Upload
anthony-brueckner
View
222
Download
2
Embed Size (px)
Citation preview
Wright on the McKinsey Problem
anthony brueckner
University of California, Santa Barbara
The McKinsey Problem concerns a puzzling implication of the doctrines of Con-
tent Externalism and Privileged Access. I provide a categorization of possible solu-
tions to the problem. Then I discuss Crispin Wright’s work on the problem. I
argue that Wright has misconceived the status of his own proferred solution to the
problem.
I would like to raise an objection to Crispin Wright’s proposed solution
to the McKinsey Problem.1 The problem can be set up as follows.
According to the doctrine of content externalism (CE), there is a depen-
dence between intentional mental states and their thinkers’ external
environments. Thus CE has the Consequence that for some external
world proposition E (a proposition which entails the existence of an
external, physical world), the proposition that, say, I am thinking that
water is wet logically ⁄ conceptually implies E. (In general, p logi-
cally ⁄ conceptually implies q iff ‘‘there is a correct deduction of q from
p, a deduction whose only premises other than p are necessary or con-
ceptual truths that are knowable a priori, and each of whose steps fol-
lows from previous lines by a self-evident rule of some adequate system
of natural deduction.’’ (McKinsey 1991: 14)) I have an a priori warrant
for believing that I am thinking that water is wet, since my warrant for
this belief does not depend upon empirical investigation. Call this claim
about a prioricity Privileged Access (PA). A plausible Closure Principle
for A Priori Warrant (CPAPW) says that if I can have an a priori war-
rant for believing p, and if p logically ⁄ conceptually implies q, then I
can have an a priori warrant for believing q. So: I have (and so can
have) an a priori warrant for believing that I am thinking that water is
wet (by PA); thus I can have an a priori warrant for believing external
1 The locus classicus of the problem is McKinsey 1991. For a recent considered dis-
cussion of the problem, see McKinsey 2002. For Wright’s proposed solution, see
Wright 2000. For a recent considered discussion, see Wright 2003.
Philosophy and Phenomenological ResearchVol. LXXVI No. 2, March 2008� 2008 International Phenomenological Society
WRIGHT ON THE MCKINSEY PROBLEM 385
world proposition E (by CE’s Consequence and CPAPW). But I do not
have, and cannot acquire, an a priori warrant for believing E. So: the
conjunction of PA and CE gives rise to an apparent problem—the
McKinsey Problem.2
Let us consider a categorization of possible solutions to the problem.
A Category I solution will consist in either the denial of PA, the denial
of CE, or the denial of both.3 Proponents of Category I solutions are
driven to deny one or the other of the plausible doctrines PA and CE
because they think that (1) the a priori warrantability of E follows
from the conjunction of the two doctrines, and (2) E is not a priori
warrantable.
As I argued in Brueckner 1992, the assessment of the McKinsey
Problem depends upon the specification of E. Clearly, proponents of a
Category I solution are thinking that there is some problematic exter-
nal world proposition which is in fact not a priori warrantable but
which would be a priori warrantable if both PA and CE were true.
Candidates for such an E are: water exists; either water exists or I inha-
bit a speech community in which speakers have theorized that H2O
exists.
Let us turn to the other possibility in the proposed categorization
of possible solutions to the McKinsey Problem. A Category II solution
will not involve the denial of either PA or CE. A Category IIA solu-
tion will consist in the claim that there is no external world proposi-
tion E which is such that E’s a priori warrantability follows from the
conjunction of PA and CE. According to Category IIA solutions,
there is therefore no pressure to deny either of the two plausible
doctrines.4
A Category IIB solution will consist in the claims that (A) there are
some external world propositions (the E’s) whose a priori warrantabili-
ty follows from the true doctrines PA and CE (the E’s being the only
external world propositions whose a priori warrantability so follows),
and (B) the E’s are in fact a priori warrantable. According to Category
IIB solutions, there is therefore no pressure to deny either of the two
plausible doctrines PA and CE. Candidates for allegedly unproblematic,
2 This setting up of the problem follows McKinsey 2003, but unlike McKinsey, I have
focussed up a priori warrant rather than a priori knowledge. This is because
Wright’s proposed solution to the problem focusses upon warrant.3 For example, McKinsey 2002 contains the denial of PA, and Boghossian 1997
contains the denial of CE. As far as I know, no one in the literature denies both
doctrines.4 See for example Brueckner 1992. In that paper, though, I consider a Category IIB
solution as well. See the discussion below in the text. A denial of CPAPW would be
sufficient to ground a Category IIA solution.
