UNCORRECTED PROOF!
KARSTEN R. STUEBER
MENTAL CAUSATION AND THE PARADOXES OF
EXPLANATION
ABSTRACT. In this paper I will discuss Kim’s powerful explanatory exclusion argument against the causal efficacy of mental properties. Baker and Burge
misconstrue Kim’s challenge if they understand it as being based on a purely
metaphysical understanding of causation that has no grounding in an epistemological analysis of our successful scientific practices. As I will show, the emphasis
on explanatory practices can only be effective in answering Kim if it is understood as being part of the dual-explanandum strategy. Furthermore, a fundamental problem of the contemporary debate about mental causation consists in
the fact that all sides take very different examples to be paradigmatic for the
relation between psychological and neurobiological explanations. Even if we
should expect some alignment in the explanatory scope of neurobiology and
psychology/folk-psychology, there is no reason to expect that all mental explanations are exempted by physical explanations, since they do not in general explain
the same phenomena.
INTRODUCTION
Due to the influential arguments of Donald Davidson, philosophers
generally accept the claim that one has to conceive of reasons as
causes and that one has to regard psychological explanations as
causal explanations. Despite this agreement there is, however, still
no consensus about how such a causal conception of mentality
is justified within a contemporary physicalistic framework. Even
though hardly anybody accepts Cartesian substance dualism, we
still have difficulty in accounting for the causal efficacy of mental
properties. A merely cursory look at the recent debate about mental
causation could leave the impression that we seem to be condemned
to conceive of mental states as causes for actions without being
able to philosophically comprehend such a conception. Our best
philosophical arguments provide prima facie strong reasons for the
claim that mental properties can have only epiphenomenal status.
Philosophical Studies 00: 1–35, 2003.
© 2003 Kluwer Academic Publishers. Printed in the Netherlands.
WEB2C
PDF-OP
Disk, CP
VICTORY: PIPS No.: 5149227 (philkap:humsfam) v.1.2
phil8321.tex; 8/09/2003; 18:50; p.1
2
KARSTEN R. STUEBER
But this is a conclusion we cannot live with since it contradicts our
self-conception as responsible agents.1
The above difficulty stems from the fact that mental properties, whether they are phenomenal or intentional properties, are for
various reasons not reducible to specific physical or neurobiological
properties of a specific organism. Reasons most often cited for nonreducibility are the multiple realizability of mental properties, the
non-individualistic constitution of mental content and the so-called
anomaly of the mental. In this article, I will take the non-reducibility
of mental properties and the non-individualistic character of mental
content for granted. I will rather focus on what I regard to be
the central challenge any non-reductionist physicalist has to meet:
Kim’s argument against the causal efficacy and causal relevance of
mental properties based on the principle of explanatory exclusion.
In the first section, I will outline the precise structure of this
powerful argument and its presupposition. The most promising
response to Kim’s challenge is, in my opinion, to be found in Baker
and Burge’s proposal to think about the causal efficacy of specific
properties in the context of established scientific and commonsensical explanatory practices. However, as I will argue in the
second section, their attack on Kim’s argument is to a certain extent
misguided. Kim’s challenge is misconstrued if one understands it
as being based on a purely pre-Kantian metaphysical conception
of causation that is “done in total isolation from epistemology.”2
I will argue that the emphasis on explanatory practices can only be
effective in countering Kim’s challenge if it is understood as being
part of the so called dual-explanandum strategy, i.e. the claim that
different phenomena are explained within the explanatory schemes
of psychology and physical sciences, including neurobiology.3
Similarly, various attempts to argue for the compatibility of
psychological and “physical” explanation of behavior by pointing
to the autonomy of psychological generalizations or counterfactuals
should be seen in this manner. Kim has some sympathy for the
dual explanandum strategy but he is worried about whether the
causal-explanatory role of mental properties can be regarded as
truly autonomous or whether it is not merely “piggy-backing” on
the underlying neurobiological mechanisms.4 Nevertheless, merely
raising this question is not sufficient to put the metaphysical
phil8321.tex; 8/09/2003; 18:50; p.2
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
3
integrity of established explanatory practices into doubt. I will
suggest that the scope of the causal inheritance principle, which
grounds Kim’s worries about “piggybacking,” is limited to certain
explanatory contexts. As I will show, it does not apply to all psychological explanations, since the contextual and social factors that
are constitutive for their explananda involve essentially a mental
dimension. Furthermore, social and intentional categories have a
wide historical supervenience base that does not map smoothly
onto the causal mechanisms that do the explanatory work for the
lower level sciences. There is thus no reason to expect that the
causal-explanatory character of mental properties can be generally
accounted for in terms of the underlying physical mechanisms, even
if we should expect some realignment in the explanatory scopes of
neurobiology and psychology or folk-psychology.
1. THE EXPLANATORY EXCLUSION ARGUMENT
Kim’s challenge to mental causation within the framework of nonreductive physicalism can be reconstructed in the following manner.
He argues that the following six assumptions form an inconsistent
set. For the purpose of this discussion, I will articulate his argument
in terms of explanation and not only in terms of causation.5
1. The thesis of explanatory realism:
a. Explanations are grounded in and true because of objective
and mind-independent relations between events in the real
world. A causal explanation of an event E by reference to
event C is only true if there exists a real causal relation
between C and E that is independent of our explanatory
practices.6
b. Causal relations between events C and E hold in virtue of
certain properties. A causal explanation of event E in terms
of C is true only if it cites the properties in virtue of which C
causes E.
2. Psychological explanations are causal explanations. Mental
properties fulfill a real causal explanatory role.
3. We provide sometimes causal psychological explanations for
bodily movements. For example, we explain why somebody
phil8321.tex; 8/09/2003; 18:50; p.3
4
KARSTEN R. STUEBER
raises his hand by referring to his desire to be noticed or by
mentioning the pain he is in. Similarly we explain why somebody climbs a ladder by saying that he wanted to retrieve his
hat.
4. By 1–3 one is committed to the existence of psychophysical
causation, i.e. to the claim that “an event, in virtue of its
mental property, causes another event to have a certain physical
property.”7
5. The principle of physical closure: “Any physical event that has
a cause at time t has a physical cause at t . . ., i.e. if we trace the
causal ancestry of a physical event, we need never go outside
the physical domain.”8 It follows that each event and its physical properties have a complete explanation within the physical
domain. Our physical theory of the physical domain is therefore
complete and self-sufficient.
6. Causal overdetermination cannot be accepted as providing us
with a general picture of the relationship between mental and
physical causation. It would imply that if the “physical cause
hadn’t occurred, the mental cause by itself would have caused
the effect.”9 This however would violate physical closure,
besides endowing the mental with mystical powers.
(5.) and (6.) however contradict the assumptions (1.)–(4.) If a physical explanation provides a complete causal explanation of a specific
physical event and if overdetermination is not an option, then we
cannot accept another mental explanation as being objectively true.
Within the context of explanatory realism a complete explanation
places an event uniquely within the causal nexus of the world. For
that reason there cannot be two complete and independent explanations of the same event. Accordingly, one cannot accept a psychological story about how the specific physical event occurred as an
objective explanation of this event. The possibility of a sufficient
and complete physical explanation preempts the causal objectivity
of mental explanations. Non-reductive physicalists can’t have their
cake and eat it too.
Kim also suggests that one can accept the explanatory exclusion
argument for merely epistemic reasons, i.e. even if one rejects the
thesis of explanatory realism. According to this line of reasoning,
explanations are supposed to provide a certain unity of knowledge
phil8321.tex; 8/09/2003; 18:50; p.4
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
5
by connecting various items in our experience. Given this function
of unification, one cannot accept two complete explanations for the
same phenomenon without creating a certain tension in one’s world
view. This tension can only be resolved through a plausible account
of how these explanations are related to each other.10
However, a merely epistemic understanding of the concept of
explanation is not necessarily committed to conceiving of explanations as unifying all domains of knowledge. The explanatory exclusion argument could not get off the ground, if one understands
explanation as an intensional, context-sensitive, and interest-relative
notion. One could even combine such an understanding of explanation with the acceptance of (1a.) in the context of conceiving of
causal relations as extensional relations between concrete events.
What is crucial for generating the exclusion problem is acceptance
of (1b.), i.e. the claim that C causes E in virtue of certain properties.
