..n
. .
H..-

2,
LT

A?

,1,
{1
£151.
7: .1
.g; . ,
..... 4 .
\

9. ,p

.6“ rw:
Wu”; #1

q i: 5

tin

.-
X.
71”.

u“,
‘. Mu... ...

 

 

 

 

 

..
0;. .31.. .iv' .
:0...» .. gay”... in
2 3:? s... 1..
.. Ibr: J
“133.4

 

 

in}...
5....
. into:
4 1...; 1
.. .1

. I. ..
.x.d.....1.. .. 39L-..
C sun:

I
v?“ I... )l -
§§¥3 «

          

opt... Q

a...

 

.. 3.!— .::r.:-. 7'...-
.:.... ;..rf.1.1.,.. 1...... .5: 4.1.

 

. .<
l, .1...“ xi 1:... .9 . ..

 

 

1: raisgrxuma :.

‘zs.Y.¢..u.l _ . ‘

 

illlllllllllllllllllllHill‘llllllllll‘lllllllHHIUI

31293 017718150

LIBRARY

Michigan State ’
E University

 

 

 

 

This is to certify that the

dissertation entitled

Mental Causation and the Pragmatics
of Explanation

presented by
Joseph Charles Totherow

has been accepted towards fulfillment
of the requirements for

Ph . D . degree in W '

 

 

Jfl/fl r// 9/

ijgr professor

Mei/2g // M77

M5 U i: an Affirmative Action/Equal Opportunity Institution 0-12771

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

UUl.‘ 1 82002

 

 

4 wt:

 

Sm :3“ 290:1.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1m WWDGS-p.“

MENTAL CAUSATION AND THE PRAGMATICS OF EXPLANATION

By

Joseph Charles Totherow

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

1999

ABSTRACT

MENTAL CAUSATION AND THE PRAGMATICS OF EXPLANATION

By

Joseph Charles Totherow

Most philosophers of mind presuppose or argue for some
version of physicalism. However, against the backdrop of
physicalism, there exists a fundamental tension between
claims of the systematic irreducibility of the mental and
claims that the mental, in virtue of its irreducible
properties, is capable of causing behavior.

An implication of physicalism is that every behavioral
event can, in principle, be given an exhaustive causal
analysis or explanation through reference to bodily states
and properties. Thus, to claim that irreducible mental
states are causally efficacious is to claim that certain
types of bodily behavior are subject to both an exhaustive
physical explanation and an exhaustive mental explanation.
This presents two problems, the solution to which is the main
concern of this dissertation: First, how are we to make
sense of the unique causal work performed by the mental if
every behavioral event can be given an exhaustive causal
analysis in the physical idiom? Second, if the mental does
indeed perform unique causal work in the production of
behavior, how are we to avoid the problem of systematic

causal overdetermination?

This dissertation approaches these problems through a
pragmatic theory of explanation, presupposing that the causal
work of either mental or physical states can be understood in
terms of the causal explanations that they provide. The
salient feature of the pragmatic theory of explanation is
that explanations are answers to why-questions. Thus, the
unique causal work performed by irreducible mental states can
be understood in terms of the causal explanations they
provide. Specifically, the mental idiom can provide answers
to particular why-questions that the physical idiom.cannot.
Further, causal overdetermination, according to the pragmatic
theory of explanation, is understood as having multiple
acceptable answers to a single why-question. Since the
mental and physical idioms each provide answers to their own
unique why-questions, the problem of causal
overdetermination, on this analysis, does not arise.

Chapter I elaborates on the central problems of mental
causation, explicating them according to the epistemological
and metaphysical motivations of Jaegwon Kim’s Explanatory
Exclusion Principle (EEP). Chapter II develops a preliminary
solution to these problems by responding to Kim's arguments.
Chapter III develops certain details of the pragmatic theory
of explanation, since it plays a vital role to the proposed
solution. Chapter IV develops a more detailed solution to
the problems of mental causation. Chapter V responds to the
solutions offered by the eliminative materialist position and

discusses the advantages of the present proposed solution.

Copyright by
JOSEPH CHARLES TOTHEROW
1999

TABLE OF CONTENTS

LIST OF FIGURES ..................................... ‘ ....... Vi

INTRODUCTION ............................................... 1

CHAPTER I

MENTAL CAUSATION AND EXPLANATORY EXCLUSION ................. 23
The Explanatory Exclusion Principle ................... 23
Mental Causation and the Levels of Analysis ........... 33

CHAPTER II

EXPLANATORY EXCLUSION AND THE PRAGMATICS OF EXPLANATION....36
Explanatory Exclusion and the Non-basic Sciences ...... 36
The Friendly Physicist and the Multiplicity of
Explanations .......................................... 4O
Epistemic Puzzlement and Why-Questions ................ 45
Questions and Contexts ................................ 53
The Epistemic Aims of Explanation ..................... 60

CHAPTER III

THE PRAGMATIC CONCEPTION OF EXPLANATION .................... 71
A Preliminary Solution ................................ 71
Understanding Questions: Alternative Space, Topic
and Relevance Relation ................................ 74
Explanatory Context: On What We Need To Know in
Order To Be Puzzled ................................... 87
Objections and Replies ................................ 92

Objection 1: Bogus Relevance Relations .......... 92
Objection 2: Explanatory Relativism ............ 101

Example and Summary .................................. 108

CHAPTER IV

MENTAL CAUSATION AND EXPLANATION .......................... 117
Problems and Resolutions ............................. 117
Justification of the Relevance Relation .............. 129
Fodor and Ceteris Paribus Laws ....................... 131

CHAPTER V

THE FUTURE OF INTENTIONALITY .............................. 141
The Prospects for Eliminativism ...................... 143
Concluding Remarks: The Value of the Pragmatic
Theory of Explanation ................................ 157

BIBLIOGRAPHY .............................................. 161

LIST OF FIGURES

Introduction
Figure 1 .............................................. 15
Figure 2 .............................................. 19
Chapter I
Figure l .............................................. 31

Chapter II
Figure l .............................................. 52

Chapter III

Figure 1 .............................................. 75
Figure 2 .............................................. 75
Figure 3 .............................................. 88

vi

INTRODUCTION

One of the truly classic problems in philosophy is the
mind-body problem. But while philosophers have long grappled
with the question of the nature of mind and the relation it
bears to the body, the nature of the problem itself has
changed dramatically over the past century.

A long-standing tradition in philosophy maintains not
only a substantive difference between mind and body, but also
a certain primacy of the mind over the body. As far back as
Anaxagoras, we see this primacy of mind. For Anaxagoras, the
origin of order, the first principle which set the physical
universe in regular motion, is Nbus, or Mind. Mind is
therefore metaphysically prior to the physical world.

For Plato, the body is the cage of the immortal soul.
The soul is the true nature of man, while the body serves
only to obscure and blind the soul from its contemplation of
the Forms. The soul, in so far as its proper place is in the
heavens, contemplating the Forms, has a certain metaphysical
priority over the body. The Forms are, after all, ultimately
real.

Not quite so long ago, Descartes insisted that the mind,
as a substance altogether different than physical substance,
is better known than the body. Indeed, Descartes'

- epistemology begins not with an analysis of the knowledge of

physical objects, but with the act of introspection on his

own faculty of thought. Thus, mind is epistemologically
prior to body.

George Berkeley found the material world altogether
expendable, retaining a metaphysics consisting only of minds
and ideas. These few are but representatives of a long and
pervasive view that the mind maintains an epistemological and
metaphysical priority over the body. While there have been
dissenters to this view throughout the history of philosophy,
only within the past fifty to one hundred years has this
conception of the mind fallen into general disrepute.

The reasons for this fall are numerous and complicated.
For instance, by the mid—twentieth century, the claim that
knowledge of the mind could be attained via introspection was
criticized as inherently subjective and unreliable. The
epistemological centrality of the mind would subsequently
fall out of favor.

One might also claim that the emergence of pragmatism
was a salient cause in the fall of the priority of mind.
Rorty, for instance, exercising a fundamental tenant of
pragmatism, rejects what he calls the ”Platonic urge“ to
abstract oneself (i.e. one's mind) away from the physical
world to the contemplation of Absolute Truth: "It is the
impossible attempt to step outside our skins--the traditions,
linguistic and other, within which we do our thinking and
self-criticism--and compare ourselves with something

absolute.” (Rorty, xix)

Certainly, a thorough and detailed analysis of the fall
of the primacy and substantive difference of the mind would
be lengthy. But, in the broadest of strokes, it is
reasonable to claim that the advent and growth of modern
science is the primary cause. Since the revolution away from
Scholastic science to the mechanistic world view of Galileo,
both the scope and the explanatory and predictive power of
modern science has grown exponentially. Along with this
growth emerged a sort of positivism. Whatever could not be
assimilated into this new scientific system must be dismissed
as nonsense. E. A. Burtt gives an example of this
positivistic attitude, drawn from Galileo's own works:

In his discussion on the tides [Galileo] severely

criticizes Kepler for explaining the moon's influence on

the tides in terms that sound like the occult qualities
of the scholastics, judging it better for people 'to

pronounce that wise, ingenious, and modest sentence, 'I

know it not' rather than suffer to escape from.their

mouths and pens all manner of extravagances.‘ (Burtt,
103).

In a world in which a (more or less) mechanistic,
physical science dominates the intellectual landscape, the
mind becomes an ethereal, alien entity. Analytic
philosophers of mind of the twentieth century, by and large,
are different from their predecessors in that a main goal of
the discipline is to scientize the mind; to assimilate the
mind into our scientific world—view. Somehow, our conception
of the mind must find a place in a scientific world. But

what impact would such an assimilation have on the idea that

the mind is epistemologically prior to the body and that,
metaphysically, the mind is of a substance wholly different
from the body?

If the above considerations are true, even in such broad
terms, then the contemporary mind-body problem is
fundamentally different from its traditional ancestor. While
the traditional attempts to understand the nature of mind and
the relation it bears to the body (such as Cartesian dualism)
typically assume (or argue for) a difference in substance as
well as a priority of mind, contemporary mind-body theorists
have all but turned such assumptions on their heads. As
such, contemporary theories of the mind have generally
presupposed some version of physicalism.

While there have been different types of physicalism,
usually varying with respect to strength, we can take Jaegwon
Kimfs notion of "minimal physicalism' to be representative of
the general idea. According to Kim, there are two principles
of minimal physicalism: ”The anti-Cartesian principle“, and
'the mind-body dependence principle":

[The anti-Cartesian principle] There can be no purely

mental beings ... That is, nothing can have a mental
property without having some physical property.

[Mind-body dependence] What mental properties a given
thing has depends on, and is determined by, what
physical properties it has. That is to say, the

psychological character of a thing is wholly determined

by its physical character. (Kim II, 10-11)1

To assimilate the mind is to make it less alien to the
physical world. Eliminating any primacy of the mental over
the physical (the conditions of which are spelled out in the
dependence principle) and precluding any version of substance
dualism (a requirement met by the anti—Cartesian principle)
goes a long way toward meeting this goal. Thus, contemporary
philosophers of mind, if they are to take seriously the need
to scientize the mind, or assimilate the mind into our
scientific world—view, must at least construct their theories
against the backdrop of minimal physicalism.

But what sorts of theories fit the constraints of
minimal physicalism? What, properly speaking, does one
study when one studies the mind? Is there a sense in which
the mind is distinct from the brain, or, given the tenets of
physicalism, must a study of the mind ultimately be a study
of the brain?

Current discussions of the mind—body problem focus
primarily on the relationship between physical states

(primarily brain states) and mental states, e.g. qualia,

 

1Kimdoes invoke a third principle of minimal
physicalism: The supervenience principle . This principle
dictates that ”any two things ... exactly alike in all
physical properties cannot differ in respect of mental
properties." However, supervenience is entailed by the
dependence principle. While Kim distinguishes between
supervenience and dependence (and lists them as separate
principles), I shall regard them as a single principle, given
that the latter entails the former.

beliefs, desires, etc.. Given that physicalism precludes
substance dualism, it seems that the mind can be nothing over
and above the brain. How, then, do we make sense of the
relationship between mental and physical states?

Generally speaking, there are two main schools of
thought on the matter. First, the reductionist argues that
the mind can be systematically reduced to the brain. Thus,
mental kinds can be identified with kinds of brain states.

On such a theory, a necessary component of a scientific study
of the mind is the discovery of the ”bridge laws" by which
mental states can be identified with brain states. Thus, a
successful study of the mind must involve a systematic
reduction of mental states to states of the brain, according
to such bridge laws. For instance, in order to understand
the mental state pain, we must understand its brain
correlate, i.e. the firing of c-fibers.

Therefore, for the reductionist, there is no sense in
which the mind is distinct from the brain. The language of
the mind, i.e. talk of wants, beliefs, qualia, etc.. is
meaningful in that it refers to types of brain states. A
thorough understanding and assimilation of the mind must
therefore proceed according to a study of the brain.

On the other hand, fhnctionaliam argues that there is
a sense in which the mind is distinct from the brain. For
the functionalist, a successful assimilation of the mind into
our scientific world view will not and cannot involve a type—

reduction of mental states to brain states. Thus, the

functionalist must walk a metaphysical tightrope: There is a
sense in which mental states are distinct from the brain, but
they cannot be distinct in virtue of an instantiation in a
second, non-physical substance; for this would clearly
violate the principles of minimal physicalism. But, in what
way can mental states be distinct without violating the
tenets of physicalism?

The functionalist claims that we can understand and
define mental states functionally. In other words, we can
understand a particular mental state in terms of its causal
role in a cognitive system, where that role consists of
relations to input stimuli, to other mental states, and to
output behavior. For instance, the mental state pain can be
understood in terms of its causal relations to input stimuli
(stubbing one’s toe), to other mental states (the desire to
cry out, the desire to protect the wounded area, etc.) and to
output behavior (crying out, rubbing one’s toe, swearing,
etc.). On this view, a study of the mind is a study of the
abstract causal network of mental states, as well as the
varieties of input stimuli and output behavior. Mental
states are therefore, on this view, functional states.

Thus, at this point, two questions arise: First, is a
functional understanding of mental states necessarily
incompatible with the reductionist’s identification of mental
states with brain states? Why is it that we cannot conceive
of pain both functionally and in terms of a certain physical

kind? Second, if they are indeed mutually exclusive, which

analysis of the mental is superior? Are there any arguments
to recommend the reductionist’s analysis over the
functionalist's, or vice versa?

First, the functionalist argues that a functional
understanding of mental states is incompatible with a
reductionist conception of the mind. To conceive of mental
states as functional states is to claim that they are not
systematically identifiable with brain states--that the mind
is somehow distinct from the brain. The most influential
argument for this claim is the multiple realizability
argument.

Suppose Jones is a mechanic and Smith is a mechanically—
illiterate friend. Smith asks Jones to explain what a
carburetor is. Jones has (at least) two choices for her
explanation: Either she explains what a carburetor is
according to a functional analysis or according to a
reductionist analysis. On a functional analysis, Jones would
explain the causal role of the carburetor relative to the
rest of the car: The carburetor responds to the driver's
pressing on the accelerator (external stimulus) by increasing
the amount of oxygen and fuel sent to the engine cylinders
(other internal states of the system), which turns the gears
of the transmission more quickly, thereby increasing the rate
of tire rotation and accelerating the automobile (behavioral
output). On a reductionist analysis, on the other hand,

Jones would explain the concept of the carburetor in terms of

its physical constitution, e.g. the material of which it is
made, namely a certain type of metal alloy.

Again, the functionalist claims that the functional
conception of the carburetor is incompatible with the
reductionist conception. This is because the carburetor,
functionally defined, can be realized in a number of
different physical types. While Jones’ carburetor happens to
be made of a particular metal alloy, it could very well have
been made of a type of hardened fiberglass and still
performed its causal role, in relation to the rest of the
automotive system. Thus, since the carburetor, functionally
conceived, is multiply realizable in an indefinite number of
physical types, the reductionist analysis of the carburetor
as identical to a particular physical type is incompatible
with the functional analysis.

But while the functionalist and reductionist analyses of
the carburetor are incompatible, it is not yet clear whether
one is superior to the other. In order to distinguish the
superior of the two analyses, we need to ascertain which
understanding more closely resembles our common conception of
the carburetor. Would a close analysis of our ordinary,
shared definition of the carburetor more closely resemble the
reductionist conception or that of the functionalist?

Again, the functionalist turns to the multiple
realizability argument. Let us return to Jones and her
automobile. Jones’ carburetor has broken down and must be

replaced. However, while Jones has the moldings for shaping

a new carburetor, she does not have a supply of the metal
alloy she needs to fashion a new metal carburetor.

Instead, she uses her moldings to fashion a device out
of a super-hardened fiberglass. Once cooled, she installs
this device in her car where the carburetor used to be. Once
installed, Jones turns the key and the car starts and
continues to run. Functionally speaking, this new fiberglass
piece of machinery bears the same causal relations to the
rest of the car as did her old carburetor. In other words,
the car with the fiberglass mechanism is functionally
isomorphic to the car with the metal carburetor. The
question is, is this new fiberglass mechanism, strictly
speaking, a carburetor? For instance, would the community of
mechanics recognize Jones’ machine as a carburetor?

The functionalist predicts (rightly, I think) that the
new mechanism, regardless of its material, will be considered
by anyone who understands automotive systems to be a
carburetor. The functionalist argues that the functional
definition of the carburetor is not only incompatible with
the reductionist’s, it is superior, because it more closely
resembles our common idea or conception of what a carburetor
is. The fact that a mechanic would define Jones’ new
fiberglass mechanism as a carburetor illustrates that our
common definition of a carburetor is a functional definition.

But what of mental states? First, are mental states
multiply realizable? Second, would a functional definition

of mental states be superior to the reductionist definition?

10

The multiple realizability of mental states can be
established by a thought experiment similar to our carburetor
story.

Let us take a common example: An alien, whose chemical
composition is based in silicon (whereas humans’ is based in
carbon), lands its spacecraft just outside of Lansing.

First, while the alien is composed of a radically different
physical type than humans, it is perfectly plausible that it
maintains internal states that, functionally defined, mirror

our own. As Paul Churchland says,

The chemistry and even the physical structure of the
alien's brain would have to be systematically different
from ours. But even so, that alien brain could well
sustain a functional economy of internal states whose
mutual relations parallel perfectly the mutual relations
that define our own mental states. (Paul Churchland, 36-
37)

Thus, it is quite plausible that such an alien could admit of

internal states that mirror (or are functionally isomorphic

to) our own.

Therefore, an alien, quite unlike us in physical makeup,
could plausibly maintain internal states that admit of
functional definition. So, for each such state, there is a
functional definition and a reductionist definition. Again,
the question arises, which definition is superior?

Let us say that the alien begins to advance towards a
group of people who have gathered to watch the ship’s

landing. As it advances, it extends an appendage towards the

11

group. Panicked, a member of the group clubs the appendage
with a nearby branch. Once struck, the alien shrieks,
retracts the appendage, shields it from further damage and
moves quickly back into the spaceship. Would we want to
claim that the alien felt pain? The functionalist reasonably
predicts that, were such a situation to occur, the internal
state pain would be attributed to the alien, even though it
is made of a radically different type of substance. The
alien felt pain and wished to protect the wounded area and
avoid any further infliction of injury. The reductionist, on
the other hand, cannot say that the alien felt pain because
it has no c—fibers.

As was the case with the carburetor example, the
functional definition of pain is superior to the reductionist
definition because the former more closely mirrors the common
use of the concept. The functionalist predicts that, if such
a being were to land on earth, we would, as a community of
speakers who maintain a mental idiom, attribute mental states
to it. Thus, not only are the functionalist and reductionist
analyses of the mental incompatible, the functionalist
analysis of mental states is superior in that it more
accurately captures our common use of mental concepts.

Hence, the functionalist is able to walk the
metaphysical tightrope: Functionalism is able to maintain
the irreducibility of the mind by denying type—
identifications of mental states with physical states. Thus,

the mind is distinct from the body, but not in virtue of

12

being a second, immaterial substance. On a functional
conception of the mind, mental states must be realized in a
physical system. But it does not follow that the mind simply
is the physical system that realizes it. Thus, humans are
physical beings, but with minds that are not systematically
reducible to their brains.

The primary virtue of functionalism is its ability to
make sense of mental states as distinct from brain states
without violating the principles of minimal physicalism"
Thus, in terms of the question of the assimilation of the
mind into our scientific world view, it appears that
functionalism is superior to Cartesian dualism. But while it
avoids the problem of invoking a second immaterial substance,
functionalism, on its own, does not escape all the
difficulties that afflict Cartesian dualism“

A primary example is the problem of mental causation.
The mind, it is usually believed, is capable of having a
causal impact on certain types of human (and possibly animal)
behavior. At bedtime, a child may ask a parent to leave the
bedroom light on because he is afraid of the dark. If one
cries, it is because he is sad. If I go to the store, I do
so because I believe I am out of milk and I believe that,
come tomorrow, I will desire that I have milk with my
breakfast. Fear, sadness, beliefs and desires are mental
states and it is not uncommon to believe that such mental

states affect behavior.

13

Further, the causal efficacy of the mind is a necessary
presupposition in many attempts to solve the problem of free
will. C.A. Campbell, for instance, in his notion of
phenomenological analysis, presupposes that deliberation, as
a subjective state of the mind, has a causal impact on
subsequent behaviorz.

Finally, it would seem obvious that claims of moral
responsibility for one's actions, as they are judged in
courts of law, must necessarily involve reference to a
defendant's state of mind. This presupposes that one’s state
of mind can impact subsequent behavior. Thus, in the
development of a theory of the mind-body relation which makes
the mind assimilable into our scientific world-view, a theory
of the irreducibility of mental phenomena can only be part of
the story. An account of the causal efficacy (or the lack
thereof) of the mental must also be provided. It is with the
question of mental causation that functionalism finds itself
in a bind similar to that of its Cartesian ancestor.

Dualismfis problem with mental causation seems
intractable and historically is one of the difficulties which
ultimately led to its current unpopularity. With dualism,
one finds two disparate types of substances which share no
common attributes. On such a theory of the mind-body

relation, how would a theory of mental causation work?

 

2 See Cambell. W, (London: George
Allen & Unwin) 1957, pp. 158-179.

14

For dualism, physical states, such as stubbing one's toe
(P1), cause mental states, such as the feeling of pain (M1).
Further, mental states, such as the desire to ease the pain
and the belief that rubbing the toe will ease the pain (M2),
could cause physical actions, such as rubbing the wounded toe

(P2).

541 '—-“* h42

/ causal gap \

p1 """"""""""""""" IP2

Figure 1

But certainly, on the dualist account, this would run
contrary to the causal closure of the physical universe and
to the law of the conservation of energy. If this
characterization of the toe-stubbing incident is correct,
then there is a ”causal gap" at the physical level between
the stubbing and the rubbing of the toe. On a physical
analysis, new energy or work is introduced into the physical
universe, since the rubbing of the toe has no direct physical
cause. This certainly runs counter to the laws of the
physical sciences.

The causal gap problem illustrates the difficulty of
assimilating Cartesian dualism into our scientific world-
view. Indeed, it is reasonable to claim that this problem is
a primary motivation for the "anti-Cartesian principle” of

minimal physicalism. However, banishing the second substance

15

does not automatically resolve the problems of mental
causation. As will be shown, functionalism, even without
invoking a second substance, is afflicted with difficulties
strikingly similar to the causal problems of dualismi It
too finds difficulty assimilating into our scientific world-
View.

Again, the primary virtue of functionalism is its.
ability to construct an understanding of the irreducibility
of the mental without invoking a second substance. Thus,
while the mind cannot be systematically reduced to a physical
system (in our case, the brain), it cannot exist unless
instantiated in a physical system. The functionalist is able
to walk this tightrope by invoking the type/token
distinction.

A systematic reduction of mental states to physical
states would involve a type-identification of mental and
physical states. Bridge laws produce this sort of type-
identification. The functionalist analysis precludes such a
type-identification, but insists that each particular (token)
mental state must be instantiated in a physical medium”
Mental states are token-identical to physical states, but not
type-identical.

In virtue of the type/token distinction, functionalism
is able to avoid the ”causal gap" problem of substance
dualism” Since mental states are token—identical with
physical states, maintaining a causal relationship between

mental states and physical states does not imply a causal gap

16

at the physical level, at least at first blush. According to
the functionalist, a mental state is causally efficacious
only if it is realized in a physical system. Roughly, mental
states inherit their causal powers from the physical states
in which they are realized. Maintaining the causal efficacy
of the mental, on this view, does not preclude the causal
continuity between physical events. However, even if the
causal gap problem is avoided, functionalism is plagued with
its own unique problem with mental causation.