386 ANTHONY BRUECKNER
a priori warrantable external world propositions include: either water
exists or I am a member of a speech community which possesses the
concept of water; there is an external, physical world.5 Once a philoso-
pher has made a convincing case for (A) and (B) above, he has done
enough to provide a convincing solution to the McKinsey problem.
Let us leave behind our categorization for a moment and consider
Crispin Wright’s proposed solution to the problem. Wright begins by
drawing a distinction between the Transmission of warrant across
known entailment and the Closure of warrant across known entail-
ment. According to Closure, if I have warrant for p, and I know
that p entails q, then I have warrant for q.6 Wright does not deny
Closure: he is skeptical about the existence of any genuine counte-
rexamples to the principle.7 According to Transmission, if I have
warrant for p, and I know that p entails q, then my warrant for p
transmits across the known entailment in the following sense: I can
begin by having a warrant for p while lacking a warrant for q, then
come to know the pertinent entailment, and then thereby come to
acquire a warrant for q which I previously lacked. Suppose, for
example, that I have a warrant for believing G (=A goal has just
been scored) and that I know that G entails S (=A game of soccer
is in progress). My warrant for believing G is that a game of soccer
is apparently in progress and a player has just driven the ball into
the net. In this case, I cannot start in the position of having that
warrant for believing G while lacking a warrant for believing S and
then end in the position of acquiring a warrant for believing S by
means of appreciating the entailment from G to S. This is because
the warrant I have for believing G depends upon having a prior
warrant for believing S and so cannot generate a new, heretofore
absent warrant for believing S.8 Still, even though such a case argu-
ably refutes Transmission, it does not refute Closure. Indeed, as
Wright points out, my warrant’s failure to transmit in the soccer
case presupposes that Closure is not violated: if I do have a warrant
for believing G, then I must have a prior, independent warrant in
place for believing S.9
5 For discussion of the first external world proposition, see Sawyer 1998. For discus-
sion of the second, see Brueckner 1992 and Brueckner 1999. For some other Cate-
gory IIB solutions, see Brewer 2000 and Miller 1997.6 Closure is a different principle from McKinsey’s CPAPW. We will discuss their con-
nections below.7 See Wright 2003, p. 58.8 This case is first discussed in Wright 2000.9 See Wright 2003, p. 67 and p. 77.
WRIGHT ON THE MCKINSEY PROBLEM 387
There is much to be said about the nuances of the various cases in
which Transmission apparently fails.10 For our present purposes, suffice
it to say that Wright holds that the crux of his solution to the McKin-
sey Problem is this: my warrant for believing that I am thinking that
water is wet fails to transmit across the entailment embodied in Conse-
quence, viz. the thesis that the proposition that I am thinking that
water is wet logically ⁄ conceptually implies (and hence entails) an exter-
nal world proposition E.
To see the basis for Wright’s claim about Transmission failure in
the context of the McKinsey Problem, we need first to see which
external proposition is Wright’s candidate for E. Though the matter
is not entirely clear, it is fair to say that Wright’s candidate is the
proposition that my speech community has had contact with water
that is sufficient for the genuine contentfulness of the apparent
thoughts I express by using the term ‘water’.11 Call this proposition
EW. On Wright’s view, I have a special sort of warrant for EW
which he calls an a priori entitlement. This a priori entitlement ‘‘is
conferred not by positive evidence for…[EW] but by the operational
necessity, so to speak, of proceeding on the basis of such so far
untested assumptions if one is to proceed at all.’’ (Wright 2003: 67)
The ‘‘so far untested assumptions’’ are presumably those embodied
in EW, regarding ‘‘my speech community’s historical interactions
with water.’’ (Wright 2003: 69) It is not completely clear which way
of ‘‘proceeding’’ on the basis of the ‘‘so far untested assumptions’’
Wright has in mind. He may be thinking of our proceeding to use
such concepts as water. Or he may be thinking of our proceeding to
form second-order beliefs about our thoughts involving the term
‘water’. It is not important for our present purposes to settle this
interpretive issue.
Let us return to Transmission. Wright holds that the warrant that I
have for believing T (=I am thinking that water is wet) does not trans-
mit to EW across the entailment embodied in Consequence. According
to Wright, in order to have a warrant for believing T, I need to have a
prior, independent warrant for believing EW. I do have such a prior
warrant: it consists in my a priori entitlement to believe EW. So I can-
not acquire a heretofore lacking warrant for believing EW by the trans-
mission of my warrant for believing T across the entailment embodied
in Consequence (using my knowledge of that entailment). Having the
10 See, e.g., Beebee 2001, Brown 2003, Brown 2004a, Davies 1993, Davies 1998,
Davies 2000, Davies 2003, and McLaughlin 2003.11 See Wright 2003, pp. 68-9.