A defender of Davidson’s anomalous monism would however deny
the truth of exactly this premise.11 Such a philosopher (i) asserts
a token-identity between mental events and physical events, (ii)
claims that only physical laws can be strict and (iii) maintains a
nomological conception of causation, i.e. events are causally related
only if they can be subsumed under some strict laws. Furthermore,
he espouses a nominalistic conception of properties and is at pains
to stress that causal relations do not hold in virtue of any properties, physical or mental. Properties do not play any genuine causal
role but are only explanatorily relevant depending on the specific
explanatory context. As Davidson himself emphasizes, the acceptability of an explanation depends on the conceptual resources that
we use to formulate them. Here lies the crucial difference between
explanations and singular causal statements. It is, for example,
perfectly in order to say that the event described on the front page
of the NY Times caused the event described on the front page of the
Frankfurter Allgemeine, but we could not provide an explanation for
the occurrence of these events under such a description. Explanations are successful not only because they succeed in identifying
causal relations between certain events but because they identify
such events in terms of concepts that play a role in the formulation of strict or non-strict lawlike generalizations. Both physical
and psychological descriptions are therefore explanatorily relevant
phil8321.tex; 8/09/2003; 18:50; p.5
6
KARSTEN R. STUEBER
since both of them allow us to formulate such generalizations,
even if psychological properties do not function in strict laws.
Psychological explanations also have an autonomous status vis-à-vis
physical explanations. They allow us to see regularities that cannot
be recognized otherwise since the intentional idiom is not reducible
to the physical idiom.
From this perspective, the existence of two explanations for a
single phenomenon does not create any epistemic discomfort. Since
explanations are relativized to contexts and there is not just one such
context, one should rather expect a multiplicity of explanations. An
explanatory contextualist does not have to deny that explanations
create some epistemic unity within one context, but he would insist
that they do not create unity across different explanatory contexts.
It should, therefore, not be surprising that we might have different
explanations of the same phenomenon in different contexts.
Even though such a position is internally consistent, its conception of the relation between causation and causal explanation is
highly problematic. The assumption that causal relations hold in
virtue of certain properties is a metaphysically plausible one, since
it is supported by our scientific practices.12 Strictly speaking within
the Davidsonian framework we could not ask for an explanation of
why event C caused event E as we ordinarily do, since events do
not cause each other in virtue of anything. We could only ask for an
explanation of why event C occurred in the sense of asking which
of the many events that happened prior to C caused C to occur.
Yet this is certainly a distortion of our scientific practices. Medical
researchers, for example, are not only interested in explaining why
peptic ulcers occur by merely pointing to the presence of Helicobacter pylori. They are also interested in finding out how and why
H pylori bacteria are able to cause ulcers in the stomach. Indeed, as
Paul Thagard has shown, understanding the underlying mechanism
of this causal relation has been a salient factor for accepting the
bacterial theory of peptic ulcers in the medical community.13 Our
metaphysical intuitions, as sanctioned by our explanatory practices thus support a position of explanatory realism as specified in
premises (1a) and (1b). Accordingly, we cannot point to the context
relativity of explanations in order to account for a multiplicity of
explanations of the same phenomenon. We would need an account
phil8321.tex; 8/09/2003; 18:50; p.6
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
7
of how these various objective properties are related to each other.
Yet it is important to recognize that the epistemic tension between
two explanations is created only in light of certain metaphysical assumptions about the causal order of the world. Since these
assumptions are plausible, we need to address Kim’s explanatory
exclusion argument, in order to account for mental causation and
the autonomy of psychological explanations.
Kim is nowadays rather skeptical about his earlier attempts to
account for the explanatory objectivity of psychological explanations in terms of supervenient causation, since within this model
each supervening mental property is causally efficacious only
because of its supervenience base and psychological explanations
still seem to be objectively preempted by physical explanations.14
Of course, within the context of traditional identity theory the above
problems would not have arisen. In that scenario, a mental property
could be identified with a physical property and no violation of the
physical closure principle would occur.15 In light of the multiple
realization arguments articulated by functionalists, the traditional
type-type identity theory cannot be defended. Nevertheless, Kim
suggests that we can only overcome the problem of mental causation and the causal-explanatory exclusion problem if it is possible to
argue for a more limited form of local reduction, i.e. an identification
of a mental property with a physical property relative to a particular
species or an even more localized and fine-grained reduction.16
Crucial for such a local reduction is the functionalization of
mental properties, i.e. the definition of a mental property in terms
of its “causal/nomic relations to other properties.”17 Such a functionalization would allow us to identify a mental property with
a physical property, which has the same functional role, without
making such talk of identification merely ad hoc. Kim is quite
aware of the fact that such functionalization does not seem to be
plausible for phenomenal mental properties. Yet he is persuaded that
such functionalization should be in principle possible for intentional
properties because intentionality supervenes logically on the physical structure of the world. It is thus “inconceivable that a possible
world exists that is an exact physical duplicate of this world but
lacking wholly in intentionality.”18
phil8321.tex; 8/09/2003; 18:50; p.7
8
KARSTEN R. STUEBER
Unlike Kim, I am however doubtful that accepting such global
supervenience forces me automatically to believe in the functionalization of intentional properties in any formal and precise manner
and in a way that will honor our folk-psychological conception of
mental content. It seems to me perfectly reasonable to insist that
our practice of ascribing intentional properties contains an irreducible pragmatic and non-formalizable element,19 while recognizing
that our common sense conception of intentionality also honors the
above supervenience intuition. We need therefore a more elaborate proposal in order to justify the hope that a functionalization
of intentional properties can be carried out. A natural candidate
for the functionalization of content properties is conceptual role
semantics. Yet as Fodor and LePore have pointed out, if conceptual
role semantics identifies the content of a mental representation with
its inferential role within the cognitive system of an individual, then,
assuming the impossibility of an analytic/synthetic distinction, each
difference in inferential role has to count as a semantic difference.
Two individuals can only share one belief if they share all beliefs,
which as a matter of fact never seems to be the case. (Notice also
that this is not only a problem for folk-psychological wide content,
but also for a notion of narrow content, at least within the context
of functionalism.) This would imply that general scientific or folkpsychological explanation of human beings in terms of intentional
states would be impossible since in that case “there are no robust
counterfactual-supporting generalizations.”20 Whether or not we are
able to replace the notion of sameness of content with a suitable
notion of similarity of content that is able to overcome the above
objection is at this stage however still an open question.
Teleofunctional theories constitute another option for the reductionist project. Yet they are, I believe, also not able to overcome the
problem of the indeterminacy of conceptual content and they are
until now not able to identify the content of a mental representation
with the specificity required for the content of a de dicto belief.
These last claims about the difficulties facing the prospect of trying
to reduce intentionality require certainly a much more detailed argument, but they suggest that we cannot take the possibility of a reduction of intentional properties for granted. If reduction turns out to be
impossible and we cannot answer the explanatory exclusion argu-
phil8321.tex; 8/09/2003; 18:50; p.8
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
9
ment, our only option would be an instrumentalist understanding of
our mentalistic explanation of behavior. From my perspective it is
therefore more promising to account for the autonomy and causal
efficacy of mental properties without demanding a reduction.21 I
will argue for such a non-reductive account in the following pages.
2. THE FLIGHT FROM METAPHYSICS INTO THE
EPISTEMOLOGIES OF EXPLANATORY PRACTICES
If the argument in the above section is correct then the problem of
mental causation seems to be intractable within the context of physicalism. In light of this conclusion, one is indeed justified to wonder
whether something has gone wrong with the above argument. For
Baker and Burge, the above conclusion should be better understood
as an argument against the metaphysical framework of physicalism.
Both take a Moorean attitude towards the problem of mental causation. It is more certain that psychological explanations are objective
causal explanations than that the metaphysical framework which
casts doubt on the reality of mental causation is correct.22 For them,
it is wrong to evaluate the objectivity of psychological explanations
according to a conception of causality that is not derived from the
explanatory practice of the special sciences but is based on a certain
materialistic prejudice. A conception of causation can be justified
only insofar as it conforms to our successful explanatory practices.23
To say it more succinctly, Baker and Burge diagnose the obsession
of contemporary philosophy with the problem of mental causation
as a failure to have learned Kant’s lesson that metaphysics requires
epistemology.
Baker and Burge argue for their diagnosis by pointing out that
Kim’s argument would not only put the objectivity of psychological
explanations into doubt. It would also challenge the explanatory
relevance of all the other special sciences, since even neurobiology
or geology are not completely causally closed. Causal closure is
only valid for the microphysical reality described by physics and
quantum mechanics. If one accepts the causal closure of the physical
in this sense it would mean that any physical event has a complete
micro-physical cause and explanation. And then one could construe
an argument that would show that all causal relations postulated
phil8321.tex; 8/09/2003; 18:50; p.9
10
KARSTEN R. STUEBER
in the special sciences like geology, chemistry, and neurobiology
etc. lack objective existence. Instead of accepting such absurd
consequences Baker rejects the causal closure principle.24 Burge on
the other hand seems to accept the closure principle for the physical
realm but he insists that it does not follow that such a closed system
“overrides causal relations or causal explanation in terms of properties from outside the system, ” especially if mental and physical
explanations answer “two very different types of inquiry.”25 Burge
thus seems to reject the principle of explanatory exclusion.
