You are on page 1of 7

The Problem of Action

Author(s): Harry G. Frankfurt


Source: American Philosophical Quarterly, Vol. 15, No. 2 (Apr., 1978), pp. 157-162
Published by: University of Illinois Press on behalf of North American Philosophical Publications
Stable URL: http://www.jstor.org/stable/20009708
Accessed: 21/10/2010 10:46

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=illinois.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

University of Illinois Press is collaborating with JSTOR to digitize, preserve and extend access to American
Philosophical Quarterly.

http://www.jstor.org
American Philosophical Paper presented at the 52nd Annual Meeting of the
Quarterly
Volume 15, Number 2, April 1978 American Philosophical Association, Pacific Division,
San Francisco, Calif., March 23-25, 1978.

X. THE PROBLEM OF ACTION


HARRY G. FRANKFURT

I certain earlier conditions. It is integral to the causal

problem of action is to explicate the contrast approach


to
regard actions and mere as
happenings
THE between what an
agent does and what merely being differentiated by nothing that exists or that is
happens to him, or between the bodily movements going
on at the time those events occur, but by
that he makes and those that occur without his something quite extrinsic to them?a difference at

making them. According to causal theories of the an earlier time among another set of events
entirely.
nature of action, which currently represent the This is what makes causal theories implausible.
most widely followed approach to the under? They direct attention exclusively away from the
of this contrast, the essential difference events whose natures are at issue, and away from
standing
between events of the two types is to be found in their the times at which they occur. The result is that it is
causal histories: a movement is an their to that a person must be
prior bodily beyond scope stipulate
action if and only if it results from antecedents of a in some
particular relation to the movements of his
certain kind. Different versions of the causal
body during the period of time in which he is pre?
approach may provide differing accounts of the sumed to be
performing
an action. The only
sorts of events or states which must in conditions insist as con?
figure causally they upon distinctively
the production of actions. The tenet character? stitutive of action may cease to obtain, for all the
they
istically share is that it is both necessary and causal accounts demand, at
precisely the moment
in order to determine whether an event when the commences to act.
sufficient, agent They require
is an action, to consider how it was brought
about. of an agent, once the causal ante?
nothing specified
Despite its popularity, I believe that the causal cedents of his
performing
an action have occurred,

approach is inherently implausible and that it except that his body move as their effect.
cannot a of the nature It is no wonder that such theories character?
provide satisfactory analysis
of action. I do not mean to suggest that actions have run of a well
istically up against counterexamples
no causes ; they are as likely to have causes, I suppose, known type. For example : a man at a party intends
as other events are.
My claim is rather that it is no to spill what is in his glass because he wants to signal
part of the
nature of an action to have a
prior
causal his confederates to begin a robbery and he believes,
history of any particular kind. From the fact that an in virtue of their prearrangements, that spilling
event is an action, in my view, it does not follow even what is in his glass will accomplish that ;but all this
that it has a cause or causes at all, much less that it leads the man to be very his makes
anxious, anxiety
has causal antecedents of any his hand and so his matter
specific type. tremble, glass spills. No
In that the essential difference between what kinds of causal antecedents are as
asserting designated
actions and mere happenings lies in their prior necessary and sufficient for the occurrence of an
causal histories, causal theories that actions action, it is easy to show that causal antecedents of
imply
and mere happenings do not differ essentially in that kind may have as their effect an event that is
themselves at all. These theories hold that the causal manifestly not an action but a mere
bodily
move?

sequences producing actions are


necessarily of a ment. The spilling in the example given has among
different type than those producing mere happen? its causes a desire and a belief, which rationalise the
ings, but that the effects produced by sequences of man's spilling what is in his glass, but the spilling
the two types are inherently indistinguishable. They as it occurs is not an action. That
example makes
are therefore committed to
supposing
that a person trouble particularly for a causal theory in which
who knows he is in the midst of performing an action actions are construed as
essentially movements
cannot have derived this knowledge from any aware? whose causes are desires and beliefs by which they
ness of what is currently happening, but that he are rationalised. Similar counterexamples
can
must have derived it instead from his understanding readily be generated to make similar trouble for
of how what is happening was caused to happen by other variants of the causal approach.
*57
AMERICAN PHILOSOPHICAL
I58 QUARTERLY

