Can Non-Reductive Materialism Escape from the Jaws
of Epiphenomenalism?
By
Erica Shumener, New York University
Some philosophers have attempted
to carve out a theory of the mental that identifies mental events with physical
events but does not identify every
feature of a mental event with a physical feature. Positions that identify mental events with
physical events yet deny that every mental feature is reducible to a physical
feature are often considered to be versions of “non-reductive
materialism.” Notably, Donald Davidson popularized
and defended his version of non-reductive materialism, anomalous monism, in a
series of influential articles. While
non-reductive materialism is an appealing view for many reasons, it is unclear
whether it is tenable. Philosophers have
claimed that non-reductive materialism is unable to preserve the causal
efficacy of mental properties. In other
words, they claim that non-reductive materialism capitulates to
epiphenomenalism. In this paper, I will
inquire into whether non-reductive materialism can defend itself against
epiphenomenalism. I will focus primarily
on some key arguments surrounding Davidson’s theory of anomalous monism as well
as on Cynthia and Graham Macdonald’s closely related version of non-reductive
materialism.
I will
begin by summarizing Davidson’s account of anomalous monism and explaining why
some believe that it denies causal power to mental properties. Then I will explore a particular line of
defense, taken by Cynthia and Graham Macdonald, which maintains that the causal
efficacy of mental properties is preserved under this type of non-reductive
materialism.1 The MacDonalds’ argument hinges upon the
acceptance of a specific metaphysical notion of an “event.” They take an event to be a
property-instantiation by an individual at a time. I will argue that their defense of anomalous
monism is incompatible with the property-instantiation account of events they
uphold. I will conclude that their
argument does not protect anomalous monism from the threat of epiphenomenalism:
if the causal efficacy of the mental is to be preserved under anomalous monism,
it cannot be preserved by the means they advocate.
In “Mental Events,” Davidson
introduces his theory of anomalous monism in order to reconcile three seemingly
contradictory principles. The three
principles are as follows: “The Principle of Causal Interaction”--which states
that mental events interact causally with physical events, “The Nomological
Character of Causality”--which claims that events related as cause and effect
fall under strict deterministic laws, and “The Anomalism of the Mental”-- which
claims that mental events cannot be predicted or explained on the basis of
strict deterministic laws. According to
anomalous monism, although every token mental event is a token physical event,
mental events cannot be given “purely physical explanations.” Furthermore, anomalous monism claims that
there are no psychophysical laws governing the mental and the physical, meaning
that there can be no reduction of the mental to the physical whereby mental
descriptions can be translated into the physical descriptions of events;
nonetheless, each token mental event has both a mental and a physical
description.2
While it seems plausible to
claim that anomalous monism reconciles these three principles, the way in which
it does so may deprive mental properties of their causal powers.3 If mental events are causally related to
physical events because both events are subsumed under strict deterministic
laws, and only the physical descriptions of mental events fall under these strict
laws, then why should one think that the mental descriptions of mental events
play any role in causal interactions with physical events? If mental descriptions of events, unlike
physical descriptions, do not play a role in causal interactions, this seems to
strongly imply that mental properties of events do not have causal efficacy,
even though physical properties do.
Although it may be too strong to claim that anomalous monism is
intrinsically incompatible with the causal efficacy of mental properties, this
account of anomalous monism fails to explicitly assign causal powers to mental
properties, and it is unclear whether it can succeed in doing so.4
Donald Davidson defends
anomalous monism against epiphenomenalism in his article “Thinking
Causes.” Davidson responds to the charge
of epiphenomenalism by stating that it is incorrect to claim that events cause
other events “in virtue of” certain properties.
Davidson asserts that, while mental events cause physical events, we
cannot speak of mental events as causing physical events “in virtue of” or
“qua” the instantiation of certain properties.
Hence, anomalous monism does not deprive mental properties of their
causal efficacy, because neither mental nor physical properties of an event
have causal efficacy. Rather, it is the
event itself that has causal efficacy.
In other words: the properties of
an event do not cause other events, the event simpliciter, causes other events.