As we have seen in the first section, a certain epistemic conception of explanation, which also emphasizes its context relativity,
does not have to worry about the causal/explanatory exclusion argument. Yet both Baker and Burge seem to argue for the causal
efficacy of mental properties. Burge claims explicitly that psychological and physical explanations “are explaining the same physical
effect as the outcome of two very different patterns of events.”26
In this case, we need to explicate why two independent explanations of the same phenomenon can be both accepted as objectively
true. However, the worry of whether the causal efficacy of mental
properties is preempted by neurobiological and physical properties
is independent of the problem of whether or not Kim’s exclusion
argument generalizes to all macroscopic properties postulated by the
special sciences or even to all properties as Block suggests.27 One
might be able to address the generalization problem without having
answered the question of why mental properties should be regarded
as having autonomous causal powers. For that purpose it might be
sufficient to acknowledge as Kim has recently done that certain
macro-properties have non-reducible causal powers, “powers that go
beyond the causal powers of their micro-constituents.”28 For Kim
such acknowledgment is compatible with his general physicalistic
commitments, because such properties are micro-based and can be
completely characterized in terms of their microstructure. They are
completely analyzable into their parts and the relations between
these parts without their causal powers “draining” – to use Block’s
metaphor – to the next lower level. The mass of a large object,
for example, can be completely accounted for in terms of the mass
of its micro-constituents. Yet as anybody knows, who had a heavy
object falling on his foot, its mass has causal powers that go beyond
phil8321.tex; 8/09/2003; 18:50; p.10
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
11
the causal powers of its constituents. Kim also insists that the relation between mental and neurobiological properties does not track
the distinction between micro-macro levels.29 Mental and neurobiological properties are properties of a person and they should therefore be regarded as belonging to the same ontological level. For that
very reason they are still within the scope of the causal explanatory
exclusion argument.
It is not important to decide in this context whether or not Kim’s
response is ultimately successful in answering his critics that he has
as much difficulties to account for ordinary macro-causation as for
mental causation.30 His response indicates however that the generalization problem has to be distinguished from the specific problem
of mental causation we are focusing on in this article. One can also
support this conclusion on a more intuitive basis. One only needs
to assume that our account of bodily movements is explanatorily
and causally closed insofar as neurobiology, chemistry and physics
are concerned. This seems to be a reasonable assumption in light of
our contemporary scientific practices. If this assumption is granted,
however, the explanatory exclusion problem as elaborated in the
prior section arises.
Consequently, Baker and Burge are wrong to claim that Kim
adheres to a pre-Kantian conception of how one should do metaphysics. One rather should understand Kim’s concerns as being
based on a plausible conception of explanation and on certain metaphysical assumptions that are indeed basic presuppositions of our
scientific practices. Within these practices we accept, for example,
that physics has a certain distinguished status insofar as its scope
is universal and it describes the most basic causal processes constituting the universe. It is therefore unfair to depict Kim’s concerns as
being raised outside the realm of epistemology and of our successful
scientific practices. He is better understood as pointing to difficulties
and inconsistencies within the context of our explanatory practices.
Kim is justified in insisting on the metaphysical question of how we
can conceive of mental causation as being possible.31
Nevertheless, if these metaphysical worries arise within the
context of our explanatory practices, then they are also constrained
by the parameter of these practices. As I will argue in the next
section, we should take worries about epiphenomenalism seriously,
phil8321.tex; 8/09/2003; 18:50; p.11
12
KARSTEN R. STUEBER
only if they can be substantiated within the context of the same
explanatory practices that give rise to these worries. We have to
have reasons to believe that, given the way that our various explanatory practices individuate causal efficacious properties and their
explananda, explanations of one practice make other explanations
superfluous. In arguing for the dual explanandum strategy, I will
suggest that this is not very likely. On the other hand, epiphenomenalist worries from a God’s eye perspective, in light of a unified
science – “we do not know what” – do not qualify as serious metaphysical doubts. The adequate philosophical response in regard to
such worries is neglect.
3. THE DUAL EXPLANANDUM STRATEGY AND THE CAUSAL
INHERITANCE PRINCIPLE
For that very reason, there is something right about Baker and
Burge’s insistence that in order to understand questions about causation and explanations one has primarily to focus on our successful
explanatory practices. Certain conceptual resources and different
ways of formulating and answering why-questions define explanatory practices. They are able to identify different autonomous
counterfactual patterns or ceteris paribus laws that are at least
prima facie indicative of the autonomous causal power of higher
order properties. To quote Terry Horgan as a representative of this
position:
On this generic approach, causal claims and causal explanations at different levels
of description do not directly compete with each other, because they advert to
different, but equally real and objective, dependency-patterns among properties.
Which pattern is contextually relevant depends on one’s specific interests and
purposes in asking for the cause or for a causal explanation.32
Because of these different conceptual resources constitutive for
the special sciences we also cannot expect to provide a causal
explanation for an economic or social fact by appealing to the
conceptual resources of physics or even neurobiology.33 Within the
realm of physics one cannot even articulate the question and problems for which one is trying to find an economic explanation. As
has often been pointed out, the conceptual resources of the different
sciences criss-cross each other.
phil8321.tex; 8/09/2003; 18:50; p.12
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
13
Nevertheless, it is important to note that merely recognizing
different levels of counterfactual dependence alone is insufficient to
counter the above explanatory exclusion argument. Merely asserting
a difference between levels still does not show us how we should
think of the relation between mental and physical/neurobiological
properties, especially if one agrees as Burge does that psychological and physical/neurobiological explanations explain in general
the exact same phenomenon. The counterfactual approach to the
problem of mental causation is therefore more plausible if one interprets it as part of the so called dual-explanandum strategy, i.e. the
claim that psychological explanations do not in general attempt to
explain the same phenomenon or aspects of the same phenomenon
as physical or neurobiological explanations.34
As a first step of my defense of the dual explanandum strategy,
one has to recognize that Kim distorts the relation between
the scope of neurobiological and psychological explanations by
focusing primarily on the explanations of specific bodily movements. Psychological explanations, at least ordinary belief/desire
explanations normally play a vital role in a much broader context
as in the explanation of actions in a specific historical, economic or
social environment. One is, for example, able to explain the build
up of the German Navy before World War I by pointing to Kaiser
Wilhelm’s desire for a place in the sun and his belief that a strong
German Navy would enable him to acquire such a desired location
and by assuming some rudimentary knowledge of the social and
political structure of the German society at the time. Obviously, the
build up of the German Navy involves a great amount of bodily
movements. Yet a neurobiological explanation of each and every
of these bodily movements does not provide an explanation for the
build up of the German Navy. From a neurobiological point of view
one is not even able to distinguish between the build up of the
German and the British Navy.
Here then is the paradox of psychological explanation that is
created by Kim. His argument suggests that psychological explanations do not have any objective status because they are in competition with and excluded by neurobiological explanations of behavior.
On the other hand neurobiological explanations do not answer the
same why questions and they do not satisfy the same explana-
phil8321.tex; 8/09/2003; 18:50; p.13
14
KARSTEN R. STUEBER
tory interests as psychological explanations. But if explanations are
answers to specific why questions that are grounded in objective
relations then Kim seems to leave us the choice between neurobiological statements, which describe objective relations but do not
answer the why questions we are interested in, and psychological
explanations, which answer our why questions but do not seem to
describe objective relations.
This is not an appealing choice. In light of the above example
it should be clear what has gone wrong with Kim’s argument. The
purpose of a psychological explanation does not in general consist in
explaining bodily movements as such but it explains the successful
or unsuccessful interaction of organisms with their natural, historical and cultural environment. In that respect psychological explanations are similar to explanations of behavior within evolutionary
biology. The evolutionary biologist is not interested in explaining
certain bodily movements of the beaver but he is interested in
explaining the dam building behavior, i.e. a certain interaction
between the organism and the environment.35 There is, however,
no systematic relationship between the type of bodily movement
and the kind of effect it has in any environment. Rather different
bodily movements can have the same effect in similar contexts, and
depending on the environment the same bodily movement can have
different effects in different environments. One can therefore not
expect that neurobiology, which is supposedly able to provide a
complete explanation of bodily movements, explains the successful
interaction with the environment, even if a particular bodily movement brings about such an interaction. Neurobiological explanations
thus do not exclude psychological explanations because they do not
explain the same phenomena or they at least do not explain the same
aspect of the same phenomena.
Kim might even agree with the above claim that neurobiological
explanations are primarily useful for the explanation of behavior
narrowly described, that is, without taking into account its role in
the larger environment.36 Yet he might ask, if we can explain all the
bodily movements involved in a human activity, such as the activity
to climb a ladder or to build a ship, what more do we need to know
in order to describe the causal powers effective for enabling a person
to climb a ladder or to build a ship? Neurobiology does not neces-
phil8321.tex; 8/09/2003; 18:50; p.14
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
15
sarily limit its theoretical scope to specific social environments, but
it can certainly be used for explaining bodily movements in such
environments. If this is the case, why should one not conceive of the
causal powers of mental properties as “piggybacking” on the causal
powers of neural states that are described in neurobiology? Or to say
it differently, if one accepts, as it is certainly plausible within the
physicalistic framework, that each psychological explanation
requires some physical implementing mechanism, why is it not the
case that the psychological properties inherit their causal powers
merely from the physical properties of the implementing mechanism?