The new problem of mental causation is given voice by
Jaegwon Kim» Kim argues that the functionalist's attempt to
account for mental causation turns on the phrase “determined
by but not identical to.“ (Kim I, 355) To put this phrase in
context, the causal powers of mental states are "determined
by but not identical to" the causal powers of physical
states. So as to avoid the Cartesian problem of mental
causation, mental states must inherit their causal efficacy
from physical states. But in order to retain the
irreducibility, as well as the causal efficacy of the mental,
the causal potency of mental states cannot be identified with
that of the physical.

One could easily imagine a functionalist version of the
"determined by but not identical to” (hereafter, DBNI)
thesis. Mental states must be realized in a physical medium
(via token—identity). Thus, the causal efficacy of mental
states is ”determined by" the causal powers of the

corresponding physical states. However, given that those

17

mental states (including their causal properties) can be
instantiated in any number of physical types, neither the
mental states nor their causal properties can be "identified
with” any particular type of physical state. But, as Kim
shows, such an account of mental causation leads to serious
difficulties3.

Suppose mental state M is realized in physical state P.
Further, M is the cause of a second physical state, P*. More
precisely, M is a sufficient cause of P*. According to the
DBNI thesis, M inherits its causal powers from its physical
realization, namely P. In other words, M is able to cause P*
because M is realized in P. But, also according to the DBNI
thesis, the causal properties of M are not identical to those
of P. Thus, M causes P* in virtue of novel causal properties
belonging to M.

Now, it would be highly problematic to claim that, while
M is a sufficient cause of P*, P is not also a sufficient
cause of P*. Indeed, such a claim would run counter to the
DBNI thesis. Again, the thesis states that the causal
efficacy of a mental state is determined by its physical
realization. But if M were a sufficient cause of P* and P
were not a sufficient cause of P*, then the causal efficacy

of M would not be wholly determined by P, since M would have

 

3 For a full explication of the following argument, see
Kim's “The Non-Reductivist's Troubles with Mental Causation“,

sec. VI. in W: (1993).

18

a causal potency that P lacks, namely, being sufficient for
P*. This would render the DBNI thesis false.

Thus, it follows from the DBNI thesis that, where M and
P are token-identical (but not type-identical) and P* is a
state subsequent to M and P, M is a sufficient cause of P*
if and only if P is a sufficient cause of P* (see fig. 2).
As Kim states it, "since P is a realization base for M, it is
sufficient for M, and it follows that P is sufficient, as a

matter of law, for P*" (Kim I, 354).

Rd

i\

P——> p*

Figure 2

Here we come to the heart of the problem as Kim sees it.
The irreducibility of M to P, in conjunction with maintaining

the causal efficacy of M, leads to causal overdetermination:

We would be allowing two distinct sufficient causes,
simultaneous with each other, of a single event. This
makes the situation look like one of causal
overdetermination, which is absurd. ... Given that P
is a sufficient physical cause of P*, how could M also
be a cause, a sufficient one at that, of P*? What
causal work is left over for M, or any other mental
property, to do? (ibid. 354)

Hence, the similarity between Cartesian dualism and

functionalism becomes readily apparent. Both theories appear

19

to resist the assimilation of causally efficacious mental
states into our scientific world view. On the one hand,
dualism violates the principles of the conservation of energy
and the causal closure of the physical universe. On the
other hand, the functionalist analysis of mental causation
apparently involves causal overdetermination.

But what's wrong with causal overdetermination?
Certainly, there is the occasional causally overdetermined
event, such as a firing squad in which a man is shot in the
heart with two bullets at the same instant, each bullet being
sufficient for his death. Why should this pose such a
problem for mental causation?

The problem is, claims of mental causation (either from
dualism or functionalism) do not involve the occasional,
anomalous case of overdetermination. Rather, mental
causation involves systematic overdetermination, where every
intentional or purposive behavioral event is causally
overdetermined. Systematic overdetermination occurs when our
theories of the world (in the present case, our theories of
the causal efficacy of the mind and that of the body) entail
overdetermination.

Kim claims that at best, systematic causal
overdetermination is "very odd" and at worst, "absurd." He
acknowledges that causal overdetermination does occur. But
it is difficult to give a causal explanation in such cases.
Indeed, if two distinct, causally unrelated events M and P

are each sufficient for the production of event P*, then it

20

is unclear how we should formulate a causal explanation of
P*. Both M and P render each other unnecessary in the
explanation of P*. So in general, genuine cases of
overdetermination make explanation difficult.

The problem for mental causation goes deeper than merely
presenting difficulties for explanation. As Kim says, “it is
at best extremely odd to think that each and every bit of
action we perform is overdetermined in virtue of having two
distinct sufficient causes." (Kim I, 247) Cases of
overdetermination, generally, are oddities-— explanatory
anomalies, if you will. It would be ”extremely odd" to have
our view of the world set up to claim that every bit of
intentional behavior (which encompasses quite a bit of bodily
movement) is causally overdetermined. Such a claim would be
akin to having a thorough understanding of the buildup and
discharge of static electricity in the atmosphere as an
explanation for lightning, as well as the theory that each
lightning bolt is hurled to earth by Zeus because he is
angry. (Whether or not common-sense psychology is on a par
with Greek mythology is a matter of some debate in
contemporary mind theory.) It appears that this is what Kim
means by calling systematic overdetermination “absurd.“

It seems that accounting for the irreducibility of the
mental without invoking a second substance does not solve all
the difficulties packed into the mind-body problem. Indeed,

against the backdrop of physicalism, the problem of mental

21

causation is as much a problem for functionalism as it is for
Cartesian dualism.

It is the goal of this dissertation to present an
account of mental causation which, within the conditions of
physicalism, is reconcilable with the irreducibility of the
mental. Such an account requires, first, a closer analysis
of Kimfls causal-overdetermination argument. This argument,
as it applies to mental causation, is found in the more
general argument for, what Kim calls, the “explanatory
exclusion principle" (hereafter, EEP). Thus, the resolution
of the functionalist problem of mental causation will begin
with a response to Kim's general arguments for explanatory

exclusion.

22

Chapter I: Mental Causation and Explanatory Exclusion

1. The Explanatory Exclusion Principle

A rather stock example, first given by Norman Malcom
(1968) and subsequently employed by Kim himself, will help
facilitate the discussion of the EEP: As Smith walks to her
car, a gust of wind blows her hat to the roof of her garage.
She ponders the situation briefly, then finds a step-ladder
in the garage, climbs it and retrieves her hat. While such a
task may seem rather mundane, the cognitive skills involved
in Smith's solving the problem are actually rather
impressive. While this behavioral series is complex, for the
sake of argument, the series will be considered as a single
event, E. E, like other non—quantum physical events, is
subject to a causal explanation. But what is that
explanation? Where would one look for such an explanation?

E appears subject to at least two types of explanation.
First, one might give an explanation X which is formulated in
the intentional idiom: Smith wanted to retrieve her hat; she
believed that she could not reach the hat without aid; she
desired to have such an aid that would assist her in the
retrieval of her hat; she believed there was a ladder in the
garage, etc. Second, an explanation Y, couched in the
language of neurophysiology, can be offered: Certain
external stimuli caused a particular series of neurons to

fire in Smith's brain, which, in turn, caused a certain

23

electro-chemical reaction, which sent various impulses down
her spinal cord to her limbs, which caused her muscles to
contract in a certain order, etc.

Kim argues that such a situation, where there are
multiple explanations of a particular event,.may be
untenable. The conditions under which such a situation is
untenable are expressed in what Kim calls “the explanatory
exclusion principle“: “No event can be given more than one
complete and independent explanation' (Kim I, 239). As
applied to the example, X and Y cannot both be complete and
independent explanations of E. In order to begin to
understand the meaning and importance of the explanatory
exclusion principle, it is necessary to understand Kim's
conception of the nature of explanation in general.

Kim makes two explicit presuppositions about
explanation. First, he assumes a strong, symmetrical
relation between causation and explanation, such that X is a
cause of Y if and only if X is an explanation of Y4. A
second, related presupposition is what Kim calls ”explanatory
realism“: “That a causal explanation of E in terms of C is a
'correct explanation' only if C is in reality a cause of E,
can be called an 'objectivist' or 'realist' conception of

explanation“ (Kim, 256). Thus, an explanation of an event E

 

4 While this construal of causation, explanation and
their relationship may seem overly strong, I will not dispute
this premise of Kim's argument, as I think that he is, in
general, correct.

24

is correct (true, accurate, or what have you) in virtue of
picking out the actual cause which brought about event E.

At the heart of the exclusion of multiple explanations
are the criteria of completeness and independence.

Therefore, in order to understand the EEP, we must understand
the meaning of these criteria. Curiously, Kim never gives an
explicit definition of either concept. Instead, an almost
purely negative definition arises from examples or cases Kim
gives in which explanations are either dependent, incomplete
or both. While this poses some difficulty in divining Kim's
exact meaning of these criteria, we should be able to
generate a loose but positive conception of completeness and
independence which will be adequate for our purposes.

First, in situations where X and Y only seem to be
distinct, but are actually type identical, neither X nor Y is
an independent explanation of E. Such would be the case if,
for example, there existed bridge-laws that allowed
scientists to type-reduce biological concepts to those of
chemistry; Explanations which invoke biological concepts
would be dependent on chemical explanations. Further, if the
events described by x supervene on those specified by Y, then
X is not independent of Y as an explanation. Also, where X
and Y are both links in the same causal chain, neither X nor
Y is an independent explanation.

Regarding completeness, Kim argues that if X is a
"proper part" of Y (i.e. if X picks out a proper part of the

causal process specified in Y), then x is neither complete

25

nor independent. Finally, Kim does give at least a partial
positive definition of completeness: I'In one sense of
complete, ... a complete explanation specifies a sufficient
set of causal conditions for the explanandum“ (Kim, 251).

These clues generate a fairly definite conception of
completeness and independence. Independence appears to be a
relational term: To say that explanation X is independent is
to say that X is independent of other possible explanations,
i.e. X neither collapses into (via an identity relationship)
nor supervenes over another explanation Y, nor is it a proper
part of another explanation. Thus, for X to be independent
as an explanation, X must pick out a causal process which
relies upon no other causal process for the production of E.

And, for X to be a complete explanation, X must
exhaustively specify each component of a particular causal
series, having E as its result, which makes that causal
series a sufficient condition for E. What Kim seems to have
in mind by these criteria, then, is that X is a complete and
independent explanation of E if and only if X exhaustively
picks out a causal process which produces E (i.e. X specifies
events X1 ... Xn which are jointly sufficient for the
production of E), without bearing a dependence relation
(supervenience, identity, and the like) to any other causal
process.

Now that the concepts of completeness and independence
are more or less clear, why can there not be more than one

complete and independent explanation of a particular event?

26

Recall Smith's hat-retrieval behavior. At least two
explanations of her behavior are forthcoming—- X, couched in
the intentional idiom and Y, expressed in neurophysiological
terms. If both X and Y are complete and independent
explanations of E, then each refers to a causal series which
is sufficient for the production of E. Further, neither
explanation can somehow be reduced to the other, via
supervenience or identity. Roughly, there are two distinct
causal processes, both of which are responsible for the
occurrence of E.

According to Kim, this situation is untenable. He first
objects to the possibility of such a situation on
metaphysical grounds. The immediate problem, again, is
causal overdetermination, which, as Kim says, is a
”metaphysically unstable“ situation. Recall that while
there are instances of causal overdetermination, they are
explanatory anomalies. They are situations which make
explanation difficult:

A man is shot dead by two assassins whose bullets hit

him at the same time; or a building catches fire

because of a short circuit in the faulty wiring and a

bolt of lightning that hits the building at the same

instant. It isn't obvious in cases like these just how

we should formulate an explanation of why or how the
overdetermined event came about. (Kim II, 252)

The problem of overdetermination (the "metaphysical
instability") emerges when we conceive of the world as
systematically overdetermined. Mental causation illustrates

this problem nicely. A conception of mental causation which

27

maintains the irreducibility of the mental violates the EEP
on metaphysical grounds, since it conceives of the world as
systematically overdetermined. It makes every piece of
intentional bodily behavior an explanatory oddity or anomaly.
It appears that the "absurdity" of systematic
overdetermination is the basis for this metaphysical
instability. It is bad enough that the occasional bona fide
case of causal overdetermination baffles our attempts at
explanation, since each explanation renders the other (or
others) superfluous, but to set up our view of the world such
that all cases of purposive action are causally
overdetermined is absurd.

Kim also claims that having multiple complete and
independent explanations of an event is an epistemically
unstable or counterproductive situation: "When we look for
an explanation of an event, we are typically in a state of
puzzlement, a kind of epistemic predicament. A successful
explanation will get us out of this state“ (ibid. 254).

Thus, a fundamental epistemic principle of explanation
is that a satisfactory explanation must alleviate puzzlement.
It is Clear that if either X or Y were exclusive explanations
of E, then our puzzlement would vanish. However, where we
have both X and Y as complete and independent explanations,
our puzzlement is not alleviated: Our old epistemic
predicament of needing an explanation for E is compounded by
the new epistemic predicament of needing to know which

explanation is correct. As Kim says, ”too many explanations

28

will put us right back into a similar epistemic predicament“
(ibid. 254). Having multiple competing explanations is
counterproductive to the epistemic goals of explanation, i.e.
removing this type of epistemic puzzlement.

So there are (at least) two motivations for the
explanatory exclusion principle: A metaphysical quandary,
where we are faced with the absurdity of systematic causal
overdetermination, and an epistemic quandary, where having
multiple explanations is counterproductive to the epistemic
goals of explanation. Based on these difficulties, Kim
claims that we cannot maintain multiple complete and
independent explanations of a particular event. All but one
explanation-candidate must be either incomplete, dependent,
or both.

How does the EEP apply to Smith's hat—retrieval
behavior? Since the intentional explanation X and the
neurophysiological explanation Y cannot both be complete and
independent, we must exclude either one or the other (or
both) as a complete and independent explanation of E.
Certainly, it would be exceedingly problematic to maintain
that the physical idiom cannot provide a complete
explanation. Further, it would be even more difficult to
maintain a dependence of physical explanations upon mental
explanations.

In the first place, Kim cites Churchland's arguments5

 

5 From "Eliminative Materialism and the Propositional
Attitudes“, Jburnal of Philosophy 78 (1981), p. 73.

29

regarding the explanatory failure of the intentional idiom:
”As examples of central and important mental phenomena that
remain largely or wholly mysterious within the framework of
[folk psychology], consider ... the faculty of creative
imagination ... the common ability to catch a fly ball

the miracle of memory...“ (Kim I, 262).

Churchland cites further examples of the intentional
language's (or, as he says, ”folk psychology's”) inability to
explain various types of psychological phenomena: Within
folk psychology, one cannot explain the need for rest, the
different levels of intelligence displayed in the human
population, nor are we able to explain or prescribe cures for
mental illness (Churchland, 45-46). Such examples prove
difficult for those who claim that physical explanations
depend on mental explanations, since in these cases, there is
no mental explanation on which the physical explanation might
depend5.

Explanatory failure of the mental idiom is not the only
problem with maintaining a dependence of the physical on the
mental. Indeed, maintaining such a dependence would amount

to a return to Descartes' “causal gap" problem. Recall the

 

5 These examples seem to preclude a dependence of
physical explanation on mental explanation. But while Kim
uses these examples to argue for a dependence of the mental
on the physical, Churchland uses them to advance his theory
of eliminative materialism. This theory predicts that
intentionality, as a means of describing human cognition,
will ultimately be replaced by a fully—developed theory of
neurophysiology. While the immediate concern is to address
Kim’s arguments for the exclusion of mental explanation,
Churchland’s arguments shall be addressed in chapter V.

30

'toe-stubbing" example. Stubbing one's toe causes (at least)
pain and the desire to alleviate that pain, which, in turn,
causes crying, swearing and the rubbing of the wounded toe.
Both the toe—stubbing and the subsequent behavior are
physical events. According to the neurophysiologist, the
stubbing causes certain impulses to be sent to the brain,
which in turn, causes the subsequent physical behavior.

To maintain the dependence of this explanation on the
mental explanation entails that the toe-stubbing and the
subsequent brain events are causally connected only in virtue
of the brain events' dependence on the mental events caused
by the toe-stubbing (i.e. pain and the desire to end the

pain):

hdl'———’ L42
P1 ----- P2 ------ P3 ----- P4
Figure 1

In such a situation, where the physical explanation is

dependent upon the complete and independent mental

explanation, the ”real" causal series is P1 -> M1 -> M2 —>
P4. If the physical series P1, P2, P3, P4 (where P1 is the
toe—stubbing and P4 is the swearing and rubbing) is to be a
causal series, it must be a causal series in virtue of its
dependence on the former. Such a dependence entails causal

gaps between physical events. This brings us back to the

31

”causal gap" problem of Cartesian dualism. Thus, even
without the problem of explanatory failure, maintaining the
explanatory dependence of the physical on the mental runs
counter to our scientific world-view in exactly the same
manner in which substance dualism does. Unless one wishes to
return to the problems of Cartesian dualism, maintaining the
dependence of physical explanations on mental explanations is
out of the question.

The most satisfactory application of the EEP, then, is
to maintain a dependence of mental explanations on physical
explanations. There are at least two accounts of such a
dependence relation: Either mental states are dependent in
virtue of being type-identical with physical states or in
virtue of an epiphenomenal relation with physical states.
The first option is ruled out by the irreducibility of the
mental. Therefore, if Kim's arguments about explanatory
exclusion are correct, the most adequate anti-reductionist
conception of mental causation is an epiphenomenal relation,
in which mental states are only causally efficacious in
virtue of their physical instantiations. As far as mental
causation is concerned, irreducible mental properties there
may be, but they are epiphenomena and can have no direct

impact on physical behavior.

32

6. Mental Causation and the Levels of Analysis

But, one might argue, doesn't this stand to reason?
Causality--the real ”pushes and pulls"——takes place at the
physical level. Mental states, by functional definition, are
not physical (at least at the level of types). Thus, mental
states can only be causally efficacious if they are
instantiated in a physical body.

The functionalist, in keeping with the conditions of
physicalism, concedes that the mind is causally efficacious
only if physically realized, but it does not logically follow
that the mind is causally efficacious in virtue of that
physical realization. Recall that, for the functionalist,
mental states are type-irreducible to physical states since
the former can be realized in a number of physical types.
Thus, no functional property is, properly speaking, a
property of the physical medium, since that functional
property can be realized in other physical types.

However, if the EEP is correct and the mental is
causally efficacious only in virtue of its physical
realization, then causal properties are not proper to
functional states. Causal properties are analyzed at the
physical level. The intentional language, therefore, can
figure only indirectly in questions of causality. Therefore,
if the EEP is correct, then if we maintain the irreducibility
of the.mental, Via functionalism, the mental is stripped of

its causal efficacy. Not only do mental states supervene

33

over physical states, but causal properties can only be
located at the subvenient (i.e. physical) level. Properly
speaking, there are no causal mental properties.

So while philosophers of mind have made great strides in
maintaining the irreducibility of the mental without
violating the principles of physicalism, it appears that
there is no room for both an irreducible and a causally
efficacious conception of the mind within a physicalist
framework. This brings us to the central issue: Is it
possible to construct a positive conception of mental
causation which can be reconciled with the irreducibility of
the mind, without violating the principles of physicalism?
If such a project were successful, the more general project
of assimilating the mind into our scientific world—view will
have taken a significant step forward. It is this project
that is the main concern of this dissertation.

It is important to note that the tension between mental
irreducibility and mental causality (a tension which
manifests itself relative to the conditions of physicalism)
is representative of a larger problem. As Fodor points out
in “Making Mind Matter More"7, to question the causal
efficacy of the mental is to question the causal efficacy of

the properties of the special sciences in general:

 

7 Fromwmmmmmmsaxs (1990), pp.
137-160.

34

If there's a case for epiphenomenalism in respect of
psychological properties, then there is the same case
for epiphenomenalism in respect of all the nonphysical
properties mentioned in theories in the special
sciences. (140)
Thus, there is more at stake here than simply a conception of
mental causation. If the constraints of physicalism apply to
mental states and properties, then it must apply to the
properties of all the special sciences. If the EEP is truly
a reasonable constraint on explanation in general, then it
must be applied across the board to all the special sciences.
It is hoped that the present proposal for the reconciliation
of the irreducibility and the causal efficacy of the mental

will provide a model for understanding the causal efficacy of

the states or entities of other special sciences.

35

Chapter II: Explanatory Exclusion and the Pragmatics

of Explanation

1. Explanatory Exclusion and the Non—basic Sciences

The scope of the explanatory exclusion principle goes
well beyond considerations of the causal properties of mental
states. There are a number of non-physical sciences (i.e.
those sciences which maintain a vocabulary and ontology
beyond that invoked by physics) that posit the existence of
states and objects which are instantiated in, but resist
systematic reduction to, physically described states and
objects. At first blush, the EEP’s broad scope might seem to
present serious problems for a wide variety of non-physical
or special sciences. Ironically, however, it is the scope of
its application that makes the EEP highly dubious.

If considerations of the application of the explanatory
exclusion principle are limited to mental causation, then
Kimfls arguments might appear to make sense. It does not
require a tremendous stretch of one's intuitions or
imagination to exclude mental states as causally efficacious,
outside of their physical realizations. After all, the mind
is an ethereal, mysterious thing, even without the invocation
of a second substance.

But if the EEP is applied to other non-physical sciences
it loses a good deal of its original luster. Certainly,

there are sciences which resist systematic reduction--usually

36

by some multiple-realizability argument——but whose states and
properties are clearly causally efficacious. Michael
Scriven, in ”Causation as Explanation“, gives examples of the
indispensability of causation in the economic sciences: "The
debate over the cause of economic events such as the present
inflation/depression is one in which the eminent economists
involved will not accept translations of the dispute into
non-causal (e.g. correlation) terms.“ (Scriven, 4)

Jerry Fodor, in "Making Mind Matter More,“ is quite
explicit about the implications of stripping the mental of
its unique causal properties: "If there's a case for
epiphenomenalism in respect of psychological properties, then
there is the same case for epiphenomenalism in respect of all
the nonphysical properties mentioned in theories of the
special sciences.“ (Fodor, 140) Fodor gives several examples
to illustrate this point, including the science of geology,
which maintains the property of 'mountainhood':

Consider, for example, the property of being a mountain;

and suppose (what is surely plausible) that being a

mountain isn't a physical property ... Untutored

intuition might suggest that many of the effects of
mountains are attributable to their being mountains ...

It is because Mount Everest is a mountain that Mount

Everest has glaciers on its top; ... that it casts such
a long shadow ... and so on. (139)

But geology appears to resist systematic reduction to the
physical sciences; for if any non-physical properties are to
be considered multiply realizable, 'mountainhood' must be

counted in their number. Without such a systematic

37

reduction, however, the attribution of causal efficacy to the
states and properties of geology must run counter to the
explanatory exclusion principle. Thus, for reasons sindlar
to those given for the exclusion of mental explanation, the
science of geology must be understood as causally inert.

A second example given by Fodor is the airfoil, a
concept contained in the science of aerodynamics. Again, the
claim is that the airfoil is multiply realizable. Thus, the
properties of the airfoil cannot be type-reduced to
properties described in the physical idiom. However, it
would seem absurd to claim that the properties of being an
airfoil are causally inert:

Typically, airfoils generate lift in a direction, and in

amounts, that is determined by their geometry, their

rigidity, and many, many details of their relations to
the (liquid or gaseous) medium through which they move.

The basic ideas is that lift is propagated at right

angles to the surface of the airfoil along which the

medium flows fastest, and is proportional to the
relative velocity of the flow. (139)

Ceteris paribus laws regarding the propagation of lift by an
airfoil go a long way towards explaining why planes are able
to fly and why sailboats are able to move through the water.
Such explanations, it would seem, are causal. But the
application of the EEP precludes such causal explanations.
For reasons similar to those given in chapter I for the
exclusion of mental explanations, we must exclude aerodynamic
explanations: They are subject to explanatory failure,

(their laws are, after all, ceteris paribus laws) and their

38

scope is more limited than that of the physical sciences.
However, to apply the EEP in this case and preclude the
causal efficacy of airfoils is, according to Fodor, "quite
mad": ”Airplanes fall down when you take their wings off;
and sailboats come to a stop when you take down their sails.”
(140)

In the search for causal analyses of the movement of
sailboats or the flight of airplanes, causal claims and
explanations which refer to the properties of airfoils would
be appropriate. As Fodor argues, 'if that isn’t the right
explanation, what keeps the plane up? If that is the right
explanation, how could it be that being an airfoil is
causally inert?" (140) Therefore, as with the case of
geology, an application of the explanatory exclusion
principle to the properties of airfoils yields absurd
results.