Davidson’s account of event
causation is highly counterintuitive, for we often want to assert that an event
causes another event in virtue of the first event’s having certain properties.5 For example, let us suppose that you lick a
purple lollipop, and, as a result, your tongue turns purple. It seems as if it is in virtue of the
lollipop’s having a specific property, purpleness, that the event of your
licking it causes your tongue to turn purple.
It also seems as though other properties of the lollipop, such as the
property of the lollipop’s being spherical, do
not play any causal role in turning
your tongue purple. So it is plausible
to claim that an event causes another event in virtue of the first event’s
instantiating certain properties.
Setting this worry aside, even
if Davidson’s account of event causation is viable, it appears that it will be
successful only under certain metaphysical conceptions of events. His account of event causation may be
problematic under many of the leading metaphysical theories of events. One leading theory of events defines an event
to be a property exemplification or instantiation by an individual at a time.6 If one were to adopt this
“property-exemplification” account of events, one would be required to claim
that events cause other events in virtue of, or “qua,” the instantiation of
certain properties. Why? Since an event, according to this account,
just is the instantiation of a property by an individual at a time, when we
speak of events causing other events, we are automatically claiming that
certain property-instantiations cause other property-instantiations. So, Davidson’s account of event causation is
incompatible with the property-exemplification account of events. This is not a problem for Davidson himself,
since he does not adopt the property-exemplification theory of events. Davidson instead conceives of an event as a
basic, non-repeatable particular that can have many non-equivalent
descriptions.7 Nevertheless, if one were to adopt the
property-exemplification theory of events, this account of event causation
would not protect anomalous monism against the threat of epiphenomenalism.
Is there
any way to save anomalous monism from succumbing to epiphenomenalism if one
upholds the property-exemplification account of events? Cynthia and Graham MacDonald argue that there
is such a way. In fact, the MacDonalds
make a stronger claim. In “Mental
Causes and Explanation of Action” and “Mental Causation and Non-Reductive
Monism,” they argue that specific features of the property-exemplification
theory of events are crucial for
ensuring the causal efficacy of mental properties under anomalous monism.8
Let us
examine how the property-exemplification theory of events is supposed to
restore the causal efficacy of mental properties under anomalous monism. The property-exemplification theory of events
defines an event as an object’s or individual’s having or instantiating a
property at a time.9 The MacDonalds claim that the
property-exemplification theory does not prohibit one event from being a
“co-instantiation” of more than one property by an individual at a time. That is, the MacDonalds claim that a single
event can be an instantiation of multiple
properties at a time. As we saw before,
the property-exemplification account of events automatically dictates that one
event causes another event in virtue of the first event’s instantiating a
certain property. Consequently, the
MacDonalds maintain that if an event is a co-instantiation of properties at a
time, then that particular event causes another event in virtue of first
event’s instantiating multiple properties.
Since their version of non-reductive materialism defines a mental event
as one that includes both mental and physical properties, the MacDonalds
believe that a mental event is a co-instantiation of a physical and mental property
at a time. Therefore, a mental event
causes another event in virtue of its being a co-instantiation of both a physical and a mental property at
a time.
The MacDonalds emphasize the
fact that it is the instantiation itself that provides an event with causal
efficacy. If two properties are
instantiated in a single event, then the MacDonalds believe that it is the
co-instantiation of both of the
properties, not an instantiation of either property individually, that makes
the event causally efficacious. Because
a mental event causes another event in virtue of its’ being a co-instantiation
of a mental and a physical property, it follows that the instantiation of both
the mental and the physical property makes the mental event causally
efficacious. So, mental properties are
not epiphenomenal after all.
The
MacDonalds’ utilization of the property-exemplification account of events may
not succeed in protecting anomalous monism from the charge of
epiphenomenalism. I will now raise two
objections to the MacDonalds’ defense of anomalous monism under the
property-exemplification theory of events.