For Kim these considerations commit us to accepting the socalled causal inheritance principle, which I articulate for the
purposes of this article in the following manner:
(CIP) The causal powers of an instance of a psychological/mental property M are
identical with all or a subset of the causal powers of the instance of the physical
properties {P1 . . . Pn } that constitute its realizing or implementing mechanism.
To reject the causal inheritance principle seems to imply that the
causal powers of mental properties somehow “magically emerge at
a higher level and of which there is no accounting in terms of lowerlevel properties and their causal powers and nomic connections.”37
And as Kim, in my opinion, rightly complains, what is the guarantee
that the above case does not exemplify the general relationship
between psychological and neurobiological explanations, i.e. what
justifies the claim that counterfactual relations in the psychological
vocabulary do not pick out merely epiphenomenal relations, that are
grounded in the underlying physical mechanism?
Proponents of the counterfactual approach to mental causation, which I utilize here for the dual explanandum strategy,
differ in their reactions to the above worries. Baker responds by
espousing a position of practical realism, according to which the
commonsense conception of reality has “its own integrity” because
commonsense categories are concerned with human flourishing
and the attempt to satisfy non-optional human interests.38 For that
very reason a “vertical integration of knowledge,” even though it
would be welcomed, is neither “necessary nor common.”39 Marras,
who explicitly defends a dual explanandum strategy, seems to be
persuaded by the general plausibility of Kim’s causal inheritance
phil8321.tex; 8/09/2003; 18:50; p.15
16
KARSTEN R. STUEBER
principle and claims that “to ascribe causal powers to a property
is, in general, to posit the existence of physical mechanisms
through which systems instantiating that property discharge their
causal power. The mechanisms explain the causal powers of the
property, they do not preempt them or displace them.”40 Similarly,
Antony and Levine require that autonomous higher order causal
explanations are true only in virtue of “mechanisms that are ultimately grounded in the basic properties and objects of fundamental
physics.”41 Yet they insist that knowledge of such metaphysical
grounding is not required for being justified in accepting psychological explanations, which support counterfactuals. Within this
spectrum, Horgan seems to occupy a middle position. He does not
regard the causal-explanatory relevance of mental properties and
psychological explanations to be grounded in physical mechanisms,
even though he requires that some such physical mechanism exists.
But Horgan insists that the “higher level generalizations must be
non-coincidental, when viewed from a lower level perspective.”42
For that purpose, he has some sympathy for an account of mental
content in terms of relational proper function. It allows us to
regard mental content as a naturalistically “kosher” notion and
generalizations in the mental idiom as not merely coincidental.
As I will elaborate in the following, the above positions can
each be understood only as a partial response to Kim’s concerns.
The fundamental problem of the debate about mental causation
consists in the fact that all sides take very different examples to
be paradigmatic for the relation between psychological and neurobiological explanations. They all thus nourish themselves from a
one-sided diet. Baker’s focus on particular autonomous cases of
folk-psychological explanations does not allow her to take the question of the constitution of causal powers by lower level properties
seriously enough, even though, as has been just shown, that question
is prima facie plausible in some cases of psychological explanations. On the other hand, Marras, Antony and Levine’s claims that
an explanation of the causal powers of mental properties in terms
of lower level properties does not threaten the causal explanatory
autonomy of psychological explanations is valid only for a limited
range of cases, since the scope of the causal inheritance principle
is more limited than they and Kim suggest. As I will show, it
phil8321.tex; 8/09/2003; 18:50; p.16
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
17
can be plausibly applied only in situations in which the causal
powers of higher order properties can be explained through the
implementing mechanism or the interaction of such mechanisms in
a certain environment. Yet, Marras, Antony, and Levine are right
that this, contrary to Kim’s interpretation of the causal inheritance
principle, does not automatically imply that the science of the implementing mechanism describes all of its specific causal powers. In
none of these cases do we however necessarily need, as Horgan has
suggested, a further explication of the higher order property in terms
of the lower order property.
I will illustrate these claims with two examples. Take Van
Gulick’s example43 of the color changing mechanism of a
chameleon, which he uses to argue that wide content can be understood as having causal powers that are not identical with or supervene on the causal powers of internal physical states. Chemistry and
physiology might be able to explain the causal mechanism responsible for the color changing mechanism of a chameleon, but that
such a color changing mechanism has the causal power of helping
the chameleon to avoid detection by predators is not recognized
on the chemical and physiological level. It is a mistake to think
that the science, which explicates the color changing mechanism,
also explains the survival of chameleons in certain environments.
The causal power of being able to escape detection by predators
is not merely constituted by the color changing mechanism but by
having this specific color changing mechanism in a certain environment with predators of specific sensory capacities. In a different
environment the color changing mechanism would not have such
a causal power. One could therefore argue that biology describes
different causal powers than physiology, even though the causal
power to avoid detection requires a physical implementing mechanism. Similarly, psychological explanations such as “He went to
the kitchen because he wanted to drink a beer” or “He opened
the box because he wanted to make tea” are not in competition
with neurobiological explanations because psychological explanations refer to the causal powers an organism has within a specific
environment. As the discussion of about the externalistic constitution of mental content has revealed, psychological explanations
take as their primary subject of explanation the person in so far
phil8321.tex; 8/09/2003; 18:50; p.17
18
KARSTEN R. STUEBER
as that person is interacting with a certain environment and whose
behavior is directed towards that environment. Neurobiological
mechanisms are necessary preconditions for the possession of these
causal powers that persons have in certain environments and that are
described in the mental idiom. Yet neurobiology is not able to reconstruct these causal powers solely within its domain, which abstracts
from the embeddedness of the organism in a certain environment.
Nevertheless, the above example is a case in which we should
find it plausible to apply the causal inheritance principle or a
plausible extension of it. Even though it is correct to maintain
that explaining the color changing mechanism does not explain the
capacity to escape detection in certain environments, that capacity
can be easily explained by or derived from knowledge of the color
changing mechanism, knowledge of common sense physical facts
of the environment and knowledge of the sensory mechanisms of
potential predators. These are strictly speaking the implementing
mechanisms of the capacity to escape detection in a certain environment. Similarly, knowledge of the neurobiological mechanisms, the
movements of the body and knowledge of the environment should
allow us to account for the movement towards the kitchen or the
opening of the box.
I would thus suggest the following modification of the causal
inheritance principle in order to analyze better how we should
understand that the causal powers of higher order mental properties
“piggyback” on the causal power of lower order properties.
(CIP*) The causal powers of an instance of a psychological/mental property M are
identical with all or a subset of the causal powers of the instance of the physical
properties {P1 . . . Pn } that constitute its realizing or implementing mechanism,
only if all the relevant effects of M can be fully explained in terms of the causal
implementing mechanisms and other facts of the environment as described in
“physical” terminology by commonsense or scientific theories.
In order to understand CIP* correctly, it is important to note that
I understand “physical terminology” rather broadly as referring to
commonsense terms like “houses” and “cars” and to more technical
terms like “cells,” “neurons,” and “quarks.” It only excludes terminology that essentially and irreducibly presupposes the existence of
mentality. In that case we certainly could not speak of the causal
powers of higher order mental properties as merely piggybacking
phil8321.tex; 8/09/2003; 18:50; p.18
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
19
on lower order physical mechanisms. Notice also that the notion of
explainability implicit in my formulation of CIP* is squarely situated in the epistemic context of our scientific practices and does not
appeal to the practice of a god like intellect not accessible to us.
It is only our scientific practices that provide us with the necessary conceptual and theoretic resources that allow us to identify
the relevant underlying causal mechanisms. Even if our scientific
practices might at the same time reflect our cognitive limitations
it is only from the perspective of these practices that the issue
of explanatory exclusion arises and not from the viewpoint of the
cosmic exile.
According to CIP* it is indeed the case that the causal powers
of the chameleon’s survival is in a broad sense piggybacking on
lower order causal mechanisms.44 Yet one has to stress against
Kim that strictly speaking, it is still true that the science that
explains the color changing mechanism, does not alone explain
the causal power of escaping detection in a certain environment.
One has to appeal to knowledge outside of its specific domain in
order to do that. For pragmatic reasons it thus might make sense
to establish a “discipline” with its own terminology that focuses
only on the interaction of certain physical mechanisms in specific
environments and that describes the causal powers such systems
have within this environment. Whether or not this implies that
the relationship between psychological explanations and neurobiological explanations of behavior should be seen at times in
analogy to the relationship between the basic sciences and the so
called applied sciences – like between physics/chemistry and the
engineering sciences that deal with the causal properties of cars,
airplanes and batteries – instead of being conceived of as being
similar to the relation between physics and biology, I will leave
undecided. It is, however, important to stress that nothing so far
forces me to accept that psychological explanations appeal to causal
powers that emerge somehow magically from the physical realm.