Further examples could be given, but the point is clear:
If we accept the explanatory exclusion principle as a
'plausible constraint on explanation in general” (Kim I,
239), then none of the non-basic sciences are capable of
making causal claims or generating causal explanations. As
far as Fodor is concerned, such implications make the
arguments for the epiphenomenalism of the mental at best,

suspect, and at worst, absurd:

There are lots ... of examples where, on the one hand,
considerations like multiple realizability make it
implausible that a certain property is expressible in

39

the physical vocabulary; and, on the other hand, claims
for the causal inertness of the property appear to be
wildly implausible. (139)

Though it's true that claims for the epiphenomenality of
mountainhood and airfoilhood and, in general, of any
nonphysical-property-you—like-hood, will follow from the
same sorts of arguments that imply claims for the
epiphenomenality of beliefhood and desirehood, it's also
true that such claims are prima facie absurd. (141)

The far-reaching implications of the EEP appear to be its
undoing. Examples such as those given by Scriven and Fodor
show that the explanatory exclusion principle is not, as Kim
maintains, "a plausible constraint on explanations in
general." Rather, it is simply too strong a condition.

However, the problems and considerations which motivate
the EEP remain steadfast: How are Kim's metaphysical and
epistemological quandaries regarding multiple explanations to
be addressed? More broadly, if the EEP is incorrect, how are
cases of multiple complete and independent explanations to be
understood? How are we to make sense of the relationship
between such explanations?

2. The Friendly Physicist and the Multiplicity of
Explanations

In order to answer these questions and, in general,
understand the problems that make the EEP untenable, it is
necessary to begin with a closer analysis of an example
(similar to those given by Fodor and Scriven) which appears
to run afoul of the EEP. By now, most philosophers have

heard Wesley Salmon's story of the "friendly physicist." A

40

physicist on an airplane, waiting to take off, makes a
friendly wager with a young boy seated nearby: He contends
that the boy's helium-filled balloon, on takeoff, will move
toward the front of the plane. The boy (along with the other
passengers) believes that, just as a passenger feels the
inertia pulling him toward the back of the plane, the balloon
will also be “pulled back." Much to their surprise, the
physicist wins the bet.

The boy and the other passengers are in Kim's "epistemic
quandary." Given that their initial expectation that the
balloon would float toward the back of the plane was foiled,
they are now in need of an explanation as to why the balloon
acted as it did. Salmon argues that there are at least two
legitimate explanations:

First, one can tell a story about the behavior of the

molecules that made up the air in the cabin, explaining

how the rear wall collided with the nearby molecules

when it began its forward motion, thus creating a

pressure gradient from back to front of the cabin. This

pressure gradient imposed an unbalanced force on the
back side of the balloon, causing it to move forward
with respect to the walls of the cabin. Second, one can
cite an extremely general physical principle, Einstein's

‘principle of equivalence, according to which an

acceleration is physically equivalent to a gravitational

field. Since helium—filled balloons tend to rise in the
atmosphere in the earth's gravitational field, they will
move forward when the airplane accelerates, reacting

just as they would if a gravitational field were
suddenly placed behind the rear wall. (Salmon, 183)

Salmon refers to the explanation based in molecular mechanics

as explanationl and that based on the principle of

equivalence as explanationz. As Salmon claims, "it is my

41

present conviction that both of these explanations are
legitimate and that each is illuminating in its own way"
(ibid., 183). It would be difficult to dispute Salmon on
this point. Both explanations seem to have equal claim on
the explanandum event (or, the event to be explained).

Thus, there are at least two explanations that
adequately account for the phenomenon of the balloon's
movement. The question now is whether each is a complete and
independent explanation of the explanandum event. There is
not much in the way of disputing that each is a complete
explanation. Neither explanation (as a specification of the
cause of the balloon's movement) is incomplete by not
referring to the principles invoked by its rival.

Is one of the explanations somehow dependent on the
other? Such an argument would be equally difficult to make.
The project of unifying contemporary space-time physics with
atomic physics, either by bridge laws or by underlying,
unifying laws, has stymied the best of physicists. So, at
this time, there do not seem to be any convincing arguments
which show that the laws and principles of general relativity
either supervene on or reduce to the laws of atomic physics,
and vice versa.

Let us return to our example: The plane accelerates,
the balloon moves forward, and the physicist wins the bet.
Of course, the young boy and his fellow passengers with whom
he had the bet, are in an epistemic quandary with regard to

the balloon's movement. An explanation from the physicist is

42

required in order to remove them from this quandary. But
which explanation ought he give? How can the physicist give
an adequate explanation of the balloon's movement when such
an event admits of two apparently complete and independent
explanations?

According to Kim, the physicist cannot simply choose one
explanation over the other: "It is not implausible to think
that failing to mention either of the overdetermining causes
gives a misleading and incomplete picture of what happened,
and that both causes should figure in any complete
explanation of the event.“ (Kim, 252) Further, the
physicist cannot simply give both explanations, since the
puzzlement of both the boy and passengers would simply be
increased. Not only would they still not know why the
balloon moved as it did, but they would have the additional
quandary of not knowing which explanation is correct (or, at
least, how they are related).

Here, we have a real problem for Kim: The physicist,
constrained by the EEP, cannot explain to the passengers why
the balloon moved as it did. Clearly, this inability is not
for lack of a good explanation. As Salmon claims, both
explanations appear to be legitimate. But having a good
explanation is not enough. According to Kim, if one appears
to have multiple explanations, then one must also know how
the explanations are related. Given that both explanations

are (apparently) inescapably complete and independent, the

43

physicist cannot give an account of how they are related,
viz. reduction, supervenience, and the like.

How is the movement of the balloon to be explained? The
physicist, according to the constraints of the EEP, must say
to the passengers, “I don't know.“ But it is obvious that he
does know. Individually, there is little that is unclear or
controversial about either explanation. How, then, does the
physicist explain the movement of the balloon? Salmon offers
an answer:

Pragmatic considerations determine which ...

[explanation] is appropriate in any given explanatory

context. In the case of the friendly physicist, for

example, an appeal to Einstein's equivalence principle
would have been totally inappropriate; however, the

[atomic] ... explanation might have been made
intelligible to the boy and the other interested adults.
(Salmon, 185)

In this response, the choice between the two explanations has
nothing to do with the relative strength or legitimacy of
either one. According to Salmon, the choice is determined by
the audience to whom the explanation is given. The
principle-of—equivalence explanation requires a high level of
abstraction. Given this level of abstraction, along with the
extreme complexity of general relativity, making such an
explanation intelligible to an untutored audience would be an
arduous and frustrating task for all involved. On the other
hand, the physicist might have a considerably easier time
getting his fellow passengers to understand an explanation

based in atomic physics (at least in broad strokes), since

44

the level of abstraction in such an explanation would not be
nearly as considerable. Further, such an explanation readily
admdts of visual metaphor.

For Salmon, the choice between explanations is governed
by ”pragmatic considerations.“ These considerations are
focused not on the explanation itself, but on the subject or
subjects to whom the explanation is given. Thus, while each
explanation has legitimate claim over the motion of the
balloon, one may be more “appropriate" than the other,
depending on the audience to whom the explanation is given.

In what remains of this chapter, it will be shown,
first, that Kim’s arguments for the explanatory exclusion
principle presuppose that the subject (i.e. the subject to
whom the explanation is offered—-the "explainee') plays a
central and necessary role in explanation theory; second, an
understanding of the subject's role in explanation will show
that the EEP is counterproductive to the epistemic goals
which Kim invokes in support of the EEP. In short, the
subject is indispensable to Kim's arguments for the
explanatory exclusion principle, and yet the subject renders

the EEP untenable.

3. Epistemic Puzzlement and Why-Questions

Salmon indicates that each explanation is legitimate

because each ”provides a different kind of understanding of

the same fact ... Each is illuminating in its own way“ (ibid.

45

183). Different (satisfactory) explanations, in some cases,
may be able to provide a different understanding of a
particular phenomenon. In the present example, Salmon labels
the molecular-based explanation a “bottom-up" explanation and
the relativity-based explanation “top-down": "Bottom-up
explanations ... appeal to the underlying micro—structure of
what they endeavor to explain", while top-down explanations
are “global ... They relate to the structure of the whole
universe.“ (ibid. 184) Thus, each explanation generates a
different perspective on the explanandum event. Each
provides a satisfactory way of understanding why the
phenomenon occurred the way it did.

What makes Salmon's remarks interesting is that,
traditionally, theories of explanation focus on what
explanations are. Such is the case with Hempel's covering-
law model: Explanations are arguments, either inductive or
deductive, where the conclusion of the argument is the event
or fact to be explained (i.e. the 'explanandum') and the
premises (which must include a general, relevant law) are the
”explanans.” In such an analysis, little attention is paid
to what explanations do, i.e. the function of the
explanation.

For Salmon, the function of explanation is to provide
understanding of or perspective on a fact or event. In his
example, the passengers' expectations regarding the movement
of the balloon are foiled. The event is anomalous for them.

The function of a good explanation of the balloon's movement

46

is to provide the passengers with an understanding of or
perspective on the event.

In considering the function of explanation, and thus
invoking concepts such as "perspective“ and “understanding",
a factor is introduced which is largely ignored as
unimportant or problematic in the more traditional models of
explanation: The subject to whom the explanation is given.
I'Perspective" and “understanding“ are relational terms, where
explanation x provides an understanding of or a perspective
on the object or topic of X to a subject S. As will be
shown, the subject is to play a vital role in the rejection
of the EEP and in a subsequent understanding of the
relationship between multiple explanations.

At this point, the concepts of ”perspective" and
”understanding“ are quite vague, but extremely important.
What constitutes a perspective? What does it mean to
"acquire" an understanding of an event or fact? What is the
relationship between correctly explaining a phenomenon and
understanding a phenomenon? In order to be useful, these
concepts need clarification.

We begin with the question, what does it mean to "have"
or “acquire" or "be in" a certain perspective? A beginning
of an answer to this question can be found in Kimfis
commentary on the purpose of explanation. Recall Kimis
epistemological considerations: Kim characterizes the
situation in which we need an explanation as a "state of

puzzlement” or an "epistemic predicament." Satisfactory

47

explanations remove one from this state of puzzlement,
unless, according to the EEP, we have too many.

What Kim has in mind here, as he himself claims, is
Sylvian Bromberger's notion of a “predicament." While there
are various types of predicaments on Bromberger's account,
they share at least one important common characteristic: A
subject A is in a predicament with regard to a particular
question Q. As Bromberger says, "'A is in a ... predicament
with regard to Q' is true if and only if the question
mentioned in it admits of a right answer, but that answer is
beyond what [A] can conceive" (Bromberger, 36).

Therefore, this puzzlement or predicament, of which both
Kim.and Bromberger speak, is relative to a particular
question: One is in such a predicament if one asks a
question Q, where Q is a meaningful question that has a
correct answer, but one cannot conceive of the satisfactory
or correct answer to Q. One finds one’s way out of this
predicament when Q has been satisfactorily answered. Thus,
it seems that, on Kim's epistemic considerations of
explanation, whether or not an explanation is adequate (i.e.
whether it removes one from this state of puzzlement)
necessarily depends on whether the explanation satisfactorily
answers the question Q, where Q is the question to which the
subject's puzzlement is relative.

These considerations jibe nicely with Kim's own
formulation of the epistemological version of the EEP: "No

one may accept both explanations unless one has an

48

appropriate account of how they are related to each other.“
(Kim I, 257) Multiple complete and independent explanations
are epistemically counterproductive, i.e. they do not remove
puzzlement. In other words, multiple complete and.
independent explanations do not answer the subject's
question. Indeed, not only do they not answer the question
which expresses the subject's puzzlement, they also serve to
introduce a new question: “Which explanation correctly
answers my original question?"

Puzzlement is compounded since new questions or
puzzlements are introduced without resolving the original
questions or puzzlements. Therefore, the epistemological
motivations for the explanatory exclusion principle must
necessarily include considerations of the questions which
express the puzzlement of the subject. These considerations
must include an account of when and how such puzzlement is
removed, i.e. they must include the criteria for satisfactory
answers to such questions. These considerations, therefore,
are directly implied by Kim's epistemic arguments for the
explanatory exclusion principle.

Clearly, "puzzlement" is a central concept in Kim's
epistemological arguments for the EEP. And its importance is
paralleled only by its vagueness. Thus, it is necessary to
clarify what exactly is meant by "puzzlement.“

There are at least two ways of construing puzzlement--
as a psychological phenomenon and as an epistemic phenomenon.

In the former case, one may “feel“ puzzled and such a feeling

49

may be relieved simply by forgetting the question that
expressed the puzzlement or by taking a drug which would
cause the feeling of puzzlement to dissipate, even if the
question lingers. Even an answer which is false might
relieve psychological puzzlement. Clearly, this is not the
type of puzzlement relevant here. The specific features
which characterize epistemic puzzlement can be divined from
Hintikka's explication of the semantics of questionsl.

Every wh—question implies two concepts: a
"presupposition" and an epistemic goalz. The presupposition
is a description of the epistemic state of affairs necessary
for the asking of a wh—question. An example of such a wh—

question, asked by a subject S, might be:

(1) "Where has Smith taken Jones' car?”

For question (1), the presupposition would be the

propositional attitude,

(2) S knows that Smith has taken Jones' car.

 

1 The examples used here to explicate the epistemic
features of puzzlement can be generalized through Hintikka's
epistemic logic of questions and answers. Such an account
can be found in "Semantics and Pragmatics for Why— Questions“,

in Wigwam (1995). p 636- 657

2 Although Hintikka refers to this latter concept as the
'desideratum," for the sake of ease and clarity, we shall use
the term "epistemic goal." See ibid. p. 637.

50

Notice that there are two necessary conditions for the
legitimacy of the question. Where A: 'Smith has Jones' car',
it is necessary that (a) ”S knows that A" is true (in other
words, the presupposition must be true) and (b) “A" is true3.
If (a) were false, then S would not be in a position to ask
question (1). If (b) were false, then (1) would not admit of
a correct answer. For instance, if Smith had not taken
Jones' car, then the question, “where has Smith taken Jones'
car?“ cannot be given a good answer. Thus, (via conditions
(a) and (b)) the presupposition of a question is a necessary
condition for the asking of question Q.

Question Q, however, indicates that there is an
epistemic state of affairs that subject S is lacking. This
state of affairs is expressed in the epistemic goal of the
question: “The [epistemic goal] of a question is a
description of the state of affairs that the questioner would
like to have brought about" (Hintikka, 638). Relative to

(1), the epistemic goal can be expressed as such:

(3) S knows where Smith has taken Jones' car.

If question (1) is answered correctly, then (3), the
epistemic goal, becomes true. The function of answers to
such questions is to bring about the epistemic state of

affairs expressed by the question's epistemic goal. As

 

3 Notice that, while (a) implies (b), (b) does not imply
(a). Thus, as conditions for question legitimacy (a) must be
distinguished from (b).

51

Hintikka says, "the outcome of a conclusive answer to a
question is the extra information ... of its [epistemic goal]
as compared with the information conveyed by its
presupposition" (ibid., 639).

While an explication of the logical features of
questions and their answers is necessary to understand the
general relationship between presupposition, wh-question and
epistemic goal, such an explication is not necessary for
present purposes. This brief sketch is enough to understand
the conditions of epistemic-puzzlement and the method by
which such puzzlement is resolved: For every question Q,
there is implied an epistemic goal of Q and a presupposition
of Q. The epistemic goal expresses the epistemic state of
affairs which will be achieved once the question is
satisfactorily answered. The presupposition expresses the
epistemic state of the questioner at the time the question is
asked. The epistemic gap between presupposition and one's

epistemic goal is the puzzlement expressed by Q.

puzzlement/why—question

 

 

[presupposition

 

........ epistemic gap-------- epistemic
goal

 

 

Figure l

The state of puzzlement (of the questioner) is expressed

by the question Q. The alleviation of that puzzlement is

52

characterized by the information gathered (i.e. a
satisfactory answer to Q) which allows the questioner to move
from the epistemic state of affairs expressed in the
presupposition of Q to the epistemic state of affairs
expressed in the epistemic goal of Q. This "epistemic
movement“ cannot be accomplished by forgetting Q or by taking
a drug which dissipates any psychological feeling of
puzzlement. These considerations of the relationship between
the presupposition and the epistemic goal of Q clearly allow
us to isolate this discussion from any psychological

conception of puzzlement.
3. Questions and Contexts

Let us consider those criteria which, at least in part,
determine whether or not an explanation is satisfactory, i.e.
when the movement from the presupposition of Q to the
epistemic goal of Q has been successfully achieved. Of
course, that the explanation—candidate is true is a necessary
condition for its being satisfactory, but is it sufficient?
Will the truth of an explanation-candidate suffice to answer
the relevant question effectively and eliminate puzzlement?

It is possible for an explanation to be true and still
not remove puzzlement, or answer the question. Alan
Garfinkel uses a famous example to illustrate this point:
“When Willie Sutton was in prison, a priest who was trying to

reform him asked him why he robbed banks. I'Well," Sutton

53

replied, “that's where the money is“.“ (Garfinkel, 21)
Sutton certainly thinks that he has answered the priest's
question, and his explanation is true. However, it is
doubtful that the priest's puzzlement over why Sutton robs
banks is abated. It seems that they have each considered
this question in different ways. Each has a different
understanding of the question.

Garfinkel characterizes this difference in terms of a
varying “space of alternatives.“ (21) Sutton presupposes a
certain set of alternatives to the priest's question: “Why
do you rob banks, as opposed to diners, grocery stores,
private homes, etc.?" So the alternatives in this question
are in contrast to banks, i.e. objects of robbing. Thus, the
space of alternatives includes banks, diners, grocery stores,
private homes, etc.

Of course, the priest is interested in a different set
of alternatives: "Why do you rob banks, as opposed to taking
up an honest, legitimate profession?” Here, the alternative
is in contrast to the activity of robbing rather than the
object that is robbed. So the space of alternatives includes
robbing banks and alternatives to the act or occupation of
robbing banks. The misunderstanding that takes place between
Sutton and the priest can therefore be characterized
initially as a difference in alternative spaces.

Certainly, the two alternative spaces understood by
Sutton and the priest are not the only possible

interpretations of the question. A wide variety of

54

alternative spaces are possible. For instance, one might
contrast robbing banks with painting banks, building banks,
demolishing banks, etc. If such a variety of alternative
spaces are possible, how does one determine which alternative
space is invoked in a particular question? In general, the
particular alternative space of a question asked is
determined by the context of the situation4.

The response that Sutton gives to the priest's question
is generally regarded as comical or humorous, since Sutton's
response is not only a clear misunderstanding of the
question, but it is also quite unexpected. It is unexpected
to a general audience, because such an audience will more
than likely have understood the alternative space implied by
the priest's version of the question. This is because a
general audience would recognize certain important features
of the context.

For instance, they would recognize the questioner as a
priest (by clues such as a collar or a bible) and understand
that the priest is interested in reforming Sutton, or at
least in getting Sutton to repent his sins. Most observers
of this situation, knowing what they know about priests, will
not make the same misinterpretation that Sutton makes. This
makes his misinterpretation both unexpected and comical.

Clearly, the priest is not interested (as Sutton interprets)

 

4 Note that the present use of the term ”context" is
non—technical. The “context of the situation" is simply the
salient features of the situation in which the question is
asked--features the might provide clues as to the intended
meaning of the question.

55

in the advantages of robbing banks instead of grocery stores,
private residences, etc.

Ascertaining certain features of the context of the
situation, namely that a priest is asking the question and
priests have certain aims and roles to play in such
situations, allows a general audience to understand the
alternative space of the priest's version of the question.
Sutton either ignores or is ignorant of such features of the
context, which allows him to misinterpret the priest's
question.

If those contextual features were changed, so would the
interpretation of the question. For example, if, instead of
a priest, the questioner is a friend of Sutton's and a thief,
Sutton's interpretation of the question (that is, Sutton's
alternative space) would be quite reasonable. Thus, changing
certain features of the context in which the question is
asked alters the conditions by which the question is
interpreted. In this latter context, Sutton's interpretation
of and answer to the question, "why do you rob banks?“ would
be reasonable and, probably, appropriate.

So far, the relation between contexts, questions and
alternative spaces has been painted in rather broad strokes.
It is, at this point, necessary to supply some details. The
Sutton example illustrates that it is possible (though, as

will be shown, not essential) to ask questions which express

 

5 A good deal of this detail is drawn from.Bas van
Fraassen's work. See specifically “A Theory of Why

Questions " in Wrasse.

56

different alternative spaces with the same utterance. Thus,
such an example allows us to distinguish between an
interrogative utterance and a question (or interrogative
proposition) where the question is characterized (in part) by
its alternative space.

The priest's utterance, “Why do you rob banks?“ is taken
to express two different questions and one can distinguish
between the questions by understanding their different
alternative spaces. Further, coming to understand the
question expressed in an interrogative sentence can often be
accomplished by understanding the context in which the
sentence is uttered.

There is still a good deal of detail which remains
unexamined regarding the relationship between interrogative
sentence, question and context6, but the detail given thus
far is enough for present purposes. The upshot of these
considerations is this: According to Kim, explanations
remove a subject from a state of puzzlement. That is their
epistemic function. Puzzlement, in order to be
philosophically respectable, must first be expressed in a
why-question Q. More specifically, puzzlement is the
epistemic gap between the presupposition of Q and the
epistemic goal of Q. To relieve a subject of his or her

puzzlement is to bridge the gap between presupposition and

 

6 Some of these details will be addressed in chapter
III.

57

epistemic goal. This is accomplished by successfully
answering the subject's question.

But in order to answer such a question and thereby
alleviate epistendc puzzlement, one must understand the
question that is asked. However, the Sutton example
illustrates that in order to understand (and thus
successfully answer) a particular question, one needs to
understand (at least) the alternative space of the question.
If such an understanding is not achieved, then one runs the
risk of misunderstanding the question and giving an
unsatisfactory answer. Such is the case with Willie Sutton's
answer to the priest's question. Clearly, Sutton's answer
does not relieve the priest of his epistemic puzzlement
regarding Sutton’s chosen profession, even though his
explanation is true.

But what, in general, are the conditions by which a
question is successfully answered? Recall that, in the
priest's understanding of the interrogative, ”Why do you rob
banks?“ the robbing of banks is set up in opposition to
earning an honest wage, or living an honest life, i.e. “Why
do you rob banks, as opposed to earning an honest wage?“ In
other words, the robbing of banks (what van Fraassen calls
the topic of the question, "Why do you rob banks?“) is a
member of the space of alternatives, amongst other possible
alternatives; in this case, earning an honest wage. Broadly
speaking, a successful answer to a question will select the

topic of the question from the space of alternatives.

58

For example, Sutton's answer, "That's where the money
is“ is a good answer to the question “Why do you rob banks?"
where the alternative space includes {robbing banks, robbing
grocery stores, robbing trains, robbing private residences,
etc.}, because it selects “robbing banks“ from the other
alternatives. The fact that banks have the most money is why
one robs banks and not grocery stores, trains, private
residences, etc.

So, in general, a good answer selects the topic from the
range of alternatives. In other words, an answer shows why
the topic is true and the alternatives are false. This
notion of a successful answer fits quite nicely with
Hintikka's idea of epistemic movement from presupposition, ”I
know that Willie Sutton robs banks,“ to epistemic goal, "I
know why Willie Sutton robs banks“, as opposed to grocery
stores, trains, etc.

Again, these considerations of the nature of puzzlement,
why-questions, context, and alternative spaces follow from
the epistemic foundations of the explanatory exclusion
principle. The epistemic aim of explanation is to remove the
subject from their state of puzzlement. The conditions of
such a removal are made explicit by the above considerations.
If it turns out that the EEP runs counter to this explication
of puzzlement removal, then it also runs counter to the
epistemic aims of explanation which are supposed to support
it. Such a situation would do more than simply show that the

EEP is problematic, as the examples given by Scriven, Fodor

59

and Salmon illustrate. Such a situation would render the EEP
untenable, since it would undercut its own justificatory

principles.

4. The Epistemic Aims of Explanation

If the above considerations are correct, then (as will
be shown) the EEP is counterproductive to the very epistemic
aims of explanation which are supposed to serve as its
justification. Thus, the EEP is untenable and therefore not
a reasonable condition on explanation in general. The final
section of this chapter is dedicated to the support of this
claim.