First, I argue that they have overlooked key features concerning the
structure of an event under the property-exemplification theory. Moreover, once these features are taken into
account, it seems as if the property-exemplification theory conflicts with the
main thesis of anomalous monism. Second,
even if there is some way to make the property-exemplification theory of events
compatible with anomalous monism, the MacDonalds still will not be able to
ensure the causal efficacy of mental properties.
While the MacDonalds are correct
to claim that the property-exemplification theory allows properties to be
co-instantiated in a single event, the property-exemplification theory of
events primarily handles events that are single property-instantiations by an
individual at a time. An event that is a
co-instantiation of properties by an individual at a time is a special type of
event under the property-exemplification theory of events. For clarity, I will call an event that is a
co-instantiation of multiple properties at a time a “complex event,” and an
event that is an instantiation of a single property a “simple event.” Multiple properties can be co-instantiated in
a single event; however, the property-exemplification theory dictates that a
complex event that is a co-instantiation of two properties by an individual at
a time can be decomposed into two simple events that are each a single
property-instantiation by an individual at a time.
For example, a ball can co-instantiate the
properties of being spherical and being red at time t. This event is a co-instantiation of two
properties by the ball at a time. But it
follows, on the property-exemplification account, that there are actually two
simple events here in addition to the complex event: the ball’s being red at
time t and the ball’s being spherical at time t. In a sense, the complex event is constructed,
or derived, from these two simple events. While this decomposition of a complex
event into constituent simple events, each of which instantiates a single
property, might be unpalatable because it leads to a “proliferation” of events,
it is nevertheless correct to claim that the property-exemplification account of
events allows, and in certain cases might even require, this decomposition.10
Since a
mental event, under the MacDonalds’ version of anomalous monism, is a
co-instantiation of a mental and physical property by an individual at a time,
it follows that a mental event can be decomposed into two distinct simple
events: the instantiation of a mental property by an individual at a time and
the instantiation of a physical property by the same individual at the same
time.11 Since the mental and physical properties are
not identical, it follows that the MacDonalds need to acknowledge the
occurrence of two distinct simple events in addition to the complex mental
event.
This decomposition of one complex mental event
into two simple events poses a problem for the MacDonalds. How should one characterize these two simple
events, the physical property-instantiation by an individual at a time and the
mental property-instantiation by the same individual at the same time? It seems uncontroversial to claim that the
simple event that is a physical
property-instantiation by an individual at a time is a type of physical event, for if an object or
individual “x” possesses a property of type T, x itself falls under, or belongs
to, type T. Similarly it seems plausible
to consider the simple event that is a mental
property-instantiation by an individual at a time as a type of mental event. It is reasonable to conclude that the
following is a correct characterization of the two simple events: the event
that is a physical property-instantiation at a time is a physical event and the event that is a mental
property-instantiation at a time is a mental
event.
This
characterization of the two simple events as a mental event and a physical
event conflicts with the main thesis of anomalous monism. The main thesis of anomalous monism states
that token mental events are identical with token physical events. If the two simple events are characterized as
a mental event and a physical event, then the MacDonalds are required to claim
that, at least some token mental events are not identical with any token
physical events. They are required to
make this claim because they believe that mental properties are not identical
with physical properties; thus, a mental property-instantiation is not
identical with a physical property-instantiation. So, the mental event that is a mental
property-instantiation by an individual at a time is not identical with the
physical event that is the physical property-instantiation by the same
individual at the same time. However,
the thesis of anomalous monism states that all token mental events are identical with token physical
events. By employing the
property-exemplification account of events along with anomalous monism, the
MacDonalds encounter a conflict: they need to uphold anomalous monism and
affirm that all token mental events are token physical events, and
simultaneously, they are required to maintain that some token mental events are
not token physical events.