They describe real causal powers that physical entities have in a
certain environment, but that are not necessarily described by neurobiology because its explanatory interest is not mainly focused on the
interaction between that physical entity and the environment.
phil8321.tex; 8/09/2003; 18:50; p.19
20
KARSTEN R. STUEBER
The above example represents only a partial range of psychological explanations. Take the following examples: “I signed the
mortgage papers because I wanted to buy a house” or “President
Clinton signed legislation proposed by Congress into law because
he wanted to end welfare as we know it.”45 Examples like these
ones are different from the cases discussed so far and cannot be
seen in analogy to the chameleon case, because they involve an
essential mental and intentional dimension. And as I will show, it
is exactly for this reason that CIP* does not apply in these cases. No
knowledge of a broad range of external and non-intentional factors
of the environment together with knowledge of the neural states of
the brain that caused the movement of the hand signing the contract
or the law will enable me to explain why that movement of the hand
has the causal power of entering into a certain contract with the bank
or of creating a new law of the country.
Proponents of the principle of autonomy like Stich and Kim,
however, might deny the last claim. The principle of autonomy
states that “explanatory psychological properties and relations
are supervenient upon the current internal physical properties of
organisms.”46 According to the intuitions supporting this principle,
psychological explanations have to be able to explain my behavior
and the behavior of my exact physical replica, since we both should
behave in an identical manner as long as we describe our behavior
autonomously or narrowly. President Clinton’s and my physical
replica would thus “sign” the exact same papers. Proponents of the
principle of autonomy readily admit that in a certain sense President
Clinton and his physical replica do not show the same behavior if
we individuate it widely. Only President Clinton’s signature has the
power to change proposed legislation into the laws of the country,
in the same way that only I (and not my physical replica) can sell
the car I own. President Clinton’s replica is not President of the
United States, since he was not elected, and my replica is not the
owner of my car since he did not buy it. Nevertheless this does not
necessarily show that there is a psychological difference between me
and my replica. All that is required for understanding the difference
in wide behavior is to recognize that our behavior is embedded in
an environment with different historical, social and legal facts. As
Stich suggests in order to explain why I sell my car, we first
phil8321.tex; 8/09/2003; 18:50; p.20
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
21
need an explanation of the behavior I exhibited while selling my car, autonomously described. Second, we need an account of why that behavior, in those
circumstances, constituted the act of selling my car. The second component of
the story will marshal historical facts, legal facts and perhaps other sorts of facts
as well.47
If these intuitions in favor of the principle of autonomy hold
up, then psychological explanations would seem to be preempted
by neurobiological explanations. The causal powers of psychological properties would piggyback on the underlying physical
mechanisms, at least in the sense articulated by CIP*.
Nevertheless, proponents of the principle of the autonomy are
mistaken in thinking that explanations of behavior can be neatly
separated in an internal component that is supervenient on the
internal physical states of an organism and external facts of the
environment that would allow CIP* to be applied. As particularly
Baker has pointed out, talking about social and legal institutions for
example has “intentional presuppositions.”48 Moreover the behavior
that we want to explain like selling a car or signing mortgage papers
is constituted partly in reference to intentional attitudes that are
widely individuated and that we assume to be causally involved
in the production of the behavior.49 It is, for example, part of our
very conception of morally responsible and legally binding behavior
that the agent knows what he is doing. If a bank employee tricks
me into signing mortgage papers while I was thinking that I was
signing a deposit slip, then I did not sign mortgage papers in any
legally binding manner. This is however not merely a difference
between external factors in the environment. How I conceive of my
action makes a difference of how we have to characterize my action.
These actions require that I have a certain conceptual repertoire that
contains a minimal understanding of the notion of a contract and the
notion of a mortgage, ownership, law, and other relevant notions. For
that very reason we also do not think of children as being capable
of signing contracts, since they lack the required mental maturity
and conceptual sophistication. The attribution of such concepts and
beliefs involves attributing content that is widely individuated. In
order to have such intentional attitudes it is necessary that I have
a history of living in a certain social and natural environment. In
this context it also does not seem to be of much help to try to
phil8321.tex; 8/09/2003; 18:50; p.21
22
KARSTEN R. STUEBER
come up with a narrow equivalent of concepts like mortgage and
contract, since for my action to be individuated in a certain manner
it is required that I have an adequate understanding of the environment I am acting in and not merely a narrow equivalent of that
understanding.
Here, we have a clear case of irreducible cross-classification
and robust counterfactual relations, which cannot in principle be
explained in light of further knowledge of the wider environment
without essential reference to mentality. For that very reason we
can also not speak of causal powers of psychological states as
merely piggybacking on underlying physical mechanisms. I would
also suggest that the cross-classification between social and intentional properties on the one hand and physical properties on the
other has to do with the fact that intentional and social categories
both delineate properties that have an essential historical aspect,
if viewed from the perspective of their supervenience base. As
has been often pointed out, in order to attribute a specific intentional state, which is individuated widely, we have to assume a
certain history of causal interaction between an individual and the
substances in his environment. Similar observations apply to social
categories that clearly define causal powers of certain individuals
like “President of the United States” or “German Kaiser.” An individual can be a US President only if he went through a certain causal
history of being elected. In order to be President of the US that
election also has to be part of the wider historical context, in which
the US established itself as an independent nation.50
It is thus hardly surprising that the causal power of an individual
in the social realm cannot be described on the physical level.51
(Only philosophers do indeed legitimately wonder about such causal
powers.) Within the realm of the physical sciences – at least an
ideal physics and chemistry – we assume that the physical states
of the world at t1 on can be explained in light of our knowledge of
the fundamental law of nature and our knowledge of the states of
the world at time t0 (t1 > t0 ).52 Even if a certain physical history
is nomologically necessary in order for a world to be in a certain
physical state at t0 , this does not seem to be conceptually necessary
for explaining the physical states of that world at t1 . Indeed for the
purpose of these explanations it does not seem to matter that much
phil8321.tex; 8/09/2003; 18:50; p.22
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
23
whether or not the physical world has a particular causal history.
Assume for a moment that there is a world that was created five
minutes ago and was at that time physically indistinguishable from
our world. What we surely know is that this world cannot be physically distinguished from our world right now, yet it will not have a
President of the US and its inhabitants will lack certain intentional
states. Certain psychological and social explanations do not apply
and are only true of the world with a certain causal history. However,
insofar as the time between t0 and t1 is concerned, the same physical
explanation applies to both worlds.
Does this position violate the general framework of physicalism
by denying that higher order properties supervene on lower order
physical properties? This is for Kim the minimal condition any
physicalist must accept. Moreover does the position argue for a
worrisome version of dualism, in which the causal power of higher
order properties somehow appears magically, since it cannot be
explained through lower order properties? Kim might also ask
whether it is indeed conceivable that an exact physical duplicate of
our world lacks any other causal power that we describe with our
social and intentional concepts.53
Let me briefly address these worries. I deny that psychological
properties supervene on the brain states of persons. This however
does not deny that social and intentional properties supervene
globally on physical properties, at least if one includes the physical
history of the world as part of the supervenience base. It also does
not exclude the possibility that mental properties and social properties might supervene on more specific but “larger spatio-temporal”
regions, which Horgan characterizes as “regional supervenience.”54
Yet, given the above argument and the different explanatory structure of physical explanations and psychological explanations, we
should not expect that such supervenience could be explained in
terms of lower level properties and categories that individuate causal
mechanisms on the lower level. Rather from the perspective of the
lower level sciences the supervenience base for social and intentional properties would have to be regarded as gerrymandered. Even
if psychological explanations require certain physical mechanisms,
this does thus not preempt psychological and social properties from
phil8321.tex; 8/09/2003; 18:50; p.23
24
KARSTEN R. STUEBER
having their own causal power, since they do not fall within the
scope of the causal inheritance principle.
If this is a version of dualism, it is neither mysterious nor worrisome from the physicalist perspective. The causal efficacy and
explanatory relevance of mental and social properties has to do with
the way the world is and the way we divide up explanatory tasks
between the various disciplines. Kim’s frequent suggestion that a
world, which is an exact physical duplicate of our world, has to
have the exact same causal powers is a rather ambiguous manner
of probing one’s physicalist intuitions. If it means that a world,
which God created ex nihilo five minutes ago, has the exact same
powers, I would suggest that a physicalist does not have to agree
with that intuition. There is indeed a physical difference between
it and our world. It lacks our physical history and origin, even if
that does not make a difference to explanations on the physical
level. It is not clear how seriously a physicalist has to take such
thought experiments, since it is nomologically not very likely that an
exact physical duplicate of our world could exist without having our
physical history. The physicalist is only committed to the claim that
worlds that are physically identical and have the exact same physical
history are identical in their causal powers. Moreover, physicalism
is compatible with the claim that having a certain physical history
makes a difference to the causal properties of events. But given the
manner in which we divide up the disciplines and given the way
the world is – i.e. a world in which difference in physical history
makes a difference in causal powers, it is not surprising, even for
a physicalist, that physics does not describe and explain all causal
powers.