Before such an argument can be made, there is an
important addendum to the considerations regarding the
relationship between interrogative sentence and question.
The Sutton example illustrates how one can understand the
utterance of an interrogative sentence and yet misunderstand
the question that is being asked. In such a case, the same
interrogative sentence is capable of indicating multiple
questions, viz. multiple alternative spaces. However,
nothing crucial hangs on the idea that Sutton and the priest
use the same utterance, or interrogative sentence, to express
different questions. Certainly, the priest's interrogative
sentence might have been phrased differently, so as to allow
Sutton to understand the question he had in mind. The

general fact that different questions can be asked through

60

the same interrogative utterance allows us to distinguish
between questions and interrogative sentences. It is not
essential that in every example, the same interrogative must
be given to express different questions.

How, then, do the above consideration serve as a
response to Kim and the EEP? Let us take another example.
Smith and Jones see their friend Roy walking down the street.
Smith is aware that Roy has a neurological condition that,
when untreated, causes his right arm to spasm. Jones is not
so aware. Suddenly, Roy begins waving his arm wildly. Smith
utters the interrogative, "Why is Roy's arm flailing about?"
Jones, unaware of Roy's condition, assumes Roy is hailing a
nearby cab and utters, “Why is Roy hailing a cab?“

Given that Smith is aware of Roy's neurological
condition, it is reasonable to assume the alternative space
{Roy's hand is waving, Roy's hand remains at his side}.
Jones, on the other hand, knows that Roy is on a fixed income
and cannot afford extravagances. Thus, it is reasonable to
assume the alternative space {Roy is hailing a cab, Roy takes
a bus, Roy walks to his destination, Roy hitches a ride}.
After all, Roy is on a fixed income and can 111 afford to
spend money on a cab. Why, then, is he hailing a cab, as
opposed to simply taking the bus or walking or hitching?

Let us suppose two possible scenarios as outcomes.
Scenariol: Roy sits down on a park bench, his arm still
waving. He takes his medicine and relaxes. After a few

minutes, the movement of his arm subsides. ScenarioZ: A cab

61

pulls up to the curb near Roy. Roy stops waving his hand,
gets in, instructs the driver and the cab departs.

It is clear that in scenariol, Roy is not hailing a cab
at all. Thus, Jones' question, “Why is Roy hailing a cab?“
is misguided. In fact, characterizing the movement of Roy's
arm in an intentional matter would be inappropriate, since it
is obviously not an intentional action. As his subsequent
actions suggests, his behavior is a result of his
neurological condition and can be appropriately characterized
as a spasm, rather than a purposive action. On the other
hand, Smith's question is entirely appropriate. The answer,
"Roy's armrmovement is a result of his neurological
condition"7 picks out the topic of Smith's question from its
alternative. These considerations appear to agree with Kimis
analysis. The neurological idiom allows Smith to ask a
legitimate question and receive a good answer, whereas the
intentional idiom has failed Jones. His characterization of
Roy's behavior as intentional, implied by his question, is
inappropriate.

ScenarioZ, however, has radically different consequences
for Smith and Jones' questions. It is clear that in
scenarioZ, Roy is indeed hailing a cab. On Hintikka’s
analysis, Jones' question is appropriate, since it meets both
conditions for question legitimacy: Where the question is,

“Why is Roy hailing a cab?" it is necessary that “Roy is

 

7 Certainly, a more detailed answer, regarding how the
neurological disorder caused the movement, would be possible,
but the given answer is sufficient for present purposes.

62

hailing a cab“ is true and "I [Jones] know that Roy is
hailing a cab“ is true. The context of scenarioz determines
that Jones' question is legitimate, whereas the context of
scenariol determines that it is not. In this way, whether or
not the topic of a particular question is true is determined
by the context of the situation.

Since, in scenarioZ, Jones' question is legitimate (i.e.
it meets both of Hintikka’s conditions) it must admit of a
correct answer or explanation. In other words, there is an
answer which makes the topic of the alternative space true
and the alternatives to the topic false. Notice, however,
that since the topic (and thus the question) and the other
members of the alternative space are characterized
intentionally, the answer which selects the topic from the
alternative space must make reference to Roy's intentional
states. For example, one might offer the answer "Roy
believed he was late for work and that a cab would be the
fastest way to get there." Such an answer (if true) would
select the topic from the space of alternatives. In other
words, it would explain why he took a cab and did not take
the bus, hitch a ride or walk.

But why must an answer to Jones' question invoke Roy's
intentional states? Wouldn't an answer such as "Roy is late
for work. The fastest way to work is by cab,“ be a good
answer? Why do Roy's intentional states have to enter the

picture at all?

63

Here, the unique character of propositional attitudes
plays a crucial role. There are a number of possible
situations (e.g. Roy's watch is fast, Roy forgot that
daylight savings time ended, etc.) in which "Roy is late for
work“ or “the fastest way to work is by cab" are false, and
yet “Roy believes that he is late for work“ or “Roy believes
that the fastest way to work is by cab" are true. Whether or
not Roy is late for work is irrelevant to Roy's hailing a
cab. It might very well be false. What is relevant is that
Roy believes he is late for work, regardless of whether or
not he is, in fact, late. Thus, in scenarioZ, an answer to
Jones' question, if it is to select the topic from the
alternative space, must involve reference to Roy's
intentional states--his beliefs and desires.

Here, we come to the heart of the problem: Jones'
puzzlement is expressed in a particular why-question. The
alleviation of Jones' puzzlement (the epistemic goal of
explanation, according to Kim) requires an answer to her
question which necessarily includes reference to Roy's
intentional states, because such an answer will pick out ~Roy
is hailing a cab" from amongst other possible intentional
actions in the alternative space. However, in principle, the
explanatory exclusion principle rules out the possibility of
invoking such intentional states in explanations, unless the
states posited in the intentional idiom are somehow reducible
to (i.e. systematically identifiable with) the states of the

neurophysiological idiom. But arguments such as multiple

64

realizability preclude any such systematic reduction.
Therefore, according to the EEP, such intentional states
cannot enter into explanations of behavior.

But this simply means that, in principle, Jones'
puzzlement, as expressed in her why—question, cannot be
alleviated, since a necessary condition for answering her
question and thus alleviating her puzzlement must include
reference to Roy's intentional states. But recall that
abating puzzlement is the epistemic goal of explanation, as
stated by Kim himself:

I take it that explaining is an epistemological activity

To be in need of an explanation is to be in an
epistemically incomplete and imperfect state, and to
gain an explanation is to improve one's epistemic

situation; it represents an epistemic gain. (Kim I,
255-6)

Kim invokes this argument as a justification for the EEP.
However, if we are constrained by the EEP, then it must be
the case that any question the answer to which must include
reference to intentional states, such as Jones' question,
cannot be answered. The puzzlement cannot be abated. The
”epistemic gains” cannot be achieved.

The problem for Kim is compounded if we consider Fodor's
arguments. The EEP can be applied equally well to all the
non-physical sciences. Thus, any why—question the answer to
which must refer to the systematically irreducible states of
any non-physical sciences cannot be answered. The EEP must

rule out, in principle, the possibility that any such

65

question can be answered and therefore the puzzlement
expressed in those questions cannot be abated.

Clearly, if the alleviation of puzzlement is the
epistemic goal of explanation, then rather than being
supported by this goal, the explanatory exclusion principle
is counterproductive to it--to an extreme degree. Therefore,
the EEP is not, as Kim believes, a reasonable constraint on
explanation in general. On the contrary, it is
counterproductive to the very goals of explanation that Kim
sets out in support of the EEP.

One might offer an objection to this argument: Doesn't
there exist a kind of asymmetry between the intentional and
neurophysiological idioms? In all contexts (scenarios), it
is possible to ask a question for which a neurophysiological
answer is appropriate. Even in scenarioZ, in which Roy's
behavior can legitimately be characterized intentionally, one
can also ask a question whose answer must involve reference
to Roy's neurophysiological states.

The same cannot be said for the intentional idiom.
There are cases, such as scenariol, in which an intentional
characterization of behavior is illegitimate. In short, for
every behavioral event, there is a possible
neurophysiological explanation. But it is not the case that
for every behavioral event there is a possible intentional
explanation. Jones' question in scenariol is just such a
case. Doesn't this suggest a certain explanatory priority of

the physical over the intentional? Don't we get better

66

explanations from the neurophysiological idiom because the
idiom is not subject to explanatory or descriptive failure?

Certainly, there is an asymmetry between the intentional
and the neuro-physical. While this asymmetry does imply a
difference in explanatory scope, it does not imply a general
priority of the neurophysiological idiom over the
intentional. In cases such as scenariol, the
neurophysiological idiom is superior to the intentional in
that Roy's behavior admits of physical description, but not
to intentional description. But this does not imply a
superiority in general.

In cases such as scenarioZ, the neurophysiological idiom
can give a description of Roy's behavior and therefore answer
certain questions about it. But it does not follow that it
can answer all possible questions about Roy's behavior.
Jones' inquiry is an example of such a question. Therefore,
while the neurophysiological idiom is superior to the
intentional in that the former does not admit (in general) of
explanatory failure, it does not follow that neurophysical
explanations are superior to intentional explanations in
principle.

Let us consider a final objection: In Smith's question,
the topic is "Roy's hand is flailing“ or “Roy's hand is
moving," whereas in Jones' question, the topic is “Roy is
hailing a cab.“ Isn't there a sense in which they are not
explaining the same thing? There is a difference between

explaining the act of cab—hailing and the act of arm-

67

flailing. Can't we simply say that they are explaining
different events?

Kim confronts a similar objection to his arguments for
the explanatory exclusion principle. Recall from chapter I
the example of Smith's climbing a ladder to retrieve her hat
from.the roof of the garage. There are two possible
explanations of this behavior, one from the intentional
idiom, the other from the neurophysiological idiom. The
difficulties of the explanatory exclusion principle arise
when both explanations are said to give complete and
independent explanations of Smith's hat-retrieval behavior.

But one might object to Kim that the problems of having
multiple explanations do not arise, since there is more than
one event to be explained. In the neurophysical explanation,
the event to be explained (the explanandum) is expressed in
non-intentional, ”mere physical movement" terms. In the
intentional explanation, the event is explained in
intentional, purposive terms. The problems of explanatory
exclusion arise only when there are multiple explanations of
a single explanandum event. Such is not the case here.

Kim responds that, while there are multiple descriptions
of Smith's behavior, there is a sense in which all
descriptions are "about“ the same event: “Although the two
explanandum statements are not equivalent or synonymous,
there is an evident sense in which they 'describe' one and

the same event, the same concrete happening.“ (Kim I, 242)

68

I believe this is a reasonable response to this type of
objection. Regarding Roy's armrmovement, while there may be
many different ways to describe Roy's behavior, and thus many
different question-topics about his behavior, they all have
the same referent; namely, Roy's bodily activity at that
time. Thus, while different questions are being answered and
different explanations are given, they all explain the same
event, given that the topic of each question has a referent
identical to the topic of the other. So answers to both
Smith and Jones' questions will be different, but they will
be explanations of the same event.

The explanatory exclusion principle is not, as Kim
claine, a reasonable constraint on explanation in general.

On the contrary, it is counterproductive to the epistemic
goals of explanation. The problems with the explanatory
exclusion principle illustrate that we must develop an
analysis of mental causation which makes sense of and makes
room for the causal efficacy of irreducible mental states in
the production of bodily behavior, given that the physical
idiom.is (in principle) capable of providing exhaustive
causal analyses of any behavioral event, intentional or
otherwise. Further, we now have the key which will allow for
such an analysis of mental causation: A pragmatic conception
of explanation, i.e. a conception of explanation in which the
subject to whom the explanation is offered plays a central

role. What is required is a fuller understanding of the

69

pragmatics of explanation, such that a positive analysis of

mental causation can be given.

70

Chapter III: The Pragmatic Conception of Explanation

1. A Preliminary Solution

The above considerations make clear that the explanatory
exclusion principle is too strong a condition on explanation
in general. Thus, we need not rule out, in principle, the
possibility of the intentional idiom providing complete and
independent explanations of bodily behavior. But the
original problem.remains: Physicalism entails that for every
behavioral event, one can give an exhaustive causal analysis
within the physical idiom. If mental states are
systematically irreducible to physical states, then their
properties (including causal properties) cannot be
assimilated into this causal analysis. Thus, while we need
not rule out intentional explanations in principle, it is not
yet clear how we can make room for such causal explanations,
relative to the exhaustive causal analyses that the physical
idiom is capable of providing.

At best, it is not clear what causal work remains for
mental states and properties if all physical behavior has a
sufficient physical cause. At worst, attributing causal
efficacy to irreducible mental states runs the risk of
encountering the metaphysical and epistemological
difficulties that (supposedly) motivate the explanatory
exclusion principle; the most important of which is causal

overdetermination. While we need not heed the explanatory

71

exclusion principle and exclude intentional explanations in
principle, giving multiple complete and independent
explanations still smacks of systematic overdeterminationl.

Thus, a rejection of the explanatory exclusion principle
does not automatically relieve us of the original
difficulties which plague the idea that mental states are
both irreducible and causally efficacious. Our question
remains: How do we make sense of the causal work done by
irreducible mental states within a physicalist conception of
the world?

The answer to this question lies within those pragmatic
considerations of explanation briefly explicated in chapter
II. The role of the mental in causing behavior can be
understood through an analysis of the role of the mental in
causal explanations of behavior. Within such an analysis,
our question can be rephrased: How do we make sense of and
make room for causal explanations provided by the intentional
idiom, given that, for any behavioral event, a complete
causal explanation can be given in the physical idiom?

The goal is to allow for the fact that the physical
sciences are capable of giving complete explanations of any
behavioral event, while maintaining the importance of
irreducible, intentional explanations. This is exactly what

the pragmatic considerations of explanation allow us to do.

 

1 Systematic overdetermination, recall, is a causal

overdetermination which is set up and maintained by our
theories of the world, as opposed to an accidental or
anomalous overdetermination. See chapter I, p. 13.

72

Physicalism, again, implies that any behavioral event is
subject to an exhaustive causal analysis in the languages of
the physical sciences. Therefore, any behavioral event can
be given a complete physical explanation. On the pragmatic
conception of explanation, such a claim can be understood to
mean that, first, for any behavioral event, there is a why-
question regarding that event (i.e. a why-question whose
topic is the event) which requires an answer from within the
physical idiom; second, any such appropriately—phrased why—
question which requires an answer from the physical language
can be answered. Such are the implications of physicalism
for the pragmatics of explanation.

However, what is not implied by physicalism.is that any
why—question in general can be given a physical explanation.
This is evident in the arm—waving example discussed in
chapter II. The question, “Why is Roy hailing a cab?"
characterizes Roy’s behavior as intentional and requires an
explanation which refers to Roy’s intentional states. A
physicalist answer can be given to explain Roy’s behavior,
but such an answer would be inappropriate, i.e. it would not
satisfactorily answer the why—question and relieve the
subject of his/her epistemic puzzlement. Roy’s mental states
are able to do causal work that his physical states cannot in
that his mental states can provide causal explanations that

his physical states cannot.

73

Notice, this does not jeopardize the implications of
physicalism. It remains true that, for any behavioral event,
a physical explanation can be given. In other words, for any
behavioral event, a why-question requiring an answer couched
within the physical idiom, can be asked and (in principle)
can be answered. But on a pragmatic conception of
explanation, this does not imply that the physical idiom.can
provide an answer to any why-question regarding a behavioral
event. Mental states do causal work by providing answers to
why-questions that cannot be answered in the physical idiom.

This analysis of the causal role of mental states in the
production of behavior is, admittedly, brief. However, it is
clear that the pragmatic conception of explanation plays an
indispensable role in this analysis. Thus, in order to
understand and justify this analysis of mental causation, we
must fill in some of the details of such a theory of
explanation. The remainder of this chapter is dedicated to
this purpose. A more complete understanding of the pragmatic
conception of explanation will lend our analysis of mental
causation both clarity and credence.

2. Understanding Questions: Alternative .Space, Topic
and Relevance Relation

Recall Kim’s claim that the epistemic function of
explanation is to remove a subject from his/her state of
puzzlement. The subject expresses this puzzlement in the

form of a why—question. As such, puzzlement can be

74

understood as an ”epistemic gap” between the presupposition

and the epistemic goal of the questionz.

puzzlement/why—question

l l

presupposition """"""" epistemic gap """"""" [epistemic
goal

 

 

 

 

 

 

Figure 1
To answer the why-question is to allow the subject to move
from presupposition to the epistemic goal. Such answers,

which remove the subject from a state of puzzlement, are

 

 

 

 

 

 

explanations.
explanation
presupposition answer : epistemic
goal
Figure 2

Therefore, the conditions by which an explanation is deemed
adequate are determined by the conditions by which a why-

question is successfully answered, i.e. by which the move

 

2 Recall that the ”presupposition" is the epistemic

state of affairs necessary for the asking of a particular why
question. Where the question is “Why is it the case that
X?', “I know that X' is the presupposition. The epistemic
state expressed by the “epistemic goal" includes "I know why
X" as well as 'I know that X.” Again, Hintikka labels this
latter concept the “desideratum' of the question. See
chapter II, p. 8.

75

from presupposition to epistemic goal is made. An
examination of such conditions was begun in chapter II. In
order to understand the pragmatic theory of explanation more
fully, a more detailed analysis is in order’.

Recall the distinction between interrogative sentence
and question. The subject desiring an explanation of a
particular phenomenon or event utters an interrogative
sentence. In order to alleviate the subject’s puzzlement,
the explainer must ascertain the specific question asked, so
that a satisfactory answer can be given. Willie Sutton’s
answer, ”That’s where the money is,” demonstrates a
particular understanding of the priest’s interrogative
sentence, but a misunderstanding of the intended question.
The present task is to enumerate the features of why-
questions which must be understood, such that an appropriate
answer can be given.

One such feature is the alternative space‘. In order to
understand a why-question, one must understand the
alternative space of the question. Recall that Garfinkel
characterizes the Sutton example as a mistake in alternative
space. Sutton’s answer selects ”robbing banks” from the

alternative space consisting of {robbing banks, robbing

 

3 The analysis given will be based primarily on van

Fraassen’ s theory of why- questions from Ihe_§gien§ifiig_1mage,
1980.

‘ Garfinkel’s concept of the alternative space is
roughly equivalent to van Fraassen’s idea of the contrast

class. For present purposes, I will take these terms as
synonymous.

76

private residences, robbing grocery stores, etc.). The
priest’s question, however, requires Sutton’s answer to
select ”robbing banks” from the alternative space consisting
of {robbing banks, leading an honest life with an honest
job}.

However, the interrogative sentence ”Why do you rob
banks?” can express either question with either alternative
space. Hearing the interrogative sentence uttered is not
enough. One must be able to grasp (at least) the alternative
space of the question which the interrogative sentence
expresses. Again, we are (ceteris paribus) able to ascertain
the intended question via the context of the situations.

In the context of the Sutton example, one notices that a
priest asks the question. There are certain facts that one
knows about priests; namely, that they are interested in
reforming sinners and not (one would hope) in learning the
best places to rob. Thus, Sutton’s interpretation of the
question is clearly inappropriate, given the clues provided
by the situation.

To make the point clear, let us take a second example.

A young girl is fascinated by a steam-engine train that rests

at a train station. The wheels of the engine begin to churn

 

5 As is the case in Chapter II, the term ”context” is

not used in any technical sense. Here, the term refers to
the situation in which the question is asked. Clues provided
by the situation allow an explainer to ascertain the question
expressed by the interrogative sentence.

77

and the train lurches forward. The girl, in awe, asks the
conductor, ”Why is the train moving?” How should the
conductor understand the girl’s question? The context of the
situation includes the fact that the inquirer is an
inquisitive young girl and that she is fascinated by the
movement of the train. From these facts, there may be
several reasonable alternative spaces that the conductor can
infer. For instance, one might assume that the girl is
interested in how the train is able to move at all. Thus, a
reasonable interpretation of the alternative space might be,
{the train moves, the train is incapable of motion, etc.).

Before the conductor has a chance to answer the girl, an
angry young man, carrying a suitcase, hurries up to the
conductor with his pocket-watch extended and demands, ”Why is
the train moving?” Here, the conductor is presented with the
same interrogative sentence, but with a radically different
situation. There are several salient features of this second
scenario: The young man is carrying a suitcase, indicating
he is a traveler and supposed to be on the train. He is
angry, so it is doubtful he is merely indulging an interest
in the mechanics of the train. Finally, he shows the
conductor his pocket-watch. A reasonable alternative space
might be, {the train is moving, the train remains at rest
until the proper departure time}.

There may be several reasonable interpretations for each
situation, and thus several different possible alternative

spaces. But it is clear that the different situations in

78

which the utterances are made require radically different
interpretations of the intended question. This difference in
interpretation can be characterized, in part, by a difference
in alternative space. Thus, the alternative space must be
understood (via clues provided by the situation in which the
question is asked) such that an appropriate answer to the
question can be given.

While the alternative spaces of the questions of the
young girl and the angry man are markedly different, their
respective questions share at least one common feature: Both
questions are about the movement of the train. In both
cases, it is the motion of the train that is in need of
explanation. The difference in the alternative space of each
question shows that the conditions by which the motion of the
train is explained will be considerably different for each
question. However, there is still a basic, intuitive sense
in which the movement of the train is the subject matter of
both questions.

Generally speaking, where ”why is it the case that X?”
is the question, X is the subject matter of the explanation.
In a broad sense, X is the event to be explained. As such,
we are able to consider X in two different ways. First, in
order for a subject to ask ”Why is it the case that X?”, that
subject must know that X is the case. Hintikka labels
knowing X (or, more specifically, ”I know X”) as the
presupposition of the question. Recall that puzzlement is

the epistemic gap between a presupposition and one’s

79

epistemic goal. In order to be in a state of puzzlement,
i.e. in order to ask a why—question, a certain epistemic
state of affairs must obtain. That epistemic state of
affairs is expressed by the presupposition. Thus, ”I know
that X”, insofar as it is a necessary epistemic condition for
the asking of ”Why is it the case that X?”, is the
presupposition of the question.

However, ”I know X”, as the presupposition, is not
itself a feature of the why-question. Since it is the
epistemic state of affairs necessary for the asking of the
question, the presupposition is antecedent to the asking of
the question. On the other hand, there is a sense in which X
itself is a feature of the why-question. Given the question,
”Why is it the case that X?”, a successful answer must
demonstrate why X is the case, as opposed to the other
possible alternatives.

Each why—question has an alternative space. The members
of this alternative space include not only possible
alternatives to X, but X as well. A good answer to ”Why is it
the case that X?” will ”select” or ”favor” x over the other
members of the alternative space. In other words, a good
answer will show why X is true, as opposed to the other
alternatives (or at least show that X is more probable than
its alternatives). As such, X is the topic of the why-
question, ”Why is it the case that X?”

The alternative space is simply a set of possible states

of affairs or events. The topic is the actual state of

80

affairs, amongst a number of possibilities. A satisfactory
answer to ”Why is it the case that X?” must demonstrate why X
is true, instead of other possible alternatives. ”Why do you
rob banks, as opposed to grocery stores, trains or private
residences?”, ”Why do you rob banks, as opposed to living an
honest life with an honest job?”, ”Why is the train moving,
as opposed to remaining motionless?”. A good answer to any
of these questions, must select the topic from its
alternatives (i.e. make the topic true), or favor the topic
over its alternatives (i.e. make the topic more probable than
its alternatives)6.

Thus, in the pragmatic theory of explanation, X serves
two different functions: First, X is contained in the
presupposition of the question, insofar as knowing X is a
necessary epistemic condition for the asking of ”Why is it
the case that X?”; second, X is the topic of the question,
insofar as a satisfactory answer must select X from the
alternative space of the question. Thus, as the topic of the
question, X is a defining feature of the question and, as
such, helps determine the conditions for a satisfactory
answer. In short, a successful answer must seleCt the topic

from the alternative space.

 

6 The general conditions by which an answer favors or

selects the topic from the alternative space is not a central
concern. However, a brief treatment will be given in section
II of this chapter. For a thorough treatment of these
issues, see van Fraassen, p. 146.

81

Finally, the conditions by which an answer is determined
to be relevant must be enumerated. Such conditions are
contained in what is generally called the relevance relation
of the question. An answer-candidate must be relevant to the
question at least in the sense that it must pick out the
topic from the alternative space. For instance, the
sentence, ”The sky is blue,” is true, but it is irrelevant to
the question ”Why is the train moving?” since it does not
address the topic of the question, viz. picking out the topic
from.the alternative space. Thus, at a minimum, an answer is
relevant if it addresses the topic of the question. But a
relevant answer must do more than simply address the topic.

Recall the example of Roy’s hand-waving behavior.