There is a
natural response that the proponent of anomalous monism under the
property-exemplification account of events can make: she can claim that a “mental event,” in the
thesis of anomalous monism, refers to the co-instantiation of a mental and a physical
property. The idea is that it is
incorrect to characterize a mental property-instantiation alone as a mental event. The mental property-instantiation is only
a part of a mental event, not a complete mental event itself. I think that this is a reasonable
response. When Davidson or the
MacDonalds refer to a “mental event” in the context of anomalous monism, it is
clear that they do not mean to pick out only a mental property but a physical
property as well. Nevertheless, if one
is to accept the property-exemplification account of events along with
anomalous monism, then it seems as if one must admit that a mental
property-instantiation by an individual at a time is an event. Since the MacDonalds deny the thesis that
mental properties are physical properties, they are forced to claim that the
event that is a mental property-instantiation is not a physical event. Even though The MacDonalds define a “mental
event” as a co-instantiation of a mental and a physical property, they are
still left with a simple event, the mental property-instantiation alone, which
is a non-physical event. Thus, it seems
as if the mental property-instantiation should be classified as a kind of
mental event under the MacDonalds’ view.
So, the proponent of anomalous monism under the property-exemplification
theory of events is still faced with a problem: she cannot maintain that all
mental events are physical events.
Unless one can provide an alternative explanation of the event that is a
mental-property instantiation that does not characterize it as a mental event,
it seems as if the main thesis of anomalous monism is incompatible with the
property-exemplification theory of events.
Nonetheless,
even if the tension between the main thesis anomalous monism and the
property-exemplification account of events can be dissolved, I am still
hesitant to assert that the MacDonalds’ argument preserves the causal efficacy
of mental properties. That is because it
is not obvious that every property that is co-instantiated in a complex event
plays a role in causal interactions. To
clarify, the MacDonalds suggest that mental properties are causally efficacious
because they are co-instantiated along with physical properties in mental
events; moreover, since it is the co-instantiation itself that has causal
power, we can be confident that the specifically mental properties are causally
efficacious because they are part of the co-instantiation.
There are
many events where multiple properties are co-instantiated, but in which not
every co-instantiated property plays a causal role. Consider the event of a rock’s
co-instantiating the property of heaviness and the property of greyness at time
t. Call this complex event “event 1”. Furthermore, let’s suppose that this rock is
sitting on a stack of papers. Then event
1 causes another event, “event 2”: the
event of the paper’s staying neatly stacked in a pile at time t + (1
second). Although it is correct to claim
that event 1 causes event 2 in virtue of the rock’s having the property of
heaviness, it is not correct to claim that the rock’s greyness plays a causal
role in keeping the papers neatly stacked.
So, it follows that not every co-instantiated property in a complex
event is causally efficacious.
If it is not the case that every
co-instantiated property in a complex event plays a role in event causation,
how can we remain confident that mental properties play a causal role in mental
events? The co-instantiation account of
events alone is not enough to protect mental properties from the charge of
epiphenomenalism. The advocate of
anomalous monism under the property-exemplification account needs to provide
reasons why the co-instantiation of mental properties is enough to ensure the
causal efficacy of mental properties. At
the present, no such reasons have been provided.
We can now draw some
conclusions. First, it is clear that the
causal efficacy of mental properties under anomalous monism cannot be preserved
through an appeal to the property-exemplification account of events. Thus, the line of defense the MacDonalds pursue
is unsuccessful, and anomalous monism is still vulnerable to the charge of
epiphenomenalism. Furthermore, not only
does the property-exemplification account fail to rescue anomalous monism from
the jaws of epiphenomenalism, it also looks as if anomalous monism and the
property-exemplification theory of events are incompatible. This, I think, is the more worrisome
conclusion for the non-reductive materialist.
This conclusion entails that, if mental properties are causally
efficacious under anomalous monism, the anomalous monist is required to adopt a
different metaphysical theory of events.
A
plausible alternative theory of events to consider is Davidson’s theory, which
defines an event as a basic non-repeatable particular. However, as we saw above, Davidson’s defense
of anomalous monism under the basic-particular theory of events leads one to
adopt a very counterintuitive and implausible theory of event causation in
order to fend off epiphenomenalism. This
theory of event causation, which claims that an event never causes another
event in virtue of the first event’s having certain properties, may be too high
of a price to pay in order to preserve the causal efficacy of mental properties
under anomalous monism. Of course, there
may be other ways to preserve the causal efficacy of mental properties in
anomalous monism under the basic-particular theory of events or perhaps under a
different metaphysical theory of events
altogether. Nevertheless, none of the
options explored in this paper provide a strongly defensible way to protect
anomalous monism from the threat of epiphenomenalism. □
©Erica Shumener, 2007
1 Davidson’s original account of
Anomalous Monism can be found in “Mental Events,” The Essential Davidson, (Oxford: Clarendon Press, 2006) p. 105-118.