Accordingly, the denial that the causal inheritance principle can
be plausibly applied in all contexts does not imply that higher
order properties have magical powers. My position only entails that
certain physical entities – in certain physical environments with a
certain physical history – have causal powers that are not described
by the physical-biological sciences, given the explanatory interests
and conceptual resources of these sciences. From that conclusion
however it does not follow that we have to assume that the world is
partly a non-physical world, one rather has to conclude that physics
does not describe all the causal powers of the physical objects in
phil8321.tex; 8/09/2003; 18:50; p.24
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
25
the world. It also does not imply that the recognition of the interest
relativity of explanations commits us to a purely epistemic and
non-realistic conception of explanation. Rather the special sciences
describe objective causal powers and causal relations that we are
able to recognize only within the context of the conceptual resources
of the particular science.
The above strategy to counter the explanatory exclusion argument, consists in denying that psychological and neurobiological
explanations in general explain the same phenomena, i.e. my
response to Kim is part of the dual explanandum strategy. So far
I have not addressed the cases in which we seem to explain mere
bodily movement like the raising of a hand in a psychological
manner. And these cases of direct mental to physical causation are
the cases that Kim focuses on, since they are in direct violation of
the physical closure principle.
In light of the above considerations, one should first recognize
that these examples of psychological explanations are limited. The
range of normal psychological explanations has to be seen as a
continuum that spans the examples, in which we seem to explain
mere bodily movements, to instances, in which it is still rational
to apply the causal inheritance principle, and cases, in which that
principle does not seem to apply. Yet there are in principle two
ways to address even the cases that Kim primarily focuses on. First,
one could insist that psychological explanations are not intended
to causally explain bodily movements as such, not even in these
cases. They try to explain the bodily movements insofar as they
are directed towards a certain environment. One could maintain that
we do not psychologically explain the hand raising as such but we
explain the hand raising insofar as it is an attempt to be noticed
within the classroom or as far as it is the attempt to participate in a
conversation. I am not sure I am allowed such reasoning, given the
manner in which I explained the limitation of the causal inheritance
principle. But even if one agrees with Kim that in these cases neurobiological explanations do exclude psychological explanations, this
does not imply that neurobiological explanations in general exclude
psychological explanations. There still remains a proper domain for
psychology and folk psychology. That psychological and neurobiological explanations overlap in these cases just seems to be an
phil8321.tex; 8/09/2003; 18:50; p.25
26
KARSTEN R. STUEBER
indication of certain realignments of explanatory domains between
neurobiology and psychology. It shows how much our knowledge of
neurobiology has advanced. Such marginal realignment of explanatory domains has to be expected in the course of the advancement of
scientific knowledge. This however does not justify the conclusion
that neurobiology and the rest of the physical sciences will take over
the whole domain of psychology.
At the end of the article I would like to address briefly a worry
that might arise from considerations of practicing neuroscientists.
So far we have rejected the causal explanatory exclusion argument
because we insisted that psychological and neurobiological explanations explain different phenomena. We seem to have assumed that
neurobiological explanations are primarily suited for explaining
behavior narrowly described but not actions of social agents. In
arguing in that manner we followed the dialectic of Kim’s explanatory exclusion argument and his implicit commitment to the principle of psychological autonomy. Yet neuroscientists do not seem
to explain behavior only in the narrow sense. Neuroscientists study
social emotions, they study socio-pathologies, mental illnesses such
as depression, the neural mechanisms of social interaction and
recognition, and so on. Nothing I have said is meant to exclude this
neuroscientific research. Indeed we do want a better understanding
of the underlying causal mechanisms of such complex and widely
individuated behavior. We probably should also expect some further
realignment between neuroscience conceived in this manner and
psychology using the intentional idiom. This seems to be especially
true for our understanding of the causes of certain mental deficiency
like autism, which in the end could be due to an abnormality of
the underlying neurobiological mechanism. Yet if my considerations so far are correct then such neuroscientific research should
not be expected to completely replace explanations in our normal
psychological idiom. If I am right in claiming that we cannot reduce
our ordinary notions of meaning and content to neurobiological and
physical states and we need to appeal to these notions in order
to explain certain social behavior, then neurobiology even widely
construed will not be able to explain why a certain behavior has
certain causal powers in certain environments with the specificity
required for serving our explanatory interests. Instead of having to
phil8321.tex; 8/09/2003; 18:50; p.26
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
27
worry about explanatory exclusion in this context, we can welcome
the additional knowledge that neurobiology will provide and that
will enrich our ordinary psychological knowledge of the world.
4. CONCLUSION
As I have argued, it is a fundamental mistake of the contemporary debate about mental causation to be focusing only on a few
examples, which are particularly instructive for one’s own reductionist or anti-reductionist position. Focusing merely on cases of
a psychological explanation of bodily movements indeed does not
allow us to recognize that for a core part of folk-psychological
explanations the dual explanandum perspective is still a viable
option. Such psychological explanations are interested in explaining
human agency in a wider social and natural context. As we have
also seen, for such explanations the causal inheritance CIP or its
extension CIP* does not apply because the social and intentional
categories that are used to define this domain are intractably linked
and have a wide historical supervenience base.
Another mistake of the debate in mental causation consists in
trying to argue either that none of our mental explanations can be
regarded as describing objective causal powers in the world or that
all of them do so. Indeed given that common sense and scientific
theory constitute an epistemic continuum and given what we know
about the advancement of science so far, such all or none positions
are not very plausible. Nevertheless , if I am right, it is also not very
plausible that explanations in the mental idiom can on the whole be
eliminated, reduced to or be regarded as merely piggybacking on the
necessary physical implementing mechanisms.55 Rather then taking
this fact as problematic, I suggest that the success and permanence of explanatory practices, which is compatible with lower level
sciences, since they explain different phenomena, should be taken
as evidence for the correctness of a realist interpretation of such
practices.56
phil8321.tex; 8/09/2003; 18:50; p.27
28
KARSTEN R. STUEBER
NOTES
1
For that very reason, some philosophers have recently argued that one should
reassess the recent orthodoxy that psychological explanations are causal explanations. See for example S. Sehon, 2000. I on the other hand am still persuaded by
Davidson’s arguments in “Actions, Reasons, and Causes.”
2 Baker, 1993, p. 91.
3 In the literature one can find numerous versions of the dual explanandum
strategy. As I will show in the third section it is in my opinion fruitful to conceive
of the counterfactual approach to mental causation in this manner. For short
summary of the literature in this context see note 34. Dretske is another prominent
proponent of the dual explanandum strategy. For him, mental events are not
efficient causes but structuring causes. See Dretske 1988 and “Mental Events as
Structuring Causes of Behavior” in Heil/Mele, 1993. For a discussion of Dretske
see also Kim “Dretske on How Reasons Explain Behavior” in his 1993. Dretske’s
conception of mental causation depends on a reductive and teleosemantic account
of mental content. I am however skeptical that such an account can succeed. See
my 1997b. I will attempt to show that the dual explanandum strategy is plausible
even without requiring a reductive account of content. Also note that I am using
the term “neurobiology” very liberally. The term refers to all sciences that could
provide explanations of behavior using only the physical, i.e. non-mentalistic,
idiom.
4 Kim, 1996, p. 206. Kim himself shows more sympathy for the dualexplanandum approach in the context of trying to provide an account of the
“nature of rationalizing explanations” insofar as they are not “causal-predictive.”
See Kim, 1984.
5 For this argument see especially Kim, “The Myth of Non-Reductive Materialism”, p. 279ff and “Mechanism, Purpose, and Explanatory Exclusion”, esp.
p. 254ff. Both articles can be found in Kim, 1993. If it is not indicated otherwise,
all of the citations of Kim’s work refer to this book.
6 See Kim, “Mechanism, Purpose and Explanatory Exclusion”, p. 256 and Kim,
1988.
7 Kim, “The Myth of Non-Reductive Materialism”, p. 279.
8 Kim, p. 280.
9 Kim “The Myth of Non-Reductive Materialism”, p. 281. See also Kim, 1998,
pp. 44/45.