Smith, aware that Roy has a neurological condition which, at
times, causes certain muscles to seizure, asks ”Why is Roy’s
arm waving about?”. A relevant answer must select the topic
from.the alternative space. However, there may be other
questions containing the topic, ”Roy’s arm is waving about.”
For instance, in Jones’ question, ”Why is Roy hailing a
cab?”, the topic, ”Roy is hailing a cab,” is equivalent to
”Roy’s arm is waving about” in the sense that both topics are
the same behavioral event, even if they describe the event
differently.

Thus, if the requirement of relevance is limited to
selecting the topic from the alternative space (i.e. making
the topic true), then an answer to Smith’s question will also

count as an answer to Jones’ question, since both questions

82

maintain the same topic. But, as shown in chapter II, a
neuro-physiological answer to Jones’ question, ”Why is Roy
hailing a cab?”, is not relevant. Therefore, a relevant
answer must do more than simply make the topic true or highly
probable.

First, an answer must address not only the topic, but
the other members of the alternative space as well. Recall
that the members of an alternative space can be enumerated by
making them explicit in the why-question: ”Why is it the
case that X, as opposed to Y, Z, etc.?”. A relevant answer
must not only make the topic true or highly probable, it must
also make the other members of the alternative space false or
less probable. Let us say, a relevant answer must ”select”
or ”favor” the topic over the other alternatives. In this
way, an answer must address not only the topic, but all the
members of the alternative space.

Second, the relevance relation does not simply require
an answer-candidate to select the topic from the alternative
space. The relevance relation determines the specific
conditions by which an answer selects the topic from.the
alternative space. As van Fraassen says, the relevance
relation ”determines what shall count as a possible
explanatory factor.” (143)

To clarify the point, let us return to Roy’s armrwaving
behavior. The topic of both Jones’ question, ”Why is Roy
hailing a cab”, and Smith’s question, ”Why is Roy’s arm

waving about?” address the same event: Roy’s arm motion.

83

However, each question differs with respect to both
alternative space and relevance relation. An answer to
Jones’ question must not only select the topic from the
alternative space, it must do so by addressing the intentions
and purposes of Roy’s actions. As such, the question
presupposes that Roy’s intentional states are causally
relevant to his bodily behavior and demands that an answer-
candidate capture the intentional character of his behavior.
Therefore, the relevance relation of Jones’ question can be
labeled ”causal-intentional.” A relevant answer to Jones’
question must select the topic from the alternative space by
addressing the intentional states which brought about Roy’s
cab-hailing behavior.

On the other hand, Smith’s question is in regards to the
neuro-physiological cause of Roy’s armrwaving behavior. As
such, a relevant answer must select the topic of Smith’s
question from its alternative space by addressing the neuro-
physiological cause of Roy’s armrwaving. Thus, we can
characterize the relevance relation of Smith’s question as
”causal—mechanical.”

In this example, Smith’s question and Jones’ question
have identical topics, in the sense that both topics describe
the same behavioral event. However, they diverge with
respect to both alternative space and relevance relation.
Consequently, answers to each question must pick out the same

topic, but from different alternative spaces and according to

84

different relevance conditions (i.e. different conditions by
which an answer is judged relevant).

What in general constitutes a relevance relation is a
difficult question. Indeed, such a question cuts to the core
of much of the contemporary controversy surrounding theories
of explanation. Establishing conditions which determine when
and how an explanans is explanatorily relevant to its
explanandum is one of the central and most debated topics in
the current literature. In principle, there may be a number
of different ways for one state of affairs to be
explanatorily responsible for another.

Nevertheless, whatever else constitutes a relevance
relation, the various types of causal relations (e.g. causal-

intentional, causal—mechanical, etc.) must certainly be

included in their number. This is the stance Michael
Scriven takes in ”Causation as Explanation”: ”Causation is
the relation between explanatory factors ... and what they

explain.” (Scriven, 11) Often, when we seek to know why one
particular state of affairs obtains, as opposed to other
possible states of affairs, we are looking for an antecedent
event which brought about or caused the present condition.

If I were to ask the question, ”Why is my window broken?”, I
presuppose not only that the window is broken, but that there
is some antecedent event which is causally responsible for
the state of my window. This antecedent event explains my
window’s being broken in that it is causally responsible for

the present state of affairs.

85

I wish to leave open the possibility of other non-causal
types of relevance relations. It is a matter of some
controversy whether there are such non-causal relevance
relations7, but this issue need not be a concern for present
purposes. Since an explication of the pragmatic theory of
explanation (including the relevance relation) is important
insofar as it allows us to understand mental causation, all
that is required is that, whatever else is counted as a
relevance relation, the ”causal-mechanical” and ”causal-
intentional” relations are to be included in their rank.

Let us sum up: In order to understand a question
expressed by an interrogative sentence, it is necessary to
ascertain the alternative space, the topic and the relevance
relation of the question. Van Fraassen expresses this idea
in the claim that why-questions can be identified with the
ordered triple <Pk, X, R>, where Pk is the topic, X is the
alternative space (or, in his terms, the ”contrast class”)
and R is the relevance relation. An answer to a question
must bear relation R to <Pk, X>. In other words, an answer
must select or favor the topic from the alternative space,
according to the conditions specified by the relevance
relation. These are the features of why—questions that must
be understood if an answer or explanation is to be given

successfully.

 

7 Cf. Chapter II, p. 6. The movement of the boy’s
balloon can be given an explanation based on the principle of
equivalence regarding acceleration and gravitational fields.
It is not clear whether this explanation is causal.

86

3. Explanatory Context: On What We Need To Know in
Order To Be Puzzled

As Kim claims, the epistemic goal of explanation is to
relieve a subject from a state of puzzlement, where such
puzzlement is expressed in the form of a why-question.
However, as was discussed in chapter II, there are ways of
relieving puzzlement which do not answer a subject’s
question. A good television show or a drug is enough to
relieve a subject’s puzzlement by making him forget both the
question and the feeling of perplexity or puzzlement. But
neither television shows nor drugs count as good
explanations. On the other hand, puzzlement can be relieved
by answering the question which expresses the subject’s
puzzlement. Such answers are explanations and it is this
method of puzzlement-removal which is our central concern.

Hintikka’s analysis of wh-questions allows us to discern
this method from the former, bogus methods of puzzlement-
removal: For every such question, there is a presupposition
and an epistemic goal. The presupposition is the epistemic
condition necessary for the asking of a particular question
(i.e. the conditions necessary for being in such a state of
puzzlement). For instance, in order to ask, ”Why did Mike
Tyson throw the fight?”, it is necessary to know that Mike
Tyson threw the fight (as well as what it means to ”throw” a
fight). The explanatory goals of a question are the
epistemic state of affairs which obtain when a question has

been successfully answered. Thus, the epistemic goal of ”Why

87

did Mike Tyson throw the fight?” would be to know why Mike

Tyson threw the fight.

 

 

 

 

 

 

 

Explanation
presupposition l T epistemic goal
“I know Mike answer _ “I know why Mike
Tyson threw the ’ Tyson threw the
fight” fight”
Figure 3

Notice, however, that on this analysis of puzzlement,
there is much that one needs to know in order to be puzzled.
To ask a question, ”why is it the case that X?”, one needs to
know ”It is the case that x.” One needs to know something
about what she does not know.

Recall the example of Roy’s hand—waving behavior. Jones
asks, ”Why is Roy hailing a cab?” Jones is puzzled about
Roy’s behavior and expresses her puzzlement in the form of a
why-question. However, in order to be puzzled at all, there
is a good deal that Jones must understand about Roy’s
behavior. In order to ask such a question, Jones must know
that Roy is, indeed, hailing a cab. Generally speaking, in
order to ask such a question, the epistemic state of affairs
expressed in the presupposition of the question must obtain
for Jones.

Further, in addition to the epistemic conditions
expressed in the presupposition, Jones must also know what it

means to hail a cab. Further, hailing a cab is an

88

intentional action. In order to ask such a question (and
understand the response), Jones must be able to characterize
(or identify) Roy's behavior as intentional. Notice also
that the alleviation of puzzlement occurs when Jones’
question has been successfully answered. But, in order to be
successful, an answer must pick out the topic from the
alternative space according to the conditions specified in
the relevance relation. In this case, the relevance relation
is ”causal-intentional.” Therefore, not only are there
epistemic conditions necessary for Jones’ question, there are
also theoretical conditions. Not only must Jones know that
Roy is, in fact, hailing a cab, Jones must also have an
ontology which includes cabs and acts such as ”hailing,” and
know that Jones has intentional states which are capable of
causing purposive action.

The upshot of these considerations is that epistemic
puzzlement cannot occur in a theoretical vacuum. On the
present conception of why—questions and the conditions by
which such questions are answered, the asking of why-
questions necessarily makes both theoretical and epistemic
presuppositions. Such a claim is similar to the now common
idea that experience, in general, is theory laden.

Puzzlement, question—asking and question—answering are also

89

theory-ladena.

The idea that epistemic puzzlement is theory—laden can
be expressed with the concept of the explanatory context
(hereafter, e-context)9. Every case of epistemic puzzlement
and its corresponding why-question is couched in an e-
context. A why—question’s e-context consists of its
theoretical and epistemic presuppositions. In the case of
Jones’ question regarding Roy’s behavior, several such
theoretical presuppositions are made. First, certain
entities are invoked, such as macroscopic objects (e.g.
cabs). Second, causal relationships are presupposed, namely,
causal relationships between intentional states and behavior.
Further, such a causal relationship itself presupposes that
mental states exist.

Thus, generally, the e-context, consisting of such
entities as mental states and such relations as the ”causal-
intentional” relation, is invoked by Jones’ question, ”Why is
Roy hailing a cab?”. In such a way, we might say that Jones’
question draws from or is couched within that e-context.
Further, by extension, Jones’ state of epistemic puzzlement

expressed by her why-question and analyzed in terms of the

 

9 This claim is similar to assertions made by Garfinkel.

He claims that questions shift or arise, depending on what
one’s ”explanatory frame” is. The notion of the ”explanatory
frame” is analogous to my concept of the e-context.
Garfinkel gives several examples to illustrate this point,
including an example of a question that arises within an e-
context of Aristotelian physics, but cannot arise on
contemporary evolutionary biology (See Garfinkel, pp. 98-10).
9 Again, the term context is used in a non-technical
sense. The e-context is similar to van Fraassen’s idea of
the ”background theory” of a question (147).

90

epistemic gap between presupposition and the set of
explanatory goals of the question, is also couched within
this e-context. Therefore, both the asking of a why—question
and the state of epistemic puzzlement expressed in the
question are theory-laden, or e-context-dependent.

Finally, it is important to note that e-contexts may
vary with respect to theoretical complexity and detail. For
instance, the question ”Why is Roy hailing a cab?” and its
subsequent answer might be drawn from folk—psychology, which
is a relatively simple and imprecise e-context, consisting
generally of an ontology of mental states and the causal-
intentional relation. On the other hand, such a question
might find a more detailed e-context, which, in addition to
drawing on a folk—psychological ontology, maintains a rather
sophisticated and detailed set of scientific theories. This
would be the case, for example, on a Freudian-psychoanalytic
explanation of a behavioral event.

Let us sum up this theory of explanation as it stands
presently: Broadly speaking, explanations serve the
epistemic function of removing a subject from a state of
puzzlement. This puzzlement is expressed in a why-question
and can be understood as the epistemic gap between the
presupposition and the epistemic goal of a why—question. The
subject expresses this why-question in the utterance of an
interrogative sentence. An explainer, via features of the
situation in which the question is asked, must ascertain the

why-question expressed in the interrogative sentence. To

91

understand a why-question, the explainer must ascertain the
topic, alternative space and relevance relation of the
question.

Further, the conditions necessary for being puzzled and
for asking a why-question are contained in an e—context, from
which the question draws its ontology, as well as its
relevance relation. Thus, the explainer, in order to
understand and answer the given why-question, must, in
addition to ascertaining the specific features of the why-
question, grasp the e-context in which the why-question is
couched. Given that the e-context supplies the features of
the why-question, it also specifies the conditions by which
the question will be successfully answered. Thus, in order
to be relevant, an answer must be couched within the e-
context in which the question was asked (i.e. make the same
theoretical, ontological and epistemic presuppositions).
Such is, in broad strokes, the pragmatic conception of

explanation.

4. Objections and Replies

A. Objection 1: Bogus Relevance Relations

Recall that one of the problems with Kim’s explanatory
exclusion principle is that its application is too broad.
The arguments for excluding the intentional idiom as capable

of providing causal explanations can be applied to all of the

92

special sciences, thus making the EEP unpalatable. On the
other hand, the fact that the implications of the pragmatic
conception of explanation can be applied to those non-
physical sciences (viz. making room for their causal laws and
claims against the backdrop of physicalism) lends credence to
such a theory of explanation.

However, regardless of its broad, beneficial
application, there are two powerful objections typically
brought to bear against the pragmatic theory of explanation.
The first is given by Wesley Salmon and Philip Kitcherm. The
core of the objection focuses on the lack of restrictions
placed on the relevance relation R. According to Salmon,
without such restrictions, it is not clear how to separate
legitimate from illegitimate relevance relations: ”If R is
not a bona fide relevance relation, then [answer] A is
’relevant’ to [the topic] Pk only in a Pickwickian sense.”
(141)

To illustrate the problem, Salmon offers an example:
Smith enters an astrologer’s office, places a 20 dollar bill
on the astrologer’s desk and utters the interrogative, ”Why
did JFK die on 11/22/63?” Given the clues provided by the
situation, one would likely be able to extract the features

of the why-question Q expressed in the interrogative: The

 

” This objection can be found in summary in ”Four
Decades of Explanation." in Minnesota.$tudies_in_the
Ehilgsgphy_gfi_§gienge, vol. 13, 1989, and in full in ”Van

Fraassen on Explanation.” in Journal_of_2hilosonhx. 84. pp.
315-330.

93

topic of Q is “JFK died 11/22/63,” a likely alternative space
is {JFK died 11/22/63, JFK died 1/1/63, JFK died 1/2/63,

JFK survived 1963}, and a likely relevance relation is
”astral influence.”

Here, an answer must pick out the topic from.the
alternative space, according to the conditions specified in
the relevance relation. Such an answer for Q would
presumably be ”a true description of the configuration of the
planets, sun, moon, and stars at the time of Kennedy’s
birth.” (142) Generally, if one were to express the
interrogative, ”Why did JFK die on 11/22/63?” and this
interrogative expressed the question Q=<Pk, X, R>, as
explicated above, it seems, in principle, possible to give a
satisfactory answer to Q.

First, such an answer would be true-—it would be an easy
matter to chart the alignment of the planets and stars on the
date of Kennedy’s birth. Further, as Salmon argues, it is
reasonable to assume that astrological theory will be able to
show how A selects (or favors) the topic, relative to the
other members of the alternative space: ”Since this
explanation, like any explanation, is given ex post facto, we
must credit the astrologer with sufficient ingenuity to
produce such a derivation.” (142)

It appears possible, on the present theory of
explanation, to give a satisfactory answer to a why-question
whose e-context is astrological theory. But it certainly

seems counterintuitive to claim that the alignment of the

94

planets and stars on the date of one's birth constitutes a
legitimate explanation of the events in one’s life. It
appears that the pragmatic theory of explanation grants
explanatory legitimacy to explanations which are clearly not
legitimate. Again, the culprit, according to Salmon and
Kitcher, is the relevance relation. If proper limitations or
conditions were placed on the relevance relation R, then such
explanations (involving such relations as astral influence)
could not be said to bear an explanatory relation to the
topic and alternative space. They would be ruled out in
principle.

The burden of proof, according to Salmon and Kitcher,
falls on van Fraassen to give a general account of the
relevance relation which allows us to rule out illegitimate
relevance relations, such as astral influence in astrological
theory: ”As many philosophers have insisted, we need to
appeal to objective nomic relations, causal relations, or
other sorts of physical mechanisms if we are to provide
adequate scientific explanations.” (145) In short, if we
want a scientific theory of explanation, then the pragmatic
theory must provide conditions by which non—scientific why-
questions (viz. having non-scientific relevance relations)

and their answers are excluded from consideration“.

 

"Salmon and Kitcher assume that the demarcation between
legitimate and bogus explanation is science. A good theory
of explanation will, ultimately, be a good theory of
scientific explanation. While it is not clear to me that
this is true, I will not dispute the point, since the central
concern here is to legitimize explanatory claims made by a
science of the mind.

95

However, as Salmon and Kitcher point out, determining
what exactly those conditions are is a matter of extreme
difficulty. Let us explicate the problem using the
vocabulary we have developed thus far. On the present
theory, why-questions draw relevance relations from.the e-
contexts in which they are couched. Again, an e—context can
be understood, in most general terms, by its ontology and
relevance relations. As such, astrology forms an e-context
for the above why-question, expressed in the interrogative,
”Why did JFK die on 11/22/63?” But it is highly doubtful
that e—contexts such as astrology provide bona fide relevance
relations.

The problem is, on the pragmatic theory of explanation,
there are no mechanisms by which such bogus relevance
relations (or, more generally, bogus e-contexts) are excluded
as incapable of providing legitimate explanations. The topic
is true and, within the e-context of astrology, an answer A
which bears the relevance relation ”astral influence” to the
topic and alternative space is forthcoming. After all, say
Salmon and Kitcher, ”to the sincere believer in astrology the
configuration of heavenly bodies at the time of Kennedy’s
birth is highly relevant to his death on that particular
fateful day in 1963.” (145) E-contexts which invoke
astrological theory appear to relieve the puzzlement of those
subjects who believe in astrology. Thus, the pragmatic
theory of explanation appears to include explanations that

are illegitimate. There are no devices built into the theory

96

which exclude obviously bogus explanations and explanatory
contexts, nor are there any such devices readily forthcoming.
On such a theory of explanation, how does one
differentiate between scientific and non-scientific
explanations? Van Fraassen anticipates this type of
objection and replies:
To call an explanation scientific, is to say nothing
about its form or the sort of information adduced, but
only that the explanation draws on science to get this
information (at least to some extent) and, more
importantly, that the criteria of evaluation of how good
an explanation it is, are being applied using a
scientific theory. (155-156)
To put this response in the present terminology, whether or
not an explanation (as an answer to a why-question) is
legitimate (i.e. scientific) is not determined by any
conditions placed on any feature of the explanation,
including the relevance relation. On the contrary, whether
or not an explanation is scientific is determined by whether
or not the e-context, from which both the question and the
answer are drawn, is regarded as a scientific e-context.
Specifically, is the theory (or theories) which couches the
why—question and its answer—candidates, and thus forms the
question’s e-context, regarded as a scientific theory?
Salmon and Kitcher acknowledge this response, but seem
to find it inadequate, although their reasons are somewhat
obscure. It appears that their rejection of this reply is

based in the idea that an adequate theory of explanation will

be a theory of scientific explanation. They require that the

97

theory of explanation itself set up conditions by which we

can judge explanations scientific or non-scientific:

Unless we can impose the demand for objective relevance
relations, we cannot arrive at a satisfactory
characterization of scientific explanation ... We need
to appeal to objective nomic relations, causal
relations, or other sorts of physical mechanisms if we
are to provide adequate scientific explanations. (145)
We must be quite clear that it is scientific explanation
with which we are concerned. The term ”explanation” is

used in many ways that have little or nothing to do with
scientific explanations. (6)

Van Fraassen’s response to Salmon and Kitcher’s objection is
that it is not the job of a theory of explanation to judge
whether or not a particular explanation is scientific. Such
a judgment is made of the e—context itself: Is the e—
context, in which the why-question and its answer are
couched, a scientific e-context? Is the e-context
constituted of presently-accepted scientific theories?
Salmon and Kitcher seem to reject this response, since (as
they see it) it does not provide them with an adequate theory
of scientific explanation. Such a theory should be able, on
its own, to exclude non—scientific explanations.

But such a requirement must be in vain. Such a
condition on the relevance relation, which excludes, for
instance, astral influence as a bona fide relevance relation,
must be misguided, since astrology might very well have been
true. It is well within the realm of possibility that there

be a causal relation between the ”alignment” of the stars and

98

planets at the time of one’s birth and the resulting course
of one’s life”.

Further, such a requirement presupposes that what counts
as good science is static. If the conditions by which an
explanation is to be judged scientific must be contained in
the theory of explanation itself, instead of simply requiring
explanations to draw from currently accepted scientific
theories, then what counts as good scientific theory cannot
change, without changing the theory of explanation.

But certainly what counts as good science does change,
without a corresponding change in our theory of explanation.
Such theorists as Thomas Kuhn and Larry Laudan have pointed
out that what counts as good science is, at a variety of

levels, subject to change:

Science offers the remarkable spectacle of a discipline
in which older views on many central issues are rapidly
and frequently displaced by newer ones. ... Moreover,
change occurs at a variety of levels. Some of the
central problems of the discipline change; the basic
explanatory hypotheses shift; and even the rules of
investigation slowly evolve” (Laudan, 4-5)

Kuhn’s idea of the scientific revolution as a shift
between incompatible scientific paradigms implies the

possibility of change in what counts as good science.

Theories such as Laudan’s and Kuhn’s illustrate the dynamic

 

” Perhaps, such a causal relation would qualify as

action at a distance, but the possibility of such causal
relations has arisen in contemporary quantum mechanics.
Action at a distance, therefore, does not exclude such causal
relations as impossible.

99

nature of scientific investigation. What counts as good
scientific theory, and thus good scientific explanation, is
subject to change. Hence, any attempt to place restrictions
on relevance relations will be in vain, since a change in
what counts as good science might well involve a violation of
such restrictions.

On the pragmatic theory of explanation, there are two
forums for judging the adequacy of explanations. First, an
answer-candidate is judged to be adequate at the level of the
e-context in which the why-question is couched. Questions of
explanatory adequacy that take place within the e-context,
according to the conditions established by the why-question
via the relevance relation, can be called (in Carnap’s terms)
”internal” questions”.

Regardless of whether the answer-candidate passes muster
from.within the e—context, one might question the legitimacy
of an explanation by challenging the relevance relation of
the why—question, to which the explanation is an answer. A
challenge to the e—context of astrology as incapable of
providing adequate explanations would fall into this

category. Such questions which challenge the legitimacy of

 

” While the ”internal/external” distinction used here

closely resembles Carnap’s distinction, they are not
identical. Cf. ”Empiricism, Semantics, and Ontology” in
Meaninmniflecessitx (1956: University of Chicago Press)
pp. 205-221.

100

the e-context in which explanations are given can be termed
”external” challenges or issues“.

This is, I believe, the basis of van Fraassen’s reply to
this type of objection and I believe it is fundamentally
correct. The pragmatic theory of explanation is a theory of
explanation in general. It does not have the resources to
adjudicate between scientific and non-scientific
explanations, nor ought we expect it to. That is not its
job. To question whether an explanation is scientific is to
challenge the theories of the e-context in which why-
questions and their subsequent answers are couched. Salmon
and Kitcher are correct: We need to be able to adjudicate
between scientific and non-scientific explanations. But such
adjudication does not (and cannot) take place within the

theory of explanation.
B . Objection 2 : Explanatory Relativism

A second common objection to pragmatic theories of
explanation regards the central role played by the subject in
such theories. The importance of the subject in explanation
has long been recognized, but most theorists attempt to

downplay or eliminate the role of the subject from theories

 

“ We might ask such an external question of the
intentional idiom. Is the intentional language a legitimate
(i.e. scientific) e-context or is it more closely akin to the
likes of astrological theory? This question shall be
addressed in chapter IV.

101

of explanation. The problem, as philosophers such as Hempel

and Gasper show, is relativism:

To explain something to a person is to make it plain and
intelligible to him, to make him understand it. Thus
construed, the word 'explanation' and its cognates are
pragmatic terms: their use requires reference to the
persons involved in the process of explaining. In a
pragmatic context we might say, for example, that a
given account A explains fact X to person P1. We will
then have to bear in mind that the same account may well
not constitute an explanation of X for another person
P2. ... Explanation in this pragmatic sense is thus a
relative notion: something can be significantly said to
constitute an explanation in this sense only for this or
that individual. (Hempel, 425-426)

Philip Gasper shares Hempel's concern over making sense of
the role of the subject in explanation:
A satisfactory explanation of an event or phenomenon
should provide us with understanding of what has been
explained (the ”explanandum“). But understanding is a
notoriously vague and subjective state. Different
inquirers may disagree about what is sufficient for
understanding and whether or not understanding has
actually been achieved. If the search for explanation
has a central role in scientific reasoning, then it is
important to ensure that our concept of explanation is
free from this kind of vagueness. (Gasper, 289)
Does taking the subject to be a central feature in
explanation theory entail that explanatory adequacy will be
relative to the subject seeking the explanation? If so, then
the pragmatic theory may be lost in radical relativism, which
will effectively preclude its being able to provide an
analysis of objective explanation. The question is, what

exactly is the role of the subject in determining explanatory

adequacy and does such a role entail relativism?