Cynthia and Graham MacDonald’s theory of non-reductive materialism can be found
in two articles: “Mental Causes and Explanation of Action,” The Philosophical Quarterly, Vol. 36,
No. 143, “Special Issue: Mind, Causation and Action,” Apr., 1986, p. 145-158;
and, “Mental Causation and Non-Reductive Monism,” Analysis, Vol. 51, No. 1, Jan. 1991, p. 23-32.
2 Davidson, “Mental Events,” ibid.
The dual mental and physical descriptions that anomalous monism provides for
mental events support all three principles. Causal interactions between mental
and physical events, when examined individually, can be seen as instantiations
of laws. However, these events can only be viewed as being subsumed under laws
when they are examined under their physical descriptions. This view supports
Davidson’s first and second principles. Under their mental descriptions, on the
other hand, these events are not considered to be instantiations of laws
because mental descriptions do not fall under psychophysical laws, or any strict
deterministic laws. Therefore, the anomalism of the mental is preserved as
well.
3 See, for example Jaegwon Kim,
“Can Supervenience and Non-Strict Laws Save Anomalous Monism?,” Mental Causation, ed. John Heil and
Alfred Mele, (Oxford: Clarendon Press), 1995, p. 19-26; and, “The Myth of
Nonreductive Materialism,” Proceedings
and Addresses of the American Philosophical Association, Vol 63, No. 3,
Nov., 1989, p. 31-47. Also, Brian McLaughlin provides a detailed account of the
debate in “On Davidson’s Response to the Charge of Epiphenomenalism,” in Heil
and Mele, p. 27-40.
4 Kim, “Can Supervenience and
Non-Strict Laws Save Anomalous Monism?,” in Heil and Mele, ibid.
5 Ernest Sosa brings up this point
in “Davidson’s Thinking Causes,” ibid., p. 41-52.
6 Jaegwon Kim, “Events as Property
Exemplifications,” Supervenience and
Mind: Selected Philosophical Essays, ed. Ernest Sosa, (Cambridge: Cambridge
University Press, 1993), p. 33-52.
7 Davidson, “Events as Particulars,”
Essays on Actions and Events, (Oxford:
Clarendon Press, 1980), p. 181-188. Prima
facie, Davidson’s basic-particular conception of an event may be compatible
with this account of event causation. Nevertheless, his account of event
causation is still be controversial, regardless of how one conceives of an
event, and it seems as if there is more to say on this issue.
8 First, I should make clear that
the MacDonalds do not defend Davidson’s exact version of anomalous monism. Instead,
they adopt a closely related theory of non-reductive materialism: non-reductive
monism. The primary difference between anomalous and non-reductive monism is
that the latter is not “committed to full mental anomalism.” In other words,
the MacDonalds wish to leave the option open that certain mental properties are
sometimes related in a law-like way to other mental properties and perhaps even
physical properties. Nevertheless, the difference between anomalous and
non-reductive monism is not relevant to this particular discussion since the
MacDonalds’ still uphold a “psychological causal anomalism.” Therefore, to
minimize confusion, I will use “anomalous monism” when discussing both
Davidson’s and the MacDonalds’ versions of non-reductive materialism.
9 Kim, “Events as
Property-Exemplifications”
10 Kim, “Events as Property
Exemplifications”
11 There is some ambiguity as to
what constitutes the “individual” in a mental event. For example, should we
take the individual to be a person, or the person’s brain, or even a specific
part of the person’s brain? These are important questions, but I’m afraid that
I won’t be able to address them in this paper.
Return to the
Table of Contents