10 See Kim, “Mechanism, Purpose, and Explanatory Exclusion”, pp. 257/258.
11 A word of caution: My purpose here is not a Davidson exegesis. I think
however that these considerations constitute the best response to the explanatory
exclusion argument that Davidson has. I also agree with Crane that the problem
of mental causation as articulated by the causal/explanatory exclusion argument
is not Davidson’s problem. See Crane, 1995, p. 226ff. Yet Crane fails to consider
the epistemic version of Kim’s argument. Some of Davidson’s own remarks come
close to the epistemic response as I have described it. See, for example, his 1993,
p. 16. Yet Davidson himself muddies the water by trying to account for the rele-
phil8321.tex; 8/09/2003; 18:50; p.28
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
29
vance of mental “properties”/predicates through a weak supervenience relation
between physical and mental “properties.” Furthermore, in Davidson’s writing
even the epistemic autonomy of psychological explanations seems sometimes
in doubt. In “Mental Events” he says that psychological generalizations provide
merely evidence for strict laws in the physical idiom. If one conceives of the
relation in this manner, it seems that the mental does not possess any explanatory
autonomy. See Horgan, 1989, p. 53. Again in Davidson (1995, p. 276), he suggests
that we can formulate strict laws only in the context of an explanatory practice
that is not constrained by any practical interest in control but is guided by the
theoretical interest in truth. But if this were the case, then Davidson would have
to admit that psychological explanations are merely pragmatically interesting,
whereas only physical explanations are objectively true. Besides the explanatory
exclusion problem, Davidson’s position of anomalous monism faces other problems in accounting for the causal character of rationalizing explanations. See L.
Antony, 1989.
12 See also Kim, 1993b, pp. 21/22.
13 See Thagard, 1999.
14 See Kim, “Postscript”, p. 361.
15 There however still remains the question whether an identity solution saves the
causal efficacy of the mental or whether it merely “buys the reality of the mental
at the cost of its autonomy” (Antony/Levine, 1997, p. 84).
16 Kim, 1998, pp. 94/95. See also Kim, “Multiple Realization and the Metaphysics of Reduction”, p. 328 and “Postscript”, p. 362ff. J. Heil also favors
this solution in his 1992. For a more detailed account of what functionalization
involves see Kim, 1999, pp. 10–11. It is important to note that Kim somehow
modifies and clarifies his physicalism in his 1998 in order to avoid a generalization
of the explanatory exclusion argument to all properties specified by the special
sciences. Since my primary focus is the problem of mental causation, we can
abstract from this complication of Kim’s position at the moment. I will briefly
elaborate on it in the next section of this article.
17 Kim, 1993a, p. 99.
18 Kim, 1998, p. 101.
19 See for example Putnam, 1988.
20 Fodor/LePore, 1992, p. 15ff. This is all very contested philosophical territory.
Kim acknowledges as much in his 1996, chap. 8, but seems to be persuaded by
the supervenience intuition that a functionalization has to be possible in principle.
As his article “psycho-physical supervenience” (in Kim, 1993) reveals, he also
seems to have strong sympathies for some notion of narrow content. For a critique
of Fodor see Block 1993. As an introduction into this subject matter consult
Block’s entry “Holism, Mental and Semantic” in the Routledge Encyclopedia of
Philosophy. I tend to agree with Fodor and LePore that holism has disastrous
consequences within the context of a conceptual role semantics. I, however,
argue that one has to distinguish between an empiricist version of holism, which
derives from Quine’s “Two Dogmas of Empiricism,” and the holism of radical
interpretation. Only the second form of holism is implied by our folk notions of
phil8321.tex; 8/09/2003; 18:50; p.29
30
KARSTEN R. STUEBER
meaning and content. It furthermore does not have the aforementioned disastrous
consequences. For these distinctions see my 1997a. For a critique of the idea
that our ordinary notions of meaning and content imply holism and an interesting
development of a molecularist position see also Devitt, 1996.
21 For a critical discussion of Dretske and Millikan’s teleo-semantic theories see
my 1997b. For a critical discussion of Dretske’s position see also the articles in
Br. McLaughlin (ed.) (1991), especially the articles by Kim and Horgan. Interestingly, Kim recognizes holism as a problem for the causal-correlational approach
to content. See his 1996, p. 193.
22 See Baker, 1995, p. 119; Burge, 1993, p. 117.
23 Burge, 1986, p. 18.
24 Baker, 1993, p. 92.
25 Burge, 1993, pp. 102 and 116. For a critique of Burge see also Grundmann,
1999. Terry Horgan argues for a similar as Burge position in his 1993.
26 Burge, 1993, p. 116.
27 See Block, 1990, p. 168 and Block, forthcoming.
28 Kim, 1998, p. 85. For some implications regarding the issue of downward
causation see Kim (1999). In my explication of Kim’s position I have also profited
from Ned Block’s “Do Causal Powers Drain Away” (forthcoming) and from the
anonymous referee’s comments.
29 Kim, 1998, pp. 86/87.
30 For some objections to Kim see Block (forthcoming).
31 See Kim, 1998, pp. 60–67.
32 T. Horgan, 1997, p. 179. See also his more extensive description of his compatibilist position in Horgan, 1993.
33 See Baker, 1995, p. 128ff.
34 As far as I know, only Marras sees his counterfactual approach to mental
causation directly as part of the dual explanandum strategy. See his 1998. I will
explain in the following where I differ from Marras. Even though Baker does
not label her approach in this manner, it seems to me that she is effectively
arguing for it in suggesting that physical explanations do not provide “deeper”
explanations of the same phenomenon as intentional explanations. See her 1995,
especially p. 135 and 1999, p. 8. Horgan comes close to it by suggesting that
different counterfactual patterns concern a phenomenon and its cause insofar as
they instantiate different pairs of properties. Burge (1986) and (1989), esp. section
V, seems to be sometimes closer to the outlined position as in his (1993). As his
later article indicates, he, however, does not seem to recognize that the emphasis
on explanatory practices is best understood as being part of the dual explanandum
strategy.
35 Cf. Millikan, 1993, chap. 7 & 8.
36 At this stage of the article I am not worrying about whether or not this
characterization of neurobiology adequately describes the actual practice of
neuroscience. One might wonder about this assertion since practicing neuroscientists do not seem to be concerned merely with explaining narrowly described
behavior but are also interested in accounting for emotions, social reasoning and
phil8321.tex; 8/09/2003; 18:50; p.30
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
31
behavior such as homosexuality on a neuroscientific level. See for example the
issues addressed in Schulkin, 2000. Rather, I am following Kim in my argumentative strategy. In the context of his discussion of the causal explanatory exclusion
argument, Kim himself seems to conceive of neuroscience primarily in a narrow
fashion, because he conceives of psychological explanations also as primarily
concerned with explaining behavior narrowly described. If this is the case then
explanatory exclusion seems to be indeed a real danger. See his “Psychophysical
Supervenience” in his (1993) and his 1996, p. 206. I will address some of the
questions regarding broad neuroscientific explanations at the end of this section.
37 Kim, “Multiple Realization and the Metaphysics of Reduction”, p. 326. See
also Kim, 1998, p. 54. Note, however, that for Kim the causal inheritance principle does not apply to micro-based properties. If I understand Kim correctly, the
difference between micro-based higher-level properties and mental higher order
properties consists in the fact that causal powers of micro-based properties do not
magically emerge. We indeed have a full account of these properties in terms of
their micro-structure.
38 Baker, 1999, p. 13ff; Baker, 1995, chap. 8, especially p. 227.
39 Baker, 1999, p. 15.
40 Marras, 2000, p. 158.
41 Antony/Levine, 1997, p. 102.
42 Horgan, 1993, p. 303.
43 See R. van Gulick, 1989, p. 156.
44 CIP* also applies to cases of externalist explanations that Peacocke emphasizes. See his 1993.
45 Insofar as the examples are here concerned, I do not claim any originality.
Signing contracts etc. is an example often used by philosophers.
46 Kim, “Psychophysical Supervenience”, in his 1993, p. 177. For the following
see particularly Stich, 1983, pp. 164–170.
47 Stich, 1983, p. 196.
48 Baker, 1995, p. 74. See also pp. 4 and 75ff.
49 See also Devitt, 1996, pp. 294 and 296 for making a similar point. Devitt’s
examples however seem to allow for an application of CIP*. For a vigorous
defense of wide psychological states as being causally explanatory see particularly Burge, 1986 and 1989.
50 The primary reason why the causal inheritance principle applied to my first
example was precisely because the explanandum of “being able to avoid detection
from predators” can be characterized without essential reference to intentional and
mental categories. It is however worthwhile noticing that this category also does
not seem to have an extensively wide historical supervenience base.
51 This distinction seems to be at least true insofar as the categories of physics
and chemistry are concerned. Certain functional categories in biology seem to
have a similar irreducible character. Pace Millikan, this however does not mean
that intentional categories can be reduced to each other. Even though central
biological and social categories are both irreducible, they do not necessarily
concern the same causal history.
phil8321.tex; 8/09/2003; 18:50; p.31
32
KARSTEN R. STUEBER
52
In this context we can, in my opinion, ignore the distinction between probabilistic and deterministic laws.
53 See Kim, 1998, pp. 69 and 101.
54 Horgan, 1993, pp. 569/570. Note here however, I leave the question of regional
supervenience undecided. In order to make sense of regional supervenience we
would seem to require a functionalization of mental properties in a wide sense
and it is not clear to me that this is possible.