388 ANTHONY BRUECKNER
latter warrant for believing T requires having an antecedent, indepen-
dent warrant for believing EW.12
Once again, we see that a failure of Transmission presupposes that
Closure is not violated: I have warrant for believing T, I know that T
entails EW (since I know that Consequence is true), and I also have
warrant for believing EW (since I have an a priori entitlement to
believe EW).13
Let us return to our categorization of possible solutions to the
McKinsey Problem. Contrary to what Wright maintains, the crux of
his proposed solution to the problem is not the failure of Transmissibil-
ity in the cases that are relevant to the problem. Indeed, the problem
has nothing to do with Transmission. So long as one is going to avoid
Category I solutions and hold on to PA and CE, the issue that a solu-
tion-provider must consider is whether PA and CE guarantee the exis-
tence of some a priori warrantable E. Proponents of Category IIA
solutions say ‘‘No,’’ while proponents of Category IIB solutions say
‘‘Yes.’’ Wright is clearly a Category IIB man. He holds that the exis-
tence of an a priori warrantable E would follow from PA and CE, viz.
EW. Since he accepts PA and CE, he maintains that EW is in fact a
priori warrantable. EW is a priori warrantable, according to Wright,
because we have an a priori entitlement to believe EW. If Wright is
correct on this score, then this is the end of the story. That is, if Wright
12 This is the analysis of Transmission failure in the McKinsey inference presented in
‘‘Cogency and Question-Begging: Some Reflections on McKinsey’s Paradox and
Putnam’s Proof’’ and in the first half of ‘‘Some Reflections on the Acquisition of
Warrant by Inference.’’ Towards the end of the latter paper (sections 5 and 6),
Wright considers a more complex analysis of Transmission failure. He holds that
my warrant for believing T requires a background entitlement regarding the ‘‘integ-
rity’’ of my concepts. In the case of my concept water, the entitlement in question
is that when I token ‘I am thinking that water is wet’, I am thinking a contentful
thought and not suffering from a ‘‘content illusion’’ engendered by a lack of an
appropriate natural kind. What I am a priori entitled to believe, then, is that my
speech community has interacted with some watery stuff or other—-H2O, XYZ,….
Call this proposition EW*. My warrant w for believing T, then, will not trans-
mit to EW* via a McKinsey-style inference, since w presupposes a prior entitlement
to believe EW*. According to Wright, the (modest) good news is that w will trans-
mit to EW: w presupposes an entitlement to believe EW*, and since ‘‘it is a priori
that water is the watery stuff of our actual acquaintance,’’ w generates warrant for
believing EW (in virtue of w’s presupposed background entitlement regarding
EW*). Thanks to Jonathan Vogel for a helpful discussion of these points.13 In the papers cited in fn. 10, Martin Davies has developed (independently of
Wright) a response to the McKinsey Problem that centers around Transmission
failures. I have not discussed Davies’ work here because it is not completely clear
whether he holds that even though a priori warrant is not transmitted in the perti-
nent cases, one nevertheless does have a priori warrant for some external world
proposition E. The significance of this difference between Davies and Wright will
become clear shortly.
WRIGHT ON THE MCKINSEY PROBLEM 389
is correct on this score, then we have a successful Category IIB solu-
tion to the McKinsey Problem. Failure of Transmission is a separate
issue. If it is plausible that I have an a priori warrant for believing
EW—whatever the source might be, ‘‘operational necessity’’ or some-
thing else—then the McKinsey Problem is solved. It does not matter
for the problem’s solution whether or not the a priori warrant is trans-
mitted across the known entailment embodied in Consequence, so long
as the a priori warrant is in fact somehow in place.14
Wright might reply that if his position is correct, then he has shown
that objectionable ‘‘armchair knowledge’’ is not possible: I cannot
acquire a warrant for believing EW by deducing it from T while in my
armchair. Nevertheless, the McKinsey Problem was this: does the con-
junction of CE and PA entail the a priori warrantability of some exter-
nal world proposition that is not in fact a priori warrantable? Wright
has provided a solution to this problem that is independent of Trans-
mission failure: according to Wright, the conjunction of CE and PA is
true and does entail the a priori warrantability of EW—an external
world proposition for which we in fact have an a priori entitlement.