102

On the present analysis of the pragmatic theory of
explanation, there are two ways in which the subject seeking
the explanation plays a crucial role. First, an explanation
is adequate only if it addresses the reason or reasons for
which the subject seeks an explanation.

Recall, this was the difficulty illustrated by the
example in which both Smith and Jones inquire after Roy’s
arm-movement behavior. An answer to Jones’ question, ”Why is
Roy hailing a cab?”, based in the physical idiom is
inadequate since it doesn’t address the reason or reasons for
Jones’ asking the question. It will not move the subject
from the presupposition to the epistemic goal of the
question. The conditions by which such movement is
accomplished are determined by the features of Jones’ why-
question, including the question’s relevance relation--the
”causal intentional” relation. A physicalist answer fails to
remove the subject, Jones, from.her state of puzzlement, so
it is irrelevant.

Notice, however, that while the subject plays a crucial
role in explanatory adequacy, in that the answer must be
relevant to the subject’s question, the criteria by which an
answer can be said to be relevant (i.e. address the reasons
for which the subject seeks an explanation) is determined by
the relevance relation. The relevance relation is drawn from
a particular e-context. Again, the legitimacy of the
relevance relation can be scrutinized by asking whether the

e-context from which it is drawn consists of bona fide

103

scientific theory (this is the ”external” question). E-
contexts and their relevance relations are not subjective.
They are theoretical entities subject to public criticism.
Since it is the e-context that couches the why-question and
it is the features of the why-question which determine the
conditions by which a satisfactory answer can be given, such
conditions are subjective only if e-contexts are subjective.
But they are not.

Explanations serve the epistemic goals of the subject.
Thus, the subject selects the why-question and so determines
the e-context from which the question is both understood
(both by explainer and explainee) and answered. However,
once the selection of the e-context is made, the conditions
by which an answer is judged adequate are determined by the
e—context in which the why-question is couched. The subject
selects an e-context, but it is the e-context that determines
the conditions for explanatory adequacy (viz. the relevance
relation). These conditions, in turn, are subject to
external public scrutiny. Thus, there is no relativism in
the conditions of explanatory adequacy.

Explanatory adequacy is objective in the sense that
scientific e-contexts are public domains of inquiryu The
theories, entities and relevance relations that make up an e-
context are not private languages. Thus, the conditions for

explanatory adequacy, determined by the e—context, cannot

104

vary from subject to subject”. Which e-context is selected
may vary from the subject to subject, but once the selection
is made, the conditions for explanatory adequacy are
determined by the elements of the e-context. Such a role for
the subject does not entail relativism.for the pragmatic
theory of explanation.

The second manner in which the subject plays a crucial
role in explanation theory tends to be a bit more tricky: An
answer must not go beyond the scope of the subject’s
knowledge. An answer might be relevant in the sense that it
meets the conditions specified in the relevance relation, but
still go beyond the scope of the subject’s knowledge. Thus,
an answer might meet all of the conditions for adequacy, but
still not move the subject to the explanatory goals of the
question.

For instance, a first-year student of physics asks a
professor to explain why objects cannot move faster than the
speed of light. Clearly, the e-context from.which the
question is asked involves theories of space-time physics and
the professor is able to answer this question. However, the
professor knows that the best possible answer, with all the
mathematics involved, will not be grasped by a first-year

student of physics. The student has enough of a grasp on the

 

” This is not to say that there is no room for dissent

in publicly constructed theories, such as those theories
found in science. Lauden gives a very nice account of
dissent in the scientific community, as well as the process
by which consensus is formed. See Lauden, Sgienge_and
yalues, U. California Press: 1984.

105

e-context to be able to ask the question, but not enough to
understand the best possible answer.

The problem is, simply because a subject is able to ask
a question from within a particular e—context does not imply
that he has enough of a mastery of that e-context to
understand an appropriate answer. Certainly, there are
varying degrees of understanding one might have, with regard
to an e-context. Thus, it appears that the conditions for
understanding an answer will vary from subject to subject,
depending on the level of mastery a subject has over the e-
context in which the question is couched.

This problem illustrates that maintaining the subject as
a central feature of explanation theory makes such a theory
complicated, but it does not introduce an ”anything goes”
brand of relativism into the picture. In most cases, an
explainer is able not only to ascertain the e-context from
which the question is asked, but also the level of mastery
the subject has over that e-context. The physics professor,
for instance, recognizes that the subject desiring the
explanation is a first-year student and thus does not as yet
have the background to understand the most complete,
sufficient answer to the question.

However, an explainer can give an answer which, although
not the best answer possible, still selects (or favors) the
topic from the alternative space in the manner specified by
the relevance relation. The answer draws on the ontological

and theoretical machinery of the e—context, but in a broader,

106

more vague manner. This type of explanation occurs
frequently, and not simply in academic settings. Doctors,
for instance, answering questions such as, ”Why do I need
surgery?”, are able to give an adequate answer without
invoking the technical vocabulary necessary to give the best
answer.

This may open the door to a variety of possible answers,
each varying from subject to subject with respect to
complexity and detail, but we are able to avoid relativism in
two ways: First, all the possible answers still draw on the
e-context from.which the question is asked and, thus, they
meet the conditions by which explanations are judged
adequate, which are determined by the e-context, rather than
the subject. Second, in cases where answer A is regarded as
the best answer, but the subject does not understand answer
A, we can still claim that, where A is the best answer to
question Q, within e-context C, if subject S were to have a
sufficient.mastery of e-context C, from which question Q is
asked, then he would understand answer A.16 In cases where a
subject’s puzzlement is not removed because the subject lacks
the requisite mastery of the e-context, and thus does not
understand an answer, it can still be claimed that he would

understand the answer if he had sufficient mastery of the

 

” Notice, this is not intended to be a definition of

the best answer. This claim merely allows us to conceive of
the idea of A being the best answer to question Q (relative
to the e-context), even though subject S does not understand
A. Such an answer would be appropriate if S were to have
sufficient mastery of the e-context of Q.

107

e—context which couches the answer.

Therefore, we are able to maintain this second
subjective condition of explanation, namely, that an answer
must not go beyond the scope of the subject’s knowledge,
without committing ourselves to a relativism of the
conditions by which answers are judged adequate (or better or
worse). So the subject can maintain a central role in the
pragmatic theory of explanation without such a theory finding

itself mired in relativism.

5. Example and Summary

Before we move on, let us sum up the theory as it
stands. The epistemic function of explanation is to remove a
subject (i.e. an explainee) from a state of epistemic
puzzlement. This state of puzzlement is expressed in a why-
question, in which puzzlement is understood as the epistemic
gap between the presupposition and the epistemic goal of the
question. An explainer, in order to remove the subject from
his state of puzzlement, must understand the why-question
expressed in an interrogative sentence. Such an
understanding is accomplished, via clues provided by the
situation in which the question is asked, when the explainer
ascertains the topic, alternative space and relevance

relation of the subject’s why-question.

108

However, epistemic puzzlement cannot occur in a
theoretical vacuum. Ontological, theoretical, and epistemic
presuppositions are necessary for the possibility of
epistemic puzzlement. These presuppositions, expressed in
the above features of why-questions, draw from a particular
set of theoretical commitments—-commitments which can range
from extremely broad, such as general relevance relations and
ontology, to extremely detailed, such as specific scientific
theories.

Such theoretical commitments form the e-context of the
why-question. The e-context, as well as the specific
features of the subject’s why-question, can be ascertained by
clues provided by the situation in which the question is
asked. Answers to why-questions are explanations and such
answers, in order to succeed as explanations, must be
relevant to the question asked. The conditions for
explanatory relevance are determined by the relevance
relation of the question. Such relevance relations are drawn
from the e—context in which the question is couched. Thus,
in order to be relevant, an answer must ultimately be
relevant to the e-context in which the why—question is
couched. Answers that are not relevant to the e—context of a
question fail, since they do not relieve the subject’s
puzzlement (i.e. they do not move the subject from the
presupposition to the epistemic goal of the why-question).

At this point, it would be instructive to demonstrate

how an example from scientific practice fits this model of

109

explanation. Such an example will be helpful for two
reasons: First, by being grounded to a particular situation,
it can be shown that the pragmatic theory of explanation is
able to demonstrate its applicability--that it is useful in
understanding how explanations work in scientific practice”;
second, an example will show that seeking and finding
explanations is often a rather complicated procedure. Such
complications, as will be shown, can be handled on the
pragmatic model.

The pragmatic theory presents a sort of story about when
explanations are required and when those requirements are
met. The story begins with a subject encountering a
phenomenon. That phenomenon is a source of puzzlement for
the subject, in that she knows that the phenomenon occurred,
but not why. The search for an explanation is a search for a
resolution to this puzzlement. The end of the story is
generally the resolution of the puzzlement through the
acquisition of an explanation of the phenomenon. Such a
story appears linear. We begin with puzzlement and end with

resolution.

 

17

That is not to say that the pragmatic theory is
universally applicable, nor is its value found purely in its
application. It is not yet clear that pragmatic theory can
account for every explanation we regard as successful.
However, even if there are anomalies that pose challenges to
the present theory (and I’m sure there are), its value is at
least shown in allowing philosophers to cope with situations
where there are multiple complete and independent
explanations of a single explanandum event. As Kim argues,
it is not clear that the standard D-N models of explanation
can cope with such a phenomenon.

110

In science, however, the situation can be more dynamic.
Often, scientists have answers (in the form of hypotheses
about the world) but search for questions to ask. In other
words, a scientist’s theory or hypothesis is vindicated by
how well it solves problems or resolves puzzlement. The more
questions a theory or hypothesis can answer, the more
valuable it is. Therefore, it often occurs in science that
the story is backwards-~the process moves from.answer to
question. Richard Feynman phrases it this way: ”We
physicists are always checking to see if there is something
the matter with the theory. That’s the game, because if
there is something the matter, it’s interesting!” (Feynman,
8) Scientists seek to be puzzled. Once they are puzzled,
they can discover how well their theories work, or how much
work their theories need. Let us take an example, of such a
situation and show how it fits the pragmatic model of
explanation”.

In the early 19th century, physicists hypothesized (had
a hunch, really) that the phenomena of electricity and
magnetism were somehow related. Both phenomena were well
documented and understood, but there were as yet no
demonstrations of any relation between the two. As George
Gamow says, ”Electric charges did not influence ... magnets

in any way; neither did ... magnets influence ... electric

 

“ The following example comes from George Gamow’s The

Wu. (Dover
Publications: 1961) pp. 135—137.

111

charges.” (Gamow, 135)

It is the physicist Hans Christian Oersted who is
credited with discovering the phenomenon that would lead to
the development of electromagnetic theory—-the unified theory
of magnetism and electricity. Oersted’s experiments involved
an electric current generated by a Volta pile19 and a compass.
Similar experiments prior to Oersted’s generally involved
interaction between a compass and a static electric charge.
Instead of using static electricity, Oersted connected a wire
to the two poles of the Volta pile, thus creating a
directional current of electricity. He then placed the
compass near the wire: ”The needle, which was supposed to
orient itself always in the north—south direction, turned
around and came to rest in the direction perpendicular to the
wire.” (ibid. 135) When the Volta pile was disconnected, the
compass again pointed towards magnetic north. Once the wire
was connected to the Volta pile, the compass again turned
away from magnetic north to become perpendicular with the
wire.

Here was an anomalous phenomenon. Present, accepted
theories of electricity and magnetism could not account for
it; for, at this point, electromagnetic theory was only a
hunch, much less a fully developed scientific theory. Hence,

Oersted found himself in a state of puzzlement, expressed by

 

” A ”volta pile” is basically a battery, consisting of
”alternating copper and iron or zinc discs, separated by
layers of cloth soaked in salt solution.” (Gamow, 133)

112

the question, ”Why did the compass needle turn perpendicular
to the wire?”. The topic is the event of the needle turning
perpendicular to the wire, away from.magnetic north. Given
what Oersted knew about magnets, a reasonable alternative
space for the question would consist of {the needle turns
perpendicular to the wire, the needle remains pointing
towards magnetic north}. Thus, a successful answer must
select the topic from the alternative space, demonstrating
why the needle turns perpendicular to the wire, as opposed to
pointing at magnetic north, as is expected.

At this point, Oersted's hunches about the relation
between electricity and magnetism are rather vague (since his
was the first proof that there was such an interaction). So
the relevance relation would be a broad causal relation:
something is causing the compass to turn away from magnetic
north.

Therefore, for question Q, ”Why did the compass needle
move perpendicular to the wire?”, Q is identical to the
ordered triple <X, Pk, R>, where the topic, X, is the event
of the needle turning perpendicular to the wire attached to
the volta meter, the alternative space Pk is {the needle
moves perpendicular to the wire, the needle remains pointing
at magnetic north} and the relevance relation R is a causal
relation having to do with the electricity.

Recall that physicists at this time had a broad, vague
hunch that there was some sort of relationship between

magnetism and electricity. In order to resolve Oersted's

113

puzzlement regarding the phenomenon of the compass, however,

that hunch would be honed into a broad hypothesis:

It became quite clear to [Oersted] that there was an
interaction between the magnets and the.moving
electricity, and the direction in which the compass
needle was oriented depended on the direction in which
the electrical current was flowing through the wire.
(ibid. 136)

The e-context of the question, ”Why did the compass
needle move perpendicular to the wire?”, consists in this
case of the epistemic presupposition of the compass’
movement, the theories of electricity and magnetism, as well
as certain hunches about the interaction of their respective
phenomena.

Notice that this case demonstrates the claim that
puzzlement cannot occur in a vacuum: The event of the
compass' moving to become perpendicular with the wire is
puzzling only if it somehow confounds our expectations about
how the magnet should have behaved. Knowing what we do about
magnetic theory, the unfettered magnet was supposed to point
toward magnetic north. That expectation, founded in our
knowledge of how magnets work, is foiled. That is what makes
the event puzzling and, according to Feynman, exciting.

Thus, puzzlement, expressed in our why-question, is
couched within the e-context containing electrical theory
(which allows us at least to understand terms like ”current”)

and magnetic theory (which allows us at least to form.foiled

expectations about the behavior of the magnet). This

114

e-context also contains the epistemic presupposition of the
activity of the magnet and the relevance relation of the
causal interaction between electricity and magnets, which is
itself couched in the hunch of the physicists of the day.

This example demonstrates not only how puzzlement,
expressed in the form of a question, must be couched in an e-
context, but that the relationship between questions and
answers is dynamic. First, scientists do not necessarily
begin with puzzlement and seek its resolution in accepted
theory; Both new and old theory find vindication in the
resolution of puzzlement. Thus, scientists actively seek to
be puzzled, not only to see if their present theories cut the
mustard, but also to develop new theories or to test and hone
hypotheses or hunches.

Further, this example demonstrates a dynamic
relationship between the arising and alleviation of
puzzlement. Once Oersted’s original puzzlement is resolved,
new questions and puzzlements arise. For instance, Oersted
might ask, why did the compass turn perpendicular to the
electrical current, as opposed to spinning like a top, or
turning parallel to the electrical current, and so on. While
old puzzlements are resolved, new puzzlements arise. In this
dynamic interaction between why-questions and answers within
an e-context, theories are developed, honed, and vindicated.
This example demonstrates that the pragmatic theory of
explanation can not only make sense of concrete cases of

explanation-seeking, but it can also tell a story about the

115

rather dynamic relationship between why—questions and answers
in scientific inquiry.

The task at this point is to show, in more detail how
such a theory of explanation allows us to understand the role
of the mental in the production of behavior. This is the aim
of chapter IV. Specifically, we shall focus on what type of
analysis of mental causation follows from the pragmatic
theory of explanation. Further, we shall entertain an
additional objection to the idea that themental plays a
systematically irreducible role in the production of bodily

behavior.

116

Chapter Iv: Mental Causation and Explanation

1. Problems and Resolutions

In order to illustrate how an analysis of mental
causation can be generated from the pragmatic model of
explanation, let us briefly return to the original
formulation of the problem. Against the backdrop of
physicalism, there is a fundamental tension between the
irreducibility and the causal efficacy of the mind. The
principles of physicalism dictate that mental states and
their properties are determined by physical states of the
body. However, given arguments that demonstrate the
irreducibility of mental states (such as multiple
realizability) such mental states, properties and powers
cannot be systematically identified with their physical
realizations. These claims are expressed in the DBNI thesis,
where the properties of mental states are ”determined by but
not identical to” physical states.

The causal power of mental states is an example of a
type of mental property that can be subsumed under the DBNI
thesis. To claim that a mental state M is sufficient for the
production of bodily behavior E implies that the physical
realization P of M is also a sufficient cause of E; for, if
M were sufficient for E and P were not, then M would have a
causal property that P does not, namely, being sufficient for

E. This would contradict the DBNI thesis. Thus, one

117

encounters a situation in which there are two distinct
events, M and P, each of which is a sufficient cause of E.

On such a view, at least two problems emerge. First, if
P is a sufficient cause of E, then it is not clear what
causal work remains for M to accomplish. How does one make
sense of M as a cause of E? What work is there for M to do
that is not already accomplished by P and how is such work to
be distinguished from that done by P?

A second, stronger problem is causal overdetenmination.
Recall that causal overdetermination does occur, but cases of
overdetermination present special problems for causal
analysis. If M and P are each separate, sufficient causes of
E, then it is unclear how a causal analysis of E should
proceed. Since both M and P are sufficient causes, each
renders the other unnecessary. We cannot, therefore, claim
that P caused E, since E would have occurred whether P had
happened or not. P is superfluous to E. The same claim can
be made of M. If M and P are each superfluous to E, then it
is unclear what caused E. E has a cause, but a causal
analysis of E is at best problematic.

Again, instances of causal overdetermination do occur,
but as oddities or anomalies. But according to the DBNI
thesis, every intentional act is causally overdetermined. In
other words, the DBNI thesis implies systematic
overdetermination, thus rendering causal analyses of all
intentional behavior problematic. A primary goal of

contemporary philosophy of mind, recall, is to assimilate the

118

mind into our scientific world—view. But one of the central
goals of science is to ascertain and make sense of the causal
structure of the world. To conceive of the mind as
systematically irreducible to the brain, and yet causally
responsible for behavioral events appears counterproductive
to that goal, in that an understanding of the causes of
”intentional” behavior is made obscure and problematic by
systematic overdetermination. Therefore, if the DBNI thesis
does indeed imply systematic overdetermination, then it is at
best problematic and at worst absurd.

So any cogent analysis of mental causation must meet two
goals: First, such an analysis must make room for and make
sense of the unique causal work done by the mental in the
production of behavior, while retaining the claim that
physical events are capable of providing sufficient causes of
such behavior; second, it must avoid the pitfall of
systematic overdetermination. The pragmatic theory of
explanation provides an analysis of mental causation which
achieves both of these goals.

First, how is the unique causal work of the mental to be
understood, given that the physical is capable of providing
sufficient causes of any behavioral event? An analysis of
and answer to this first problem (which was begun in chapter
III) can be given by the pragmatic theory of explanation.
Recall that the implications of physicalism, namely, that any
behavioral event is subject to an exhaustive causal analysis

within the physical idiom, can be understood to mean, first,

119

that for any behavioral event, there is a why-question
regarding that event which requires an answer from within the
physical idiom; second, any such question can be given a
correct answer from.within that idiom.

What is not implied by physicalism is that, for any
behavioral event, any why-question regarding that event can
be answered in the physical idiom, Why-questions whose
relevance relations are causal-intentional, for example,
demand an answer/explanation of a behavioral event which
refers to the mental states of the subject performing the
behavior. Such is the case with the example of Roy’s hand-
waving. Jones’ question, ”Why is Roy hailing a cab?”
characterizes Roy’s behavior as purposive or intentional,
thus requiring an explanation which refers to his mental
states .

The unique causal work done by mental states in the
production of behavior can be understood through an analysis
of the role such states play in explanations of behavior.
The causal analysis of behavioral events which involve
irreducible mental states is capable of providing answers to
why-questions which cannot be answered by reference to
physical states. In other words, in particular sorts of
cases, the mental idiom is capable of providing causal
explanations where the physical idiom fails to do so.

One might object that this seems to imply a sort of
explanatory failure for the physical idiom, One of the

implications of physicalism is that the physical idiom is not

120

subject to explanatory failure. That is, the principles of
physicalism dictate that every behavioral event has a
sufficient physical cause. So, hasn’t the pragmatic theory
of explanation, as an analysis of the role of the mental in
causing behavior, somehow violated the principles of
physicalism?

On the contrary, this analysis of mental causation
suggests that there are two distinct types of explanatory
failure. By way of demonstration, let usreturn to the
example of Roy's hand—waving. Recall Jones' question, ”Why
is Roy hailing a cab?”, where such a question has a causal—
intentional relevance relation. On the other hand, Smith’s
question is ”Why is Roy's arm moving about?”, where such a
question has a causal—mechanical relevance relation.

Suppose, first, that Roy’s behavior is not intentional--
that it is a spasm or seizure caused by a neurological
condition. In this case, any intentional characterization of
Roy's behavior (as the topic of a question) will be false.
Recall, however, that in order for a question to be
answerable (i.e. to admit of a right answer) its topic must
be truel. So, in the present scenario, any why-question who’s
topic or presupposition characterizes Roy’s behavior as
intentional cannot be answered since such a characterization

is false.

 

1 Similarly, in Hintikka's terminology, the claim.can be

made that in order for a wh-question to admit of a correct
answer, it must have a true presupposition. See chapter II,
p. 8.

121

In this example, there is a failure to explain in a
rather broad sense. The failure is a sort of breakdown of
the e-context itself. The intentional e-context (maintaining
an ontology of mental states and the causal—intentional
relevance relation) is incapable of describing Roy's behavior
as the topic of a why~question. Thus, any such why~question
couched within the intentional e-context cannot be answered,
since its topic is false. The e-context itself is incapable
of formulating an answerable why-question regarding Roy's
behavior, since it is incapable of describing his behavior.
Therefore, no explanation of Roy’s behavior can be given from
this e-context. Such broad explanatory failure we might call
a failure of description, since the e—context is incapable
of describing the event.

On the other hand, suppose that Roy’s behavior is
subject to an intentional description. In particular, let us
say that the topic, ”Roy is hailing a cab” is true. Since
the topic (and hence the presupposition of the why-question,
”Why is Roy hailing a cab?”) is true, the intentional e-
context is (ceteris paribus) capable of providing an
explanation of Roy's behavior. Since the relevance relation
of the question is causal-intentional, any non—intentional
answer, including any answer from the physical idiom, will be
irrelevant. Here, the physical idiom fails to explain in
that any physicalist answer to the question ”Why is Roy
hailing a cab?” will be irrelevant. This narrower type of

explanatory failure can be labeled, failure of relevance.

122

Physicalism implies that the physical idiom cannot be
subject to a failure of description. For any behavioral
event, a why-question, couched in the physicalist e-context,
whose topic is true, can be asked and answered. Such a claim
cannot be made for the intentional idiom. Thus, the
intentional idiom can be subject to a failure of description.
The physical idiom, however, is not subject to such failure.

While physicalism implies that the physical idiom is not
subject to a failure of description, there are no such
implications of physicalism for failure of relevance. The
claim that there are cases in which any answer from the
physicalist e—context will be irrelevant to a question asked
from the intentional e—context is compatible with
physicalism, since a failure of relevance does not imply a
failure of description.

Hence, there are two general types of explanatory
failure that involve a failure of the e-context itselfz.
First, if e-context A is incapable of providing an answer to
question Q because it cannot provide an answer which meets
the conditions of the relevance relation of Q, then A has
suffered a failure of relevance. Since A is incapable of
providing an answer to Q, it is the e-context itself that has

failed.

 

2 There is, of course, the possibility that an
explanation could fail because it is false. This type of
failure, however, does not entail a failure of the e-context.
It simply implies that the answer given is not correct.

123

Secondly, if e-context A is incapable of describing or
characterizing an event as the topic of a question, then the
e-context fails to explain in that it cannot formulate an
answerable question which has that event as its topic. This
too is a failure of the e-context. Such a failure is a
failure of description, since the e-context is incapable
of describing the event as the topic of a why-question.