55 Strictly speaking, my argument for the dual explanandum strategy applies
only to psychological explanations in terms of intentional mental states. It seems
that nothing I have said can be directly applied to phenomenal states, at least
if one conceives of them as purely qualitative states that are devoid of intentionality. Again I do not regard this as a shortcoming of my argument, since I
am not persuaded that one can from a philosophical perspective save all of our
commonsense psychological explanations from the advancement of the physical
sciences. Maybe one has to make one’s peace with epiphenomenalism insofar as
phenomenal states are concerned, if one believes in the irreducibility of qualia. Or
if one believes in a form of reduction, rejoice in the eliminability of psychological
explanations in terms of phenomenal mental states. I myself do not think that
matters are that bleak for qualitative mental states, especially if one does not focus
only on such “purely” phenomenal states as pain but also takes into account more
complex emotional states such as pride, guilt or even more basic emotions such
as fear and sadness. Such states can be only conceived of insofar as they contain
some intentionality. Even “purely” phenomenal states such as having pain seem
to be somewhat integrated within the realm of intentionality. Indeed broadening
one’s examples allows us again to recognize that pain does not only cause me to
wince and cry but it also constitutes a reason for me to consult a doctor or go to
the dentist. In this article, I do not commit myself to any of the above possible
“solutions” in regard to phenomenal states. I hope, however, to have shown that
the dual explanandum strategy constitutes a philosophically adequate response to
the powerful explanatory exclusion argument.
56 Earlier and rather different versions of this article were read at the APA
Central Division meeting, the Southern Society for Philosophy and Psychology,
the philosophy department of City College of the City University of New York
and at a departmental colloquium at the College of the Holy Cross. I would like
to thank the audience at these talks for their probing questions. I would also like
to thank Thomas Grundmann, Bob Feleppa, David Henderson, Michael Lynch,
Noah Latham, Al Mele (my commentar at the APA meeting), Manisha Sinha,
Claudine Verheggen, and especially the anonymous referee for this journal for
their helpful comments and suggestions.
REFERENCES
Antony, L (1989): ‘Anomalous Monism and the Problem of Explanatory Force’,
Philosophical Review 48, 153–187.
phil8321.tex; 8/09/2003; 18:50; p.32
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
33
Antony, L. (1999): ‘Making Room for the Mental’, Philosophical Studies 95, 37–
44.
Antony, L. and Levine, J. (1997): “Reduction with Autonomy’, Philosophical
Perspectives 11, 83–105.
Baker, L.R. (1995): Explaining Attitudes, Cambridge: Cambridge University
Press.
Baker, L.R. (1993): ‘Metaphysics and Mental Causation’, in Heil/Mele (1993),
pp. 75–95.
Baker, L.R. (1999): ‘What is the Thing Called “Common Sense Psychology”?’,
Philosophical Explorations 1, 3–19.
Block, N. (1993): ‘Holism, Hyper-analyticity, and Hyper-Compositionality’,
Mind and Language 8, 1–26.
Block, N. (1990): ‘Can the Mind Change the World’, in G. Boolos (ed.), Meaning
and Method (pp. 137–170), Cambridge: Cambridge University Press.
Block, N. (forthcoming): ‘Do Causal Powers Drain Away?’, Philosophy and
Phenomenological Research.
Burge, T. (1986): ‘Individualism and Psychology’, The Philosophical Review 95,
3–45.
Burge, T. (1989): ‘Individuation and Causation in Psychology’, Pacific Philosophical Quarterly 70, 303–322.
Burge, T. (1993): ‘Mind-Body Causation and Explanatory Practice’, in Heil/Mele
(1993), pp. 97–120.
Corbi, J. and Prades, J. (2000): Minds, Causes and Mechanisms, Oxford: Blackwell Publishers.
Crane, T. (1995): ‘The Mental Causation Debate’, Proceedings of the Aristotelian
Society 69(suppl.), 211–236.
Davidson, D. (1980): Essays on Actions and Events, Oxford: Clarendon Press.
Davidson, D. (1995): ‘Laws and Cause’, Dialectica 49, 263–279.
Devitt, M. (1996): Coming to Our Senses, Cambridge: Cambridge University
Press.
Dretske, F. (1988): Explaining Behavior, Cambridge, MA: MIT Press.
Grundmann, Th. (1999): ‘Burge’s Antirealistic Argument Against Epiphenomenalism’, in J. Nidda-Rümelin (ed.), Analyomen 3 (pp. 521–528), Berlin/New
York: De Gruyter.
Heil, J. (1992): The Nature of True Minds, Cambridge: Cambridge University
Press.
Heil, J. and Mele, A. (eds.) (1993): Mental Causation, Oxford: Clarendon
Press.
Horgan, T. (1989): ‘Mental Quasation’, Philosophical Perspectives 3, 47–76.
Horgan, T. (1993a): ‘Nonreductive Materialism and the Explanatory Autonomy
of Psychology’, in St. Wagner and R. Warner (eds.), Naturalism (pp. 295–320),
Notre Dame: Notre Dame University Press.
Horgan, T. (1993b): ‘From Supervenience to Superdupervenience’, Mind 102,
555–586.
phil8321.tex; 8/09/2003; 18:50; p.33
34
KARSTEN R. STUEBER
Horgan, T. (1997): ‘Kim on Mental Causation and Causal Exclusion’, Philosophical Perspectives 11, 165–184.
Kim, J. (1984): ‘Self-Understanding and Rationalizing Explanations’, Philosophia Naturalis 82, 309–320.
Kim, J. (1988): ‘Explanatory Realism, Causal Realism, and Explanatory Exclusion’, Midwest Studies in Philosophy XII, 225–239.
Kim, J. (1993): Supervenience and the Mind, Cambridge: Cambridge University
Press.
Kim, J. (1996): Philosophy of Mind, Boulder: Westview Press.
Kim, J. (1998): Mind in a Physical World, Cambridge, MA: MIT Press.
Kim, J. (1999): ‘Making Sense of Emergence’, Philosophical Studies 95, 3–36.
LePore, E. and Loewer, B. (1987): ‘Mind Matters’, Journal of Philosophy 84,
630–642.
LePore, E. and Loewer, B. (1989): ‘More on Making Mind Matter’, Philosophical
Topics 17, 175–191.
Marras, A. (2000): ‘Nonreductive Materialism and Mental Causation’, Canadian
Journal of Philosophy 24, 465–494.
Marras, A. (1998): ‘Kim’s Principle of Explanatory Exclusion’, Australasian
Journal of Philosophy 76, 439–451.
Marras, A. (2000): ‘Critical Notice’, Canadian Journal of Philosophy 30, 137–
160.
Malcolm, N. (1968): ‘The Conceivability of Mechanism’, in G. Watson (ed.), Free
Will (pp. 127–149), Oxford: Oxford University Press, 1982.
McLaughlin, Br. (1991), Dretske and His Critics, Oxford: Basil Blackwell.
Millikan, R.G. (1993): White Queen Psychology and Other Essays for Alice,
Cambridge, MA: MIT Press.
Peacocke, Chr. (1993): ‘Externalist Explanations’, Proceedings of the Aristotelian
Society 93, 203–230.
Pereboom, D. and Kornblith, H. (1991): ‘The Metaphysics of Irreducibility’,
Philosophical Studies 63, 125–145.
Putnam, H. (1988): Representation and Reality, Cambridge, MA: MIT Press.
Poland, J. (1994): Physicalism: The Philosophical Foundations, Oxford:
Clarendon Press.
Sehon, S. (2000): ‘An Argument Against the Causal Theory of Action Explanation’, in Philosophy and Phenomenological Research 60, 67–85.
Stueber, K. (1997a): ‘Holism and Radical Interpretation’, in G. Meggle (ed.)
Analyomen 2 (pp. 290–298), Berlin/New York: DeGruyter.
Stueber, K. (1997b): ‘Psychologische Erklärungen im Spannungsfeld des Interpretationismus und Reduktionismus’, Philosophische Rundschau 44, 304–328.
Thagard, P. (1999): How Scientists Explain Disease, Princeton: Princeton University Press.
Van Gulick, R. (1989): ‘Metaphysical Arguments for Internalism and Why They
Don’t Work’, in St. Silvers (ed.), Rerepresentations (pp. 151–159), Dordrecht:
Kluwer Academic Publishers.
phil8321.tex; 8/09/2003; 18:50; p.34
MENTAL CAUSATION AND THE PARADOXES OF EXPLANATION
35
Van Gulick, R. (1993): ‘Who is in Charge Here? And Who is Doing All the
Work?’. in Heil/Mele (1993), pp. 232–256.
Yablo, St. (1992): ‘Mental Causation’, The Philosophical Review 101, 245–280.
Department of Philosophy
College of the Holy Cross
Worceste, MA 01610-2395
USA
phil8321.tex; 8/09/2003; 18:50; p.35