

212

In addition to these and the other philosophical criticisms I’ve made throughout this
work, this philosopher still believes that what determines both the conditions of knowing,
as well as the conditions of their ﬁrlﬁllment, is not the average IQ or linguistic habits of
one’s peers. Such a means of determining knowledge’s truth conditions is no more
rational than determining the conditions of sound argument, the morally good, or the
beautiﬁrl by popular opinion polls. Furthermore, this philosopher still believes (especially
in light of no argument to the contrary) that a necessary condition of knowledge is truth,
that a corrollary of correctly claiming to know that P is that one’s belief that P is justiﬁed,
and that there is a legitimate distinction between mere belief and knowledge. This
philosopher is not satisﬁed with knowing that her dog is real, nor that his Pyrronean origins
are true, nor that he really is a drooler, without ﬁrst knowing she is not merely an envatted
brain, not the victim of evil and cunning genius, nor merely dreaming. This philosopher
is not satisﬁed with knowing simply by ignoring, presupposing, or making ad hoc
exceptions for what is difﬁcult or impossible to prove.‘1 Finally, this philosopher is not
satisﬁed by arguments made -- at the expense of reason -- the sole purpose of which is to

avoid the consequences of a position one does not like.

"The objective standards of knowledge; the nature of the external world; the reliability of
Perception, and the like.

213

 

 

WORKS CITED

Annis, David. 1978. “A Contextualist Theory of Epistemic Justiﬁcation.” American
Philosophical Quarterly 15 : 105-1 12.

Audi, Robert. 1993. “Contemporary Foundationalism.” The Theory of Knowledge. Ed.
Lewis Pojman. Belmont, CA: Wadsworth Publishing.

Austin, J.L. 1979. “Other Minds.” Austin: Philosophical Papers. Ed. 10. Urrnson, and
G.J. Warnock. Oxford: Oxford University Press.

Bonjour, Lawrence. 1993. “A Critique of Foundationalism.” The Theory of Knowledge:
Classic and Contemporary readings. Ed. Louis P. Pojman. Belmont, CA:
Wadsworth Publishing.

Bouwsma, OK. 1965. “Descartes’ Evil Genius.” Bouwsma: Philosophical Essays.
Lincoln: University of Nebraska Press.

Cohen, Stewart. 1987. “Knowledge, Context, and Social Standards.” Synthese 73: 3-26.

DeRose, Keith. 2000. “How Can We Know that We’re Not Brains in Vats?” The Southern
Journal of Philosophy XXXVIII: 121-149.

. 2000. “Ought we to Follow Our Evidence?” Philosophy and Phenomenological
Research LX: 697-706.

. 1995. “Solving the Sceptical Problem.” The Philosophical Review 104: 1-52.

1992. “Contextualism and Knowledge Attributions.” Philosophy and
Phenomenological Research LII: 916-929.

Descartes, Rene. 1911. Meditations. Trans. Elizabeth S. Haldane, and G.R.T. Ross.
Cambridge: Cambridge University Press.

Dretske, Fred. 2000. “Entitlement: Epistemic Rights Without Epistemic Duties?”
Philosophy and Phenomenological Research 60, No. 3, 591-611.

214

 

. 1970. “Epistemic Operators.” Journal of Philosophy 67: 1007-23.
210
Empiricus, Sextus. 1933. Outlines onyrrhonism, Book]. Trans. R.G. Bury. Cambridge:
Harvard University Press.

Goldman, Alvin. 1976. “Discrimination and Perceptual Knowledge.” Journal ofPhilosophy
73: 771-91.

Hume, David. 1988. A Treatise of Human Nature. Ed, L. A. Selby-Bigge. Oxford: Oxford
University Press.

Kim, Jaegwon. 1988. “What is Naturalized Epistemology?” The Theory of Knowledge:
Classic and
Contemporary Readings. Ed. Louis P. Pojman. Belmont, CA: Wadsworth
Publishing.

Lewis, David. 1996. “Elusive Knowledge.” Australasian Journal of Philosophy 74: 549-
567.

. 1979. “Scorekeeping in a Language Game.” Journal of Philosophical Logic 8: 339-
3 3 5.

Moore, GE. 1962. GE. Moore: Philosophical Papers. New York: Collier Books
Nozick, Robert. 1981. Philosophical Explanations. Cambridge: Harvard University Press.
Popkin, Richard. 1980. “Skepticism and Anti-Skepticism in the latter part of the Eighteenth
Century.” The High Road to Pyrrhonism. Ed, Richard Watson and James E. Force.
Cambridge: Hackett Publishing Company.

1965. “The High Road to Pyrrhonism.” American Philosophical Quarterly 2,
January: 1- 15.

Stine, Gail. 1976. “Skepticism, Relevant Alternatives, and Deductive Closure.”
Philosophical Studies 299: 249-61.

Stroud, Barry. The Significance of Philosophical Scepticism. Oxford: Clarendon Press.

Unger, Peter. 1975. Ignorance: A Case for Scepticism. Oxford: Oxford University Press.

. 1971. “A Defense of Skepticism.” Philosophical Review 80.

215

 

WORKS CONSULTED
Cohen, Stewart. “Two Kinds of Skeptical Argument.” Philosophy and
Phenomenological Research 68, No.1, (1998): 143-159.

DeRose, Keith. “Knowledge Assertions and Lotteries.” Australasian Journal of
Philosophy 74, No.4, (1996): 568-580.

Feldman, Richard. “Contextualism and Skepticism.” Philosophical Perspectives 13
(1999): 92-113.

Heller, Mark. “The Proper Role For Contextualism in an Anti-Luck Epistemology.”
Philosphical Perspectives 13 (1999): 115-129.

Putnam, Hillary. Reason Truth and History. Cambridge: Cambridge University Press,
1981.

Reiber, Steven. “Skepticism and Contrastive Explanation.” Nous 32, No. 2, (1998):
189-204.

Schiffer. S. “Contextualist Solutions to Scepticism.” In The Proceedings of the
Aristotlean Society held in the Senior Common Room 20 May (1996): 317-334.

216