Thus, we are able to answer the first problem of mental
causation. First, as was argued in chapter II, causation can
be understood and analyzed in terms of explanation. 'A
causes 8’ can be understood in terms of 'A explains B.’
Second, the pragmatic theory of explanation allows us to
understand the unique causal work done by the mental. In
other words, the pragmatic theory allows us to make sense of
the causal efficacy of the mental without violating the
principle of physicalism. For on the pragmatic theory, a
causal analysis of a behavioral event which involves mental
states can provide causal explanations that a physical
analysis cannot. A why—question whose relevance relation is
causal—intentional demands an answer which draws from the
mental states of the subject performing the behavior. Since
the physical idiom is in principle incapable of answering
such a question (i.e. the physical idiom is subject to a
failure of relevance), the mental idiom is capable of
providing causal explanations of behavior that the physical

idiom cannot. In this way, we are able to understand the

124

unique work done by the mental in production of behavior.
Further, since such a failure of relevance by the physical
idiom.does not imply a failure of description, this analysis
of the role of the mental in the production of behavior in no
way violates the principles of physicalism”

But what of the second problem; namely, that of causal
overdetermination? Let us suppose a behavioral event E,
which is subject to an intentional explanation M and a
physical explanation P, both of which are seen as irreducible
and sufficient in the production of E. The problem in cases
of overdetermination is that there is causal competition
between M and P. Since each is a sufficient cause of E, each
renders the other unnecessary for E, making a causal analysis
of E highly problematic.

But how is causal overdetermination, and thus this
notion of causal competition, to be understood within the
pragmatic theory of explanation? Recall Kim's notion of
causal overdetermination: Overdetermination occurs when
there are two causal explanations, P and M, for the same

event, E9. On the present analysis of explanation, E is an
event to be explained in virtue of its being the topic of a
particular why—question. The parameters of the why-question
(i.e. the alternative space and the relevance relation)
determine the conditions by which an explanation is deemed

successful.

 

3 The fact that M and P are each complete and independent
explanations of E is presupposed. In other words, neither M
nor P reduce to the other.

125

Thus, to claim that M and P are each successful
explanations of E is to claim that they each satisfy the
conditions specified by the why-question. In other words,
in order for M and P to be successful explanations of E, they
would each have to be successful answers to a particular why-
question, whose topic is E. Therefore, on the pragmatic
theory of explanation, to have multiple explanations of a
particular event is to have multiple satisfactory answers to
a particular why-question.

In such a way, we are able to understand the causal
competition inherent to overdetermination: Multiple
explanations compete only if they are each successful answers
to the same why-question. Again, overdetermination involves
some sort of causal competition. On the pragmatic theory,
there is such competition only if there are multiple answers
to a particular why-question from within the same e-context.
Such is the case with Kim’s examples:

A man is shot dead by two assassins whose bullets hit

him.at the same time; or a building catches fire

because of a short circuit in the faulty wiring and a

bolt of lightning that hits the building at the same

instant. It isn't obvious in cases like these just how
we should formulate an explanation of why or hOW’the

overdetermined event came about. (Kim.II, 252)

In each example, there are multiple causal analyses of a
particular event. But on the pragmatic theory of
explanation, having multiple causal analyses does not, by

itself, entail overdetermination. There is causal

competition amongst these analyses only if they each serve as

126

answers to a single why—question which has the event in
question as its tOpic.

”Why is this man dead?” can be sufficiently answered by
reference to either assassin's bullet. ”Why did the house
burn down?” can be sufficiently answered by reference either
to the lightning strike or to the faulty wiring. In each
example, there is causal competition because there are
multiple answers to each question. Each causal analysis,
since each is mutually exclusive, rules out the other.

Broadly speaking, such causal competition occurs when an
e-context offers multiple answers to a particular question.
Thus, on the pragmatic theory of explanation, causal
overdetermination can be understood as having multiple
answers, within a particular e-context, to a particular why—
question.

Further, this analysis of overdetermination can also
make sense of Kim's claim that causal overdetermination
produces an ”epistemic predicament” for the subject seeking
an explanation of an event. Recall Kim’s argument: A
subject is, first, in such a predicament with regards to a
particular event. An explanation of the event will relieve
the subject of his or her predicament. However, where such
an event is causally overdetermined, not only does the
original predicament remain, but a new type of predicament
emerges: Which of the two explanations of the event is true?
So, according to Kim, there is an epistemic problem with

overdetermination. Instead of relieving them,

127

overdetermination compounds a subject's epistemic
predicaments about the world, which is counterproductive to
the epistemic goals of explanation.

On the pragmatic theory of explanation, the subject
expresses her epistemic puzzlement through a why-question. In
situations of causal overdetermination, there are multiple,
sufficient answers to that why-question. Kim is, therefore,
correct: Having multiple answers to a particular why-
question is counterproductive to the goals of explanation:
NOt only does the original why-question remain unanswered,
but a new question arises, namely, ”Which answer is the
correct one?” In this way, puzzlement is compounded.

However, such puzzlement is compounded only if there are
multiple answers to a particular why—question. The present
analysis of mental causation avoids this problem. The mental
and the physical answer different why-questions about
behavior. Thus, puzzlement is not compounded. So, although
Kim is correct about the epistemic problems of causal
overdetermination, such problems do not arise on the present
analysis of mental causation.

If this analysis of mental causation is accurate, then
there is no causal competition between mental states and
physical states in the production of behavior. Each causal
analysis does different causal work in that each is capable
of answering different why-questions. Where a mental
explanation answers such a why-question, a physical

explanation is irrelevant. Where a physical explanation

128

answers such a why-question, a mental explanation is
irrelevant. Each answer is given in a different e-context
and neither answer makes any commitments about the other.

Hence, the pragmatic theory of explanation is able to
make sense of the anomalous cases of causal
overdetermination, as well as the difficulties such cases
present for explanation. Further, the same theory shows that
mental causation does not imply systematic overdetermination.
By generating an analysis of the unique work done by the
mental in the production of behavior and by showing how
mental causation avoids systematic overdetermination, the
pragmatic theory of explanation is able to make room.for and
make sense of the causal efficacy of systematically

irreducible mental states.

2. Justification of the Relevance Relation

It is the unique character of the mental that allows the
pragmatic theory of explanation to make sense of the causal
work done by the mental and to show how the problem of
systematic overdetermination can be avoided. But in the
solution to these problems, there has been a major
presupposition; namely, that the mental can be regarded as
causally efficacious at all.

The primary goal of this thesis is to make sense of and
make room for the causal efficacy of irreducible mental
states, given that the physical sciences can, in principle,

offer an exhaustive causal analysis of any behavioral event.

129

To this end, it has been presupposed that there is such a
bona fide causal relation between mental states and physical
behavior. However, the argument can be advanced that such a
presupposition is, itself, unwarranted.

Recall the problem posed by Salmon and Kitcher for the
pragmatic theory of explanation: Without a priori conditions
on the relevance relation of a why-question, how are we able
to discern genuine inquiry, such as that found in the e-
contexts of science, from the illegitimate type, such as that
found in astrology? In other words, by what conditions can
genuine relevance relations, such as the causal—mechanical
relation, be demarcated from bogus relevance relations, such
as astral influence? The reply given in chapter III is that
it is not the task of a theory of explanation to discern
genuine from bogus relevance relation. On the contrary, such
a problem is an external concern, which investigates the
legitimacy of the e-context itself-~whether or not the e—
context is capable of providing adequate relevance relations.

But such a response implies another objection: If
inquiries into the legitimacy of relevance relations are
external questions, then such a question requires asking: Is
the intentional idiom capable of providing genuine relevance
relations? In particular, is the causal-intentional
relevance relation a bona fide causal relation, or is it to
be relegated to the likes of astral influence. Is the mental
genuinely efficacious, or is mental causation an anachronism

with which science will eventually dispense?

130

While it is not the central concern of this thesis to
ask and answer this external question, it appears necessary
to offer strategies by which one might demonstrate the
legitimacy of the causal-intentional relation. Philosophers
such as Jerry Fodor have offered arguments in support of the
view that the mind bears a causal relation to certain cases
of bodily behavior. In order to dispel qualms about the
legitimacy of the causal-intentional relevance relation,

these arguments will be discussed briefly.

3. Fodor and Ceteris Paribus Laws

As noted in chapter I, the question of causal efficacy
is not simply about the relation between events, i.e. whether
one event is causally responsible for another. The question
is about properties of events. A red ball, for instance,
collides with a blue ball with a certain amount of force.
Subsequently, the blue ball begins to move. It is clear that
the red ball causes the blue ball to move in virtue of its
momentum. The property red, on the other hand, is causally
irrelevant.

In general, then, events involved in causal relations
will maintain properties which are causally relevant and
properties which are not. The goal is, first, to establish a
set of criteria by which the causally relevant properties are
clearly distinguished from the causally irrelevant, and

second, to show that mental properties meet such criteria.

131

Meeting these goals will go a long way towards removing any
remaining worries over the causal efficacy of the mental.

According to Fodor, in the antecedent event, what
distinguishes the ball’s being red from the ball’s having a
certain momentum (i.e. having a certain mass and moving at a
certain rate of speed) is that the latter property is
nomologically sufficient for the blue ball’s subsequent
movement, while the former is not. Ordinarily, any other
ball with the same mass and speed, when it collides with a
stationary blue ball, will produce the same subsequent
movement. However, there are other red balls, with different
masses and/or speeds, that are incapable of moving the blue
ball. Since the ball’s redness is not nomologically
sufficient for the blue ball's movement, the redness is
causally inert. On the other hand, since the red ball’s mass
and speed are nomologically sufficient for the blue ball’s
movement, they are causally efficacious. So causal efficacy
can be understood in terms of causal law.

Not only does the present example illustrate that causal
efficacy can be understood in terms of causal law, it also
shows that events are subsumed under such causal laws in
virtue of certain properties. The event of the red ball's
colliding with the blue ball is sufficient for the event of
the blue ball’s motion in virtue of particular properties of
the antecedent event. The relevant law would cover those

antecedents which have the property of such-and-such mass and

132

speed. Thus, objects are subsumed by causal laws in virtue
of certain properties ”projected” by those laws.

Therefore, it is in virtue of a property (or properties)
that an antecedent event is nomologically sufficient for its
effect. As Fodor claims, ”P is a causally responsible
property if it's a property in virtue of which individuals
are subsumed by causal laws.” (Fodor, 143) So property P is
causally responsible for the production of events with
property E if events with property P are nomologically
sufficient for events with property B.

So, as Fodor claims, ”intentional properties are
causally responsible in case there are intentional causal
laws.” (143) For purposes of justifying the causal-
intentional relevance relation, which simply states that
certain mental properties are causally relevant to certain
behavioral events, it is enough to show that there are such
intentional laws. If Fodor is right, the establishment of
the existence of such laws, in addition to the work done in
this thesis to allay any fears instilled by physicalism (such
as systematic overdetermination), should be sufficient for
the vindication of mental causation.

However, while it is generally maintained that
intentional laws between mental events and behavioral events
do exist, it is not clear that such laws are strong enough to
support causal claims. The claim that an event can be

regarded as a cause of a second event only if the former is

133

nomologically sufficient for the latter seems to imply that

such events must be covered by strict laws. As Fodor claims,

Causal transactions must be covered by exceptionless
laws; the satisfaction of the antecedent of a covering
law has to provide literally nomologically sufficient
conditions for the satisfaction of its consequent so
that its consequent is satisfied in every nomologically
possible situation in which its antecedent is satisfied.
(Fodor, 147)
It is doubtful, however, that mental events are covered by
such exceptionless laws. Most philosophers of the mind
(including Fodor) embrace the ”anomalism” (or anomaly) of the
mental: There are neither strict psycho-physical laws
relating mental events and bodily events, nor strict
psychological laws relating mental events to one another.

An example will illustrate the point. If I perform the
behavior of walking to the refrigerator and getting a beer,
it is reasonable to assume that I might very well have done
so because I believed there was beer in the refrigerator and
I desired to drink a beer. Thus, if it can be claimed that
there are situations in which such a belief/desire pair
successfully explains beer-retrieving behavior, it must also
be claimed that the belief/desire pair regarding beer in the
refrigerator is causally related to my subsequent behavior in
virtue of being covered by a strict law: Such a
belief/desire pair is nomologically sufficient for the
production of this type of behavior.

The problem is, such a strict connection between such

belief/desire pairs and subsequent behavior will jeopardize

134

our common conception of agency, viz. precluding
deliberation, any notion of free will and the possibility of
countervailing beliefs and desires. On such a conception of
causality, behavior must follow, unfailingly, upon my belief
and desire. This rules out the possibility that I might have
beliefs and desires that overrule my belief that there is
beer in the refrigerator and my desire for beer.

For instance, I might have a drinking problem. In such
a situation, I will more than likely have the belief/desire
pair regarding the contents of the refrigerator. However,
one could easily imagine a situation in which I resist my
desire for beer, due to countervailing beliefs and desires;
namely, the belief that if I drink beer, I will fall back
into patterns of binge drinking and self-destructive
behavior, and the desire not to return to such patterns. It
is quite conceivable that, due to this countervailing
belief/desire pair, I will refrain from drinking the beer.

Not only are the capacity to deliberate and the
possibility of countervailing beliefs and desires vital to
our conception of agency, we encounter them in our daily
lives. Sometimes, one might put off the possibility of
immediate gratification for the sake of long—term.gain.
Other times, one’s immediate beliefs and desires are
victorious. I might believe that my favorite TV show is on
and I might desire to watch it. However, I have the
countervailing desire to finish my thesis, as well as the

belief that if I watch my show, I won't do the work necessary

135

to complete my thesis. On any particular day, this desire
may or may not prevail over my desire for immediate
gratification.

Given such common experiences, along with the necessity
of deliberation and countervailing beliefs and desires to our
notion of agency, the claim that mental events (e.g.
belief/desire pairs) and subsequent behavior are covered by
strict laws is highly dubious. In other words, belief/desire
pairs cannot be regarded as nomologically sufficient for
subsequent behavioral events. But if causal relations must
be backed by strict causal laws, then it appears that there
is no basis to the claim that particular mental events are
causally responsible for behavioral events. The project for
Fodor is to make sense of the causal efficacy of the mental,
relative to the anomaly of the mental and the standard
requirement that causal claims be backed by strict causal
laws.

Fodor's strategy is to claim that, while mental events
and subsequent behavioral events are not covered by strict
laws, they are covered by ceteris paribus laws (i.e. laws
who's antecedent events have attached ceteris paribus
clauses). According to Fodor's argument, ceteris paribus
laws do the same work as strict laws in covering causal
events: ”If it's a law that M —> B ceteris paribus, then it
follows that you get 85 whenever you get Ms and the ceteris
paribus conditions are satisfied.” (Fodor, 152, his emphasis)

In other words, where M and B are events covered by a ceteris

136

paribus law, as long as the ceteris paribus clause is
fulfilled, M is sufficient for B.

Fodor’s main point is that, if strict covering laws are
sufficient for establishing a causal relation, then ceteris
paribus laws will also be sufficient, since they both do the
same work. Strict laws establish a causal relation, since
the occurrence of the antecedent event is sufficient for the
occurrence of the consequent event. But, in cases where the
ceteris paribus clause of such a hedged law is discharged,
the antecedent becomes sufficient for the consequent event.
Thus, even if intentional laws are not strict and all we can
ever expect from the mental are ceteris paribus laws, such
laws establish causal relations just as well as strict laws.
Hedged laws only require that the ceteris paribus conditions
of that law be discharged for the antecedent to guarantee the
consequent.

There is much to recommend Fodor’s strategy. First, it
appears that ceteris paribus clauses are not unique to
psychology. All of the laws in the special sciences involve
ceteris paribus clauses. Thus, if ceteris paribus laws do
not back causal claims for psychology, then there is no
reason to believe they do so for any special science.

Second, experience shows that ceteris paribus laws in
psychology provide effective (although not flawless)
predictions of human behavior: ”We do use commonsense

psychological generalizations to predict one another's

137

behavior; and the predications do——very often——come out
true.” (Fodor II, 4)

Not only are these predications reliable, they are often
more accessible than those that the physical idiom is capable
of providing. For instance, in order to predict where I will
be next Saturday night, using only the language of physics,
one would have to be Laplace’s demon: There is far too much
one needs to know in order to predict my whereabouts, using
only the physical sciences. There is a far better, more
useful strategy in predicting where I will be on Saturday
night: As Fodor says, ”far the best way to find out
(usually, in practice, the only way to find out) is: ask
me!” (Fodor II, 6, his emphasis).

Such a strategy requires that one invoke a ceteris
paribus law: ”If he says (believes) that he will be at the
Green Door on Saturday night, then (ceteris paribus) he will
be at the Green Door on Saturday night.” Our best, and
usually only, tools for predicting complex behavior are the
ceteris paribus laws of the intentional idiom. According to
Fodor’s strategy, these laws back the causal claims made in
psychology. And what better vindication of these ceteris
paribus laws than their successful employment in practice.

Fodor’s strategy for justifying the causal efficacy of
the mental will not be defended here. Our present purpose is
merely to offer strategies by which the causal-intentional
relevance relation might be justified. Again, the causal-

intentional relevance relation implies that the mental is

138

causally relevant when seeking an explanation of certain
types of behavior (i.e. behavior subject to intentional
description). If Fodor's strategy is successful, then the
causal-intentional relation will be justified, in that
causal-intentional laws will justify (certain) causal claims
regarding mental states.

The most compelling doubts about the causal efficacy of
the mind are raised by the implications of physicalism: How
are we to understand the unique work done by the mental,
given that the physical sciences are capable of generating
exhaustive causal analyses of any behavioral event? How can
one claim that the mental is both irreducible and causally
efficacious and still avoid the problem of causal
overdetermination? This thesis, I believe, presents a
defensible solution to these problems. Further, Fodor offers
a solid strategy for justifying the causal-intentional
relevance relation. These arguments and strategies go a long
way toward allowing us to understand and justify the claim
that the mind is both irreducible and causally efficacious.

However, an important question remains: Might there not
be a simpler solution to the problem of mental causation?
Instead of making room for the causal efficacy of the mind,
might we not instead look to the arguments of the eliminative
materialist, who rejects the intentional idiom as a
legitimate means of characterizing human cognition. If
eliminativism is correct, then the problem of mental

causation dissolves: Without mental states, there is no

139

problem of mental causation. The difficulties which motivate
Kim”s explanatory exclusion principle are avoided.

Thus, does eliminative materialism provide a simpler,
more effective way to resolve the problem of mental causation
than the one provided here? This question is the subject for
the fifth and final chapter of this thesis. A response to
the eliminativist position will not only demonstrate the
value of the intentional idiom, but it will provide a final
demonstration of the value of the pragmatic theory of

explanation.

140

Chapter v: The Future of Intentionality

The foregoing considerations demonstrate that the
pragmatic conception of explanation provides the basis for a
plausible analysis of mental causation. Through the tools of
this theory of explanation, we are able to make sense of and
make room for the unique causal work performed by the mental
in the production of behavior, without violating the central
tenants of physicalism. But while this view is plausible,
might there not be a better way to resolve the difficulties
associated with mental causation?

It could be argued, for instance, that the contemporary
problems with mental causation, such as causal
overdetenmination, are indicative of a deeper, more
fundamental difficulty. Folk psychology (or common-sense
intentionality) has been the central theory for predicting
and explaining behavior for thousands of years. Talk of
beliefs, desires, wants and wishes has been the primary way
to understand ourselves as agents and to explain and predict
the behavior of both strangers and close relations.

Intentionality has been so successful as a theory of
human action that, as Fodor claims, it has become
”practically invisible.” (Fodor II, 3) Just as the expert
tennis player, due to years of successful practice and use,
'no longer notices the distinction between herself and her
tennis racket, we recognize no distinction between ourselves

and the theory of mental states by which we explain and

141

predict human behavior. But to what do we attribute the
success of the mental idiom? Do we credit the power and
scope of folk psychology for its success or has it succeeded
because it has long held a monopoly over characterizing,
explaining and predicting human behavior?

While intentionality has enjoyed a great deal of
success, it has also enjoyed a noticeable lack of
competition. Slowly, that is beginning to change. Only
recently have scientists begun to scratch the surface of the
mysteries of the human brain. While neuroscience is a
relatively new field of study, the prospects for the growth
of its predictive and explanatory power are tremendous. If a
fully articulated science of the brain were developed, would
we still require a theory of the mental?

To put the matter another way, are the epistemological
and metaphysical difficulties that motivate Kim’s explanatory
exclusion principle problems that require resolution, or are
they rather an indication that intentionality has outlived
its usefulness; that it must be eliminated as a legitimate
characterization of human cognition and as a source for
predicting and explaining behavior? Without a legitimate
alternative, intentionality can survive and flourish.
Problems such as causal overdetermination pose no real threat
to the mental if there is no alternative. Now, there appears
to be such an alternative, at least in principle.

Therefore, it could be argued that the pragmatic theory

of explanation does not provide the best solution to the

142

problems associated with mental causation. Instead, the
problems of mental causation can be dissolved by eliminating
the mental as a legitimate means of characterizing human
behavior and cognition. The contemporary difficulties of
mental causation indicate that the mental has simply outlived
its usefulness.

Thus, should philosophers of the mind embrace the
eliminativist position? Is the elimination of the mental as
a legitimate means of characterizing human behavior and
cognition a superior solution to the problems of mental
causation? In short, is intentionality scientific ”dead

wood”?

1. The Prospects for Eliminativism

Broadly speaking, the eliminativist believes that folk
psychology is fundamentally wrong or misleading. Anti—
reductionist attempts to retain an irreducible, useful theory
of the mental are wasted philosophical effort. But why
exactly is folk psychology wrong? Fodor, to take one
example, is quick to point out that folk psychology has
proven to be a useful theory for describing human cognition
and predicting behavior (with astonishing accuracy) for
thousands of years. What are the motivations behind
eliminative materialism that could possibly justify
eliminating folk psychology as a legitimate theory of human

cognition and behavior?

143

Specifically, there are two types of eliminativism, each
varying in degrees of strength. The first type we might call
{prescriptive eliminativism. Representative of this view
is Paul Feyerabend. Feyerabend's criticisms are directed at
identity theory--the idea that mental states can be
systematically reduced to physical states. Feyerabend claims
that identity theory is a result of understandable but
misguided presuppositions about how the development of a

theory of the brain should proceed:

Such [identity] hypotheses are usually put forth by
physiologically inclined thinkers who want also to be
empiricists. Being physiologically inclined, they want
to assert the material character of mental processes.
Being empiricists, they want their assertion to be a
testable statement about mental processes. (Feyerabend,
266)
According to Feyerabend, for the identity theorist, the
progress of a science of the brain will be gauged by how well
it accounts for our mental lives. In other words, the goal
of neuroscience is to generate a better, fuller understanding
of the human mind. A fully developed and articulated theory
of the brain will provide a deeper understanding of, for
example, the fear of heights, addiction to drugs, the desire
for chocolate, etc. Thus, while we satisfy our materialist
inclinations by maintaining the physical character of the
mental, we satisfy our ”empiricist” inclinations by claiming

that hypotheses about the brain should be ”testable

statements about mental processes.”

144

Both our ”physiological” and ”empirical” inclinations
are combined in an identity theory, where mental kinds are
type-identical to physical kinds. A successful theory of the
brain will allow us to develop the specific bridge laws
between physical and mental states. But, according to
Feyerabend, if we are to be materialists, such a theory is in
principle incorrect.

To invoke an identity between mental states and physical
states is to return to dualism: ”[Identity theory] not only
implies, as it is intended to imply, that mental events have
physical features; it also seems to imply ... that some
physical events ... have non—physical features. It thereby
replaces a dualism of events with a dualism of features.”
(ibid, 266) Even with a type-identification between mental
and physical states, there remains a dualism of properties or
features. Thus, to embrace the identity theory is to embrace
a sort of dualism. Therefore, since materialism is a
monistic theory, identity theory is in principle incompatible
with materialism, since it implies a dualism of properties or
features.

Feyerabend claims that, if we are to be consistent
materialists, a theory of the brain must be developed without
regard to the theory of the mental. Whether we conceive of
the mental as irreducible to the brain, or as type-reducible
to brain states, the mental is a priori incompatible with our

monistic metaphysics. Thus, to be consistent materialists,

145

we must give up talk of mental states altogether and develop
a purely physiological conception of the human being.

Prescriptive eliminativism is an in ”principle
argument”: We ought to give up talk of the mental, because
such talk is a priori inconsistent with our materialist
metaphysics. If Feyerabend’s position turns out to be
correct, then the problem of mental causation is dissolved.
There is no problem of mental causation because a causal
analysis of a behavioral event based in the mental idiom is
fundamentally incorrect. Strictly speaking, there is no
causal analysis involving mental states, since there are no
mental states. In this way, Feyerabend’s eliminativism:
provides an easy solution to the problem of mental causation.