References
Beebee, H. 2001. ‘‘Transfer of Warrant, Begging the Question, and
Semantic Externalism.’’ Philosophical Quarterly 51: 356–74.
Boghossian, P. 1997. ‘‘What the Externalist Can Know A Priori.’’ Pro-
ceedings of the Aristotelian Society 97: 161–75.
Brewer, B. 2000. ‘‘Externalism and A Priori Knowledge of Empirical
Facts.’’ In New Essays on the A Priori, edited by P. Boghossian
and C. Peacocke. Oxford: Oxford University Press, 415–32.
Brown, J. 2003. ‘‘The Reductio Argument and Transmission of
Warrant.’’ In New Essays on Semantic Externalism, Scepticism, and
Self-Knowledge, edited by S. Nuccetelli. Mass.: MIT Press, 117–30.
——. 2004a. ‘‘Wright on Transmission Failure.’’ Analysis 64: 57–67.
——. 2004b. Anti-Individualism and Knowledge. Mass: MIT Press.
14 McKinsey’s own take on Wright in McKinsey 2003 is as follows. Wright accepts
PA and CE and denies neither Closure nor CPAPW. Wright is thus committed to
the a priori warrantability of some external world proposition. (In my parlance,
Wright is a Category IIB man.) Even if Transmission does fail in the context of the
McKinsey Problem, this is irrelevant because no external world proposition is in fact
a priori warrantable. So, McKinsey concludes, Wright has provided no solution to
the McKinsey Problem even is it is true that Transmission fails in the cases relevant
to the problem. The foregoing view of Wright, however, dismisses out of hand all
Category IIB solutions to the McKinsey Problem. An anonymous referee drew my
attention to a discussion of Wright by Jessica Brown in Brown 2004b. On p. 257,
she makes a point that is close to McKinsey’s in this footnote.
390 ANTHONY BRUECKNER
Brueckner, A. 1992. ‘‘What an Anti-individualist Knows A Priori’’.
Analysis 52: 111–18.
——. 1999. ‘‘Transcendental arguments from content externalism.’’ In
Transcendental Arguments: Problems and Prospects, edited by R.
Stern. Oxford: Oxford University Press, 229–50.
Davies, M. 1993. ‘‘Aims and claims of externalist arguments.’’ Philo-
sophical Issues 4: 227–49.
——. 1998. ‘‘Externalism, architecturalism, and epistemic warrant.’’ In
Knowing Our Own Minds, edited by C. Wright, B. Smith and C.
Macdonald. Oxford: Oxford University Press, 321–61.
——. 2000. ‘‘Externalism and armchair knowledge.’’ In New Essays on
the A Priori, edited by P. Boghossian and C. Peacocke. Oxford:
Oxford University Press, 384–414.
——. 2003. ‘‘The problem of armchair knowledge.’’ In New Essays on
Semantic Externalism, Scepticism, and Self-Knowledge, edited by S.
Nuccetelli. Mass.: MIT Press, 23–55.
McKinsey, M. 1991. ‘‘Anti-individualism and Privileged Access.’’ Anal-
ysis 51: 9–16.
——. 2002. ‘‘Forms of externalism and privileged access’’. Philosophical
Perspectives (Language and Mind) 16: 199–224.
——. 2003. ‘‘Transmission of warrant and closure of apriority.’’ In
New Essays on Semantic Externalism, Scepticism, and Self-Knowl-
edge, edited by S. Nuccetelli. Mass.: MIT Press, 97–115.
McLaughlin, B. ‘‘McKinsey’s Challenge, Warrant Transmission, and
Skepticism.’’ In New Essays on Semantic Externalism, Scepticism,
and Self-Knowledge, edited by S. Nuccetelli. Mass.: MIT Press, 79–
96.
Miller, R. 1997. ‘‘Externalist Self-knowledge and the Scope of the A
Priori’’. Analysis 57: 67–75.
Sawyer, S. 1998. ‘‘Privileged Access to the World.’’ Australasian Jour-
nal of Philosophy 76: 523–33.
Wright, C. 2000. ‘‘Cogency and Question-begging: Some Reflections on
McKinsey’s Paradox and Putnam’s Proof.’’ Philosophical Issues 10:
140–63.
——. 2003. ‘‘Some Reflections on the Acquisition of Warrant by Infer-
ence.’’ In New Essays on Semantic Externalism, Scepticism, and
Self-Knowledge, edited by S. Nuccetelli. Mass.: MIT Pres, 57–77.
WRIGHT ON THE MCKINSEY PROBLEM 391