A second form of eliminativism we might call predictive
eliminativism. Representative of this theory is Paul
Churchland. Like Feyerabend, Churchland advocates the
elimination of the mental as an adequate theory of human
cognition. However, Churchland’s arguments are not ”in
principle,” a priori arguments. Instead, Churchland bases
his eliminativism on the demonstrated inadequacies of folk-
psychology and the historic parallels with other outdated,
discarded theories.

The difference between Feyerabend’s brand of
eliminativism and that of Churchland can be demonstrated
through their respective reactions to identity theory.
Churchland, like Feyerabend, believes that a reduction of

mental states to brain states is in principle impossible.

146

But while Feyerabend denies identity theory because of its
dualistic implications, Churchland does so for different

reasons :

The one-to-one match-ups [between physical and mental
types] will not be found, and our common-sense
psychological framework will not enjoy an intertheoretic
reduction, because our common-sense psychological
framework is a false and radically.misleading conception
of the causes of human behavior and the nature of
cognitive activity. (Paul Churchland, 43, his emphasis)
Churchland’s brand of eliminativism proceeds from the
problems and limitations of folk psychology. Folk psychology
has survived, not because it is the best theory for
characterizing human cognition and behavior, but because it
has been the only theory. As Churchland claims, once a
science of the brain has been fully developed, ”we must
expect that the older framework will simply be eliminated,
rather than be reduced, by a matured neuroscience.” (ibid.
43)

But what exactly are these supposed limitations of folk-
psychology? First, Churchland cites what he calls the
”explanatory poverty” of the intentional idiom. There are a
number of capacities, requirements and characteristics of

human beings that folk-psychology has not been able to

explain successfully:

We do not know what sleep is, or why we have to have it,
despite spending a full third of our lives in that
condition. ... We do not understand how learning
transforms each of us from a gaping infant to a cunning
adult, or how differences in intelligence are grounded.

147

We have not the slightest idea how memory works, or how
we manage to retrieve relevant bits of information
instantly from the awesome mass we have stored. We do
not know what mental illness is, nor how to cure it.
(ibid. 46, his emphasis)

Fodor, on the other hand, repeatedly points out how well
folk psychology has worked as a theory for predicting how
others will behave (Fodor II, 3). Churchland admits that
under limited conditions, folk psychology appears to give
good behavioral predictions. But as soon as we venture
beyond this limited scope (into the arena of patients who
suffer from brain damage, for instance), folk psychology
loses its explanatory and predictive power. It's predictive
and explanatory scope are simply too narrow to be of any real
value (Churchland, 46).

In addition to citing the supposed explanatory,
predictive and descriptive poverty of folk psychology,
Churchland draws parallels to the history of science, where
the creation of a successful scientific theory led to the
demise of a previous theory. For instance, Churchland cites
the example of caloric, the theory of heat that held sway for
much of 18th and 19th century physics. Heat was thought to
be the substance caloric, a fluid contained in all objects.
The transfer of heat from one body to another was thought to
be a transfer of substance-~a transfer of a certain quantity
of caloric.

This theory enjoyed a degree of predictive and
explanatory success, but was ultimately replaced by the

theory of kinetic energy. Instead of heat being a substance

148

contained in an object, it had become accepted that heat is
the motion of molecules of the object itself. This latter

theory enjoyed greater explanatory and predictive success.

Eventually, the theory of caloric was simply dropped by the
scientific community.

As Churchland claims, it was not that the theory of
caloric was incomplete. It was simply mistaken. Hence, it
was eliminated as a theory of heat, even though it had
enjoyed a limited degree of success. Similarly, Churchland
predicts, once a more complete theory of the brain is
developed, mental states, like caloric, will drop from our
ontology.

Clearly, Feyerabend’s version of eliminativism.makes
bolder claims than Churchland's. While the latter merely
predicts the elimination of folk psychology, based on its
demonstrated lack of explanatory and predictive scope and
power, the former version eliminates the mental in principle,
regardless of its value as a theory of human cognition and
behavior. Let us begin by responding to Feyerabend's
arguments.

NOt only is Feyerabend’s prescriptive eliminativism
stronger, but it appears too strong to be plausible.
Arguments similar to those used in chapter II to call into
question the plausibility of the explanatory exclusion
principle can be applied to Feyerabend's arguments: If
Feyerabend’s a priori arguments for the elimination of the

mental are successful, then there seems to be no reason why

149

they won’t be successful for other special sciencesl.

Sciences such as aerodynamics (with concepts such as the
airfoil) invoke non-physical properties, i.e. properties not
contained in the nomenclature of basic physics. So if we
apply Feyerabend’s arguments (and there appears to be no good
reason why we should not), all such sciences and their
ontologies must be eliminated. Surely this is absurd.

Thus, prescriptive eliminativism suffers from the same
drawback as Kim’s explanatory exclusion principle: As long
as we apply them only to folk psychology, such theories
appear plausible. But the scope of both Kim's and
Feyerabend’s arguments is too broad. Nothing precludes
their application to any of the special sciences-—sciences
which repeatedly demonstrate their predictive and explanatory
value. Once such an application is made, the untenability of
their arguments is revealed.

The fact that prescriptive eliminativism becomes
untenable when applied to less controversial special sciences
provides a diagnosis of the problem with Feyerabend's
argument: Arguments that, a priori, eliminate theories or
explanations based solely on metaphysical concerns put the
proverbial cart before the horse. One of the primary guides
for theory retention or elimination is the explanatory and

predictive power of a theory. Tyler Burge makes this type of

 

1 See Chapter II, pp. 34—36.

150

argument regarding the value of the mental idiom in providing

causal explanations:

The probity of mentalistic causal explanation is deeper

than the metaphysical considerations that call it into

question. ... Materialist metaphysics is not the most

plausible starting-point for reasoning about mind-body

causation. Explanatory practice is. (Burge, 117-118)
Feyerabend’s arguments for eliminativism fail because they
presuppose that the value of a scientific theory is
determined primarily by how well it fits into a monist
metaphysics. If a theory does not fit neatly into such a
metaphysics, it must be eliminated, regardless of its
practical value.

When we apply his arguments to other special sciences,
those arguments lose their luster because it is evident that
such theories are valuable, not because they fit nicely into
a materialist metaphysics (for if Feyerabend is correct about
the incompatibility of property dualism and substance monism,
then surely they do not), but because they have considerable
predictive and explanatory power. Further, they can provide
explanations that the physical sciences cannot. Thus, we can
broaden Burge’s claim: The probity of the explanations (and
explanatory value) of any special science is deeper than the
metaphysical considerations that call them into question.

That is not to say that metaphysical considerations are
completely removed from theory evaluation. One ought not

think that the door is open to whatever occult property one

wishes. We can consider our metaphysical commitments as one

151

set of criteria for theory evaluation, but we cannot see them
as the primary or sole criteria. Feyerabend's arguments make
such an assumption. Subsequently, his arguments are rendered
untenable.

On the other hand, Churchland's predictive eliminativism
is a much more plausible argument. On his argument, the
mental idiom will be eliminated because of its supposed
explanatory and predictive poverty. Because it has failed to
characterize much of human cognition (such as learning,
memory, imagination, etc.) it will be eliminated in favor of
a developed theory of the brain which, it is expected, will
not suffer the same explanatory and predictive limitations.

Churchland’s arguments avoid the problems encountered by
Feyerabend’s, since there are no implications for eliminating
other special sciences. On predictive eliminativism,
theories can be evaluated on an individual basis according to
their descriptive, explanatory and predictive value, relative
to the prospects of a (potential) replacing theory or
science. Thus, predictive eliminativism.makes a weaker but
more tenable argument than prescriptive eliminativism.

But while predictive eliminativism cannot be ruled out
in principle, it is highly doubtful that Churchland’s
predictions will come to fruition. First, the bulk of this
thesis has been designed to show that the mental idiom is
capable of generating causal explanations of behavior that
the physical idiom cannot. Further, not only are these

causal explanations unique to the mental, they are vital to

152

our conception of ourselves as rational agents. Tyler Burge
defends the uniqueness and importance of the intentional
idiom via its ability to generate unique causal explanations

of behavior:

Much of the interest of psychological explanation, both
in,psychology and in ordinary discourse, lies in helping
us understand ourselves as agents. ... We think that
we make things happen because we make decisions or will
to do things. We think that we make assertions, form
theories, and create cultures, because we think certain
thoughts and have certain goals—-and we express and
fulfill them. In this context, we identify ourselves
primarily in terms of our intentional mental aspects

Our agency consists in our wants’, willings’, thoughts’,

values as such ... having some sort of efficacy in the

world. Our mental events having the intentional
characters that they have is, in individual instances,

what we define our agency in terms of. (Burge, 118-119)
The intentional idiom provides a unique way of viewing human
behavior and the payoff of this mode of inquiry is in the
unique explanations and characterizations of human behavior
that it provides. It is primarily through the intentional
idiom.that we understand ourselves as rational agents.

Such an understanding is not captured in the physical
idiom. Rationality is unique to the mental and is not to be
found in the physical idiom. For instance, arguments have
been put forth (primarily by Donald Davidson) to suggest that
mental states cannot be reduced to physical states simply
because the rationality of beliefs and actions is not

captured by reference to physical states. Rationality is

unique to the intentional idiom and cannot be assimilated by

153

the physical. Cynthia and Graham Macdonald offer examples of

this ”rationality of beliefs”:

Mental states, such as belief states, have normative
connections with each other: if one believes that
2+2=4, and that 1+1+1+1=4, then one ought to believe
that 2+2=1+1+1+1. If one wants to be a good chess
player and believes that studying chess openings will
improve one’s game, then one ought to study chess
openings. This would be rational, given this belief and
desire. ... The normativity involved in these
connections is essential to the mental and is not to be
found in the physical domain. No purely physical
property has rational connections to other physical
properties, so a reduction is in principle impossible.
(Macdonald, 9)
It is the intentional aspect of the mental that allows for
the characterization and explanation of human behavior as
purposive and rational. While such a characterization is
unavailable to the physical idiom, it is vital to our common
conception of ourselves as agentsz.

Thus, the comparison Churchland draws to historical
examples is not as persuasive as one might imagine. For
instance, the theory of caloric makes a poor comparison to
the mental idiom because it is difficult to find the
explanatory or predictive power that is unique to that
theory. In other words, it is not clear that caloric, as a

theory of heat, does any explanatory or predictive work that

is not accomplished by the theory of kinetic energy. If this

 

2 It can be argued by extension that intentionality is
necessary to a successful analysis of free will. See, for
instance, C.A. Campbell’s notion of ”phenomenological
analysis” in ”Has the Self a ’Free Will’?” in Qn_§elfihggd_and
figdhggd (London: George Allen & Unwin) 1957, pp. 158-179.

154

thesis is correct and the intentional idiom is capable of
generating unique and important causal explanations of
behavior, then comparisons to examples such as caloric simply
do not hold.

Therefore, while Churchland’s arguments are not in
principle untenable, it is doubtful that his predictions
regarding the eventual elimination of the mental idiom will
come to fruition. If, indeed, a mature neuroscience were to
capture both the unique character of human cognition and the
explanations of human behavior found in intentionality, then
Churchland’s predictions might very well be realized.
However, for such a study of the brain, this would require a
radical break from the physical sciences, given that
rationality and agency are not concepts currently contained
in such sciences.

Further, Churchland is correct in claiming that the
explanatory and predictive scope of a mature neuroscience
will be much broader than that of folk psychology. This fact
is acknowledged in chapter IV: The e-contexts of the mental
ididm can suffer from descriptive failure, while in
principle, the physical idiom cannot. The physical idiom, it
is granted, is capable of generating a true explanation of
any behavioral event, while the intentional idiom is not. On
the other hand, while the descriptive, explanatory and
predictive capabilities of the mental idiom are narrow, they
are nonetheless valuable. They are vital to our conception

of ourselves as rational agents and they cannot be captured

155

in the physical idiom. Thus, regardless of the narrow
predictive and explanatory scope of folk psychology, it is
doubtful that it will simply be eliminated in favor of a
fully articulated neuroscience.

Given our analysis of and response to both prescriptive
and predictive eliminativism, we are now in a position to
answer the central question of this chapter; namely, can
eliminativism.offer a superior resolution to the problems of
mental causation? While eliminativism is capable of
generating a plausible solution to the problem of mental
causation, that solution is by no means superior. The cost
of embracing that solution is our common conception of
ourselves as rational agents and that cost, most would agree,
is much too high.

On the other hand, the solutions offered in this thesis
resolve the difficulties of mental causation. Further, they
retain the important, albeit limited, explanatory,
descriptive and predictive work performed by the intentional
idiom- The fact that, with the pragmatic conception of
explanation, we are able to resolve the difficulties of
mental causation and still retain the valuable work performed
by the intentional idiom makes the solution offered in this

thesis superior to the solution offered by the eliminativist.

156

2. Concluding Remarks: The Value of the Pragmatic
Theory of Explanation

If anti—reductionist arguments are correct, then mental
states will resist a systematic reduction to physical states.
But while a reduction of the mind to the body may be
avoidable, an assimilation of the mind into our scientific
world-view is a necessity. A modern conception of the mind
must find its place in a physical world.

While there are a number of difficulties and obstacles
to such an assimilation, none are more obstinate and
important than the problems of mental causation: Given that
any behavioral event can be given an exhaustive causal
analysis in the physical idiom, what work is left for the
mental to do? Further, if there is a sense in which the
mental does unique causal work in the production of behavior,
how are we to avoid the problem of causal overdetermination?

The pragmatic conception of explanation provides the
foundation for an analysis of mental causation that answers
both of these questions. Clearly, however, this theory of
explanation is still in its infancy, despite the work of
theorists such as van Fraassen and Garfinkel. It will need
to be refined in order to deal with the inevitable objections
and counterexamples faced by any theory of explanation. But
while it requires further development and refinement, there
is much to recommend this pragmatic theory of explanation.

First, the pragmatic theory of explanation is valuable

because the analysis of mental causation that it provides is

157

valuable. Such a theory of explanation provides an
understanding of mental causation that makes sense of and
makes room for the unique causal powers of the mental. The
causal efficacy of the mental, in turn, gives value and

meaning to human agency. As Burge argues,

If intentional psychological explanation ’made sense’ of
what we did ... but did not provide insight into the
nature of any causal efficacy, it would lose much of its
point. It would provide no insight into the various
forms of agency that give life its meaning and purpose,
and psychology its special interest. (Burge, 119)

If we are unable to answer the problems of mental causation,
then we must either embrace epiphenomenalism, or eliminate
the mental altogether. As has been argued in this thesis,
both of these latter options come at too high a price.
Hence, regardless of the difficulties that the pragmatic
theory might face, it demonstrates its value and usefulness
in the understanding of mental causation that it generates.
Further, the pragmatic theory of explanation can avoid
the explanatory exclusion principle, but it is not clear that
other theories of explanation are so able. Recall that the
explanatory exclusion principle requires that there be only
one complete and independent explanation for any event or
phenomenon. As shown in chapter I, Kim offers metaphysical
and epistemological arguments that motivate the explanatory
exclusion principle, effectively precluding the possibility

of having multiple complete and independent explanations.

158

However, it has been shown that the explanatory
exclusion too strong a condition on explanation in general.
The implications of the explanatory exclusion principle are
such that the irreducible properties of any special science
must be regarded as causally inert. Thus, instead of being a
solution to the problem of mental causation, it is a problem
for explanation theory: In order to avoid the implications
of the explanatory exclusion principle, a theory of
explanation must be able to resolve the metaphysical and
epistemological problems of having (apparently) multiple
complete and explanations. In other words, a satisfactory
theory of explanation must be able to cope with the
difficulties of compounding epistemic puzzlement and causal
overdetermination.

As shown in chapter III, the pragmatic theory is capable
of diffusing these difficulties--it is able to maintain a
multiplicity of explanations without compounding puzzlement
and without implying causal overdetermination.’ On the other
hand, it is not clear that other standard theories of
explanation, such as Hempel’s D-N model, are able to deal
with these epistemological and metaphysical difficulties.
Therefore, regardless of the difficulties or objections that
it may encounter, the pragmatic theory of explanation shows
its value in that it is able to avoid the difficulties which
entail explanatory exclusion.

Churchland may be right. The intentional idiom may

eventually find itself on the scientific scrap heap. But

159

there are good reasons for believing that Churchland is
wrong--that the mind and its unique causal powers are not
only assimilable into a scientific world-view, but also that
they are vital to our conception of ourselves as agents. A
successful analysis of mental causation will maintain the
causal efficacy of the human agent in the world and help
assimilate the mind and its unique causal properties into our
scientific world-view. I believe this thesis has provided

such an analysis.

160

BIBLIOGRAPHY

161

BIBLIOGRAPHY

Achinstein, Peter .Ihe_Natu;e_pf_ExplanaLiQn. Oxford: Oxford

University Press, 1983.

Barnes, Jonathan. Ba;ly_§;eek_£hilgspphy. London: Penguin
Books, 1987.

Batens, Diderik. ”Do We Need a Heirarchical Model of
Science?” Inferencel_Exnlanationl_and_cher
ErustratiQnsl_Essaxs_in_the_2hilcscnh¥_9f_Science. Ed.

John Earman. Berkeley: University of California Press,
1992. 199-215.

Bromberger. Sylvian» CKLlflruLJELJQEHLIELJXXLHLJmea.

.10 «to. co rag to: v.00. v01 0 - at 029‘

Them. Chicago: The University of Chicago Press, 1992.

Burge, Tyler. ”Mind-Body Causation and Explanatory Practice.”
Mental_§ausation. Ed. John Heil and Alfred Mele. Oxford:
Clarendon Press, 1995. 97—120.

Burtt. E.A.. Ihe_Metanhzsical_Eoundations_of;MQdern_acience.
Atlantic Highlands, NJ: Humanities Press, 1952.

Campbell. C.A.. Qn.$elfh99d_and_fiodhood. London: George
Allen & Unwin, 1957.

Carnap, Rudolf. ”Empiricism, Semantics and Ontology.” The

Philosonhx_of_Science. Ed. Richard Boyd. et
al. Cambridge, Mass: The MIT Press, 1991. 85-98.

Churchland, Patricia S. Nguxgphilgspphy. Cambridge, Mass.:
The MIT Press, 1993.

Churchland, Paul. Ma;;e;_and_§gn§gipu§ness. Cambridge, Mass:
The MIT Press, 1993.

162

——-. ”Eliminative Materialism and Propositional Attitudes.”

Journal_of_2hilosonhx. 78. 1981. p. 73-

Davidson, Donald. ”Thinking Causes.” Men;al_§au§atipn. Ed.
John Heil and Alfred Mele. Oxford: Clarendon Press,
1995. 3-18.

---. ”Mental Events.” Essaxs_on_Actions_and_Exents.
Oxford: Clarendon Press, 1980. 207-224.

-—-. Inauiries_into_Iruth.and_lnternretation. Oxford:

Clarendon Press, 1984.

Dretske, Fred. ”Reply: Causal Relevance and Explanatory
Exclusion.” Philosonhx_of_£s¥chologx;_nebates_on

Esxchological_Exnlanation. Ed. Cynthia MacDonald and
Graham MacDonald. Oxford: Blackwell Publishing, 1995.

142-155.

Feyerabend, Paul. ”Mental Events and Brain Events.” The
Natuze_of_mind. Ed. David M. Rosenthal. Oxford: Oxford
University Press, 1991. 266-267.

Feynman. Richard P. QEDi_Ihe_Strange_Iheor¥_of_Light_and_
Manger. Princeton, NJ: Princeton Science Press, 1985.

Fodor, Jerry. ”Making Mind Matter More.” A_Ihepzy_pfi_§9nggng

and_QLhe;_Essays. Cambridge, Mass.: The MIT Press,
1992. 137-160.

---. Egyghgggmanpips. Cambridge, Mass.: The MIT Press, 1993.

--—. ”You Can Fool Some of The People All of The Time,
Everything Else Being Equal; Hedged laws and
Psychological Explanation” Mind 100. (1991): 19—33.

Friedman, Micheal. ”Explanation and Scientific

Understanding.” Theories_of_Exnlanation. Ed. Joseph
Pitt. New York: Oxford University Press, 1988. 188-198.

163

Gamow. George- Ihe_Great_2hysi2ists_Erom_Galileo_to_Einstein.
New York: Dover Publications, 1961.

Garfinkel Alan Eorms_of_Exnlanation1_8ethink1ng_1he

Questions.of.fiocial_2heory New Haven: Yale University
Press, 1981.

Gasper, Philip. ”Causation and Explanation: Introductory

Essay.” Ihe_£hilosonhy_of_Science. Ed Richard Boyd et
al. Cambridge, Mass: The MIT Press, 1991. 289— 298.

Hempel Carl G .Asnects_of_Scientific_Exnlanation_and_cher

Essays_in_the_£hilosonhy_of_$cience New York: The Free
Press, 1965.

Hintikka, Jaakko. ”Semantics and Pragmatics for Why—

Questions ” Ihe_Qournal_of_£hilosonhy 1995 pp 636-
657.

Holcomb, Harmon. ”Logicism and Achinstein’s Pragmatic Theory
of Scientific Explanation.” Dialeggiga 41, no. 3. (1987)
239-248.

Kim, Jaegwon. Superyenienge_and_Mindi__S§l§£L§d
Bhilpspphigal_fissays. Cambridge: Cambridge University
Press, 1993.

_——.I2hilgfignh¥_Q£JMind. Bolder: Westview Press,
1996.

——-. ”Explanatory Exclusion and the Problem of Mental
Causation ” Ehilosonhy_of_£sychologyi_nebates_on_

Esychological_fixnlanation Ed Cynthia MacDonald and
Graham.MacDonald. Oxford: Blackwell Publishing, 1995.
121—141.

-—-. ”Can Supervenience and ’Non-Strict Laws’ Save Anomalous

Monism?” Mental_gau§ation. Ed. John Heil and
Alfred Mele. Oxford: Clarendon Press, 1995. 19—26.

ImmLimmms.flmJflnummebfidemifinJmmnmimm.
Chicago: The University of Chicago Press, 1962.

164

Laudan. Larry W
W Berekely:

University of California Press, 1984.

MacDonald, Cynthia and Graham.MacDonald. ”Introduction:
Supervenient Causation”. WWW
Won Ed. Cynthia

MacDonald and Graham.MacDonald. Oxford: Blackwell
Publishing, 1995. 4—28.

Mackie, John L. ”Causes and Conditions.” Causapipn. Ed.
Ernest Sosa and Michael Tooley. Oxford: Oxford
University Press, 1993. 33-55.

Menzies, Peter. ”Against Causal Reductionism” Mind 97.
(1988): 551-574.

Nagel Ernest W
Wage New York: Columbia

University Press, 1979.

Priest, Stephen. Ihepries_pf_theguind. Boston: Houghton
Miflin Company, 1991.

Putnam, Hilary. ”The Nature of Mental States.” The_naguxe_gf
Mind. Ed. David M. Rosenthal. Oxford: Oxford
University Press, 1991. 197-203.

Quine, W;V.O.. Bursuit_pj_$ruth. Cambridge: Harvard

University Press, 1992.

Rorty. Richard. W. Minneapolisz

University of Minnesota Press, 1982.

Rosenthal. David M. ed. W
.Erpplemp Indianapolis: Hacket Publishing, 1987.

165

Salmon, Wesley C. ”Four Decades of Scientific Explanation.”
Sgientifiie_ExpleneLien. ed. Philip Kitcher and Wesley

C. Salmon. Minnesota Studies in the Philosophy of

Science, Vol. XIII. Minneapolis: University of
Minnesota Press, 1989.

Scriven, Micheal. ”Causation as Explanation.” Rene 9. (1975):
3-16.

Sosa, Ernest. ”Mind- -Body Interaction and Supervenient

Causation” M1dwest.£tud1es_1n_£hilosonhy IX, (1984):
271-281.

Stillings, Neil A., et al. Cegnigiye nggngg° AD
Intredueeien. Cambridge, Mass.: The MIT Press, 1995.

Van Fraassen, Bas C. ”The Pragmatic Theory of Explanation.”
Iheories.of_zxnlanation Ed Joseph Pitt New York:

Oxford University Press, 1988.136-155.

—--. The scientific Imege. Oxford: Clarendon Press, 1980.

Wittgenstein, Ludwig. Qn_£e;;ein;y. Ed. G.E.M. Anscombe and

G.H. von Wright. Trans. Denis Paul and G.E.M. Anscombe.
New York: Harper and Row, 1969.

Worley, Sara. ”Mental Causation and Explanatory Exclusion.”
Exkennenie 39. (1993): 333-358.

166

MICHIGAN STRTE UNIV. LIBRARIES
1|11|H111"I1111111111N1111111111111111111
31293017718150