maneuvers are actions are mere


I shall not examine the various by and those which bodily movements
means of which causal theorists have to without as a criterion the presence or absence
attempted using
of the relevant desire. ... It is true that there are
cope with these In my
counterexamples.1 judgment
are to such various intrinsic characteristics of bodily movements
causal theories unavoidably vulnerable
which do some indication of their classification.
because locate the distinc? give
counterexamples, they a very movement was
For example, complicated
tively essential features of action exclusively in
. . . the
probably produced by a desire. But simplicity
states of affairs which may be past by the time the of a movement does not even make it probable that it
action is supposed to occur. This makes it impossible was not a desire.
produced by
for them to any account whatever of the most
give
Because we cannot find inherent character?
salient differentiating characteristic of action: any
the time a is an action he istic of action which permits us to distinguish it
during person performing
in touch with the movements from mere we must
is necessarily of his reliably bodily movement,
in a certain whereas is necessarily not therefore, in Pears' view, some
body way, he "classify bodily
in touch with them in that way when movements movements as actions virtue of their
of solely by
his body are occurring without his making them. A origins."2
to describing causes prior to Pears observes that the movements of a
theory that is limited correctly
the occurrences of actions and of mere bodily
move? person's body do not definitively reveal whether he
an is performing an action : the very same movements
ments cannot include of these
possibly analysis
a person to may occur when an action is being or
two ways in which may be related the performed
when a mere is occurring. It does not
movements of his body. It must inevitably leave happening
follow from that the to
open the possibility that a person, whatever his this, however, only way
discover whether or not a person is acting is by con?
involvement in the events from which his action
was on
arises, loses all connection with the movements of sidering what going before his movements
the causes from
his body at the moment when his action begins. began?that is, by considering
which they originated. In fact, the state of affairs
while the movements are is far more
occurring
II
pertinent. What is not merely pertinent but deci?
In order to a more way of is to consider whether or not the move?
develop promising sive, indeed,
thinking about action, let us consider the notion that ments as occur are under the person's It
they guidance.
actions and mere happenings are indistinguishable is this that determines whether he is performing an
in themselves. This notion is an important element action. the of whether or not
Moreover, question
in the motivation for causal theories. If it were movements occur under a is not a
person's guidance
thought that actions and mere happenings differ matter of their antecedents. Events are caused to
then it would be obvious that the way to occur states of but an event
inherently, by preceding affairs,
explicate how they differ would be by identifying cannot be guided through the course of its occur?
this inherent difference between them. It is because rence at a distance.
temporal
causal theorists think that there is no other way to It isworth noticing that Pears ismistaken when he
differentiate between actions and mere concedes that
happenings very complicated movements, though
that they seek a differentiating difference among the they may possibly be mere happenings, are probably
events that them. to be classified as actions. The move?
precede complicated
David Pears, who believes that desires play an ments of a pianist's hands and do, to be sure,
fingers
essential causal role in the of actions, that are not mere
production compellingly suggest they hap?
makes this explicit: penings. Sometimes, however, may
complexity
We simply do not possess the general ability to quite as compellingly suggest the likelihood of mere
between those movements which movement. The about of a per
distinguish bodily bodily thrashings
1
For discussion of the problem by adherents to the causal approach, cf. Alvin Goldman, A Theory of Human Action (Princeton,
1970), pp. 61-63; Donald Davidson, "Freedom to Act," inT. Honderich (ed.), Essays onFreedom of Action (London, 1973), pp. 153-154;
Richard The Philosophical Review, vol. 86 (1977), pp. 58-69. Goldman and Davidson believe
Foley, "Deliberate Action," evidently
that the problem of avoiding the counterexamples is an empirical one, which is appropriately to be passed on to scientists. Foley's
"solution" renounces the obligation to provide suitable analysis in another way: he specifies conditions for acting and, when he
that may be met spasms and twitches, he declares that such movements are nonetheless actions if they
recognises they by simply
satisfy his conditions.
2
David Pears, "Two Problems about Reasons for Actions," in R. Binkley, R. Bronaugh, A. Marras (eds.), Agent, Action and Reason
(Oxford, i97i),pp. 136-137, 139.
THE PROBLEM OF ACTION 159

son's an seizure, for example, rence of this movement does not mark the
body during epileptic perfor?
are very complicated
movements. But their com? mance of an action by the person ;his pupils dilate,
plexity is of a kind which makes it appear unlikely but he does not dilate them. This is because the
to us that the person is performing an action. course of the movement is not under his guidance.
When does complexity of movement suggest The guidance in this case is attributable only to the
action, and when does it suggest its absence? This of some mechanism with which he cannot
operation
depends, roughly speaking, upon whether the be identified.
movements in cohere in creating a Let us the term "intentional" for referring
question pattern employ
which us as
strikes meaningful. When they do, as in to instances of purposive
movement in which the
the case of the pianist, we find it difficult to imagine guidance
is
provided by the agent. We may say,
that the movements would have occurred, in just that action is intentional movement. The
then,
those complicated ways required by the meaningful notion of intentional movement must not be con?

pattern they have created, unless the pianist had fused with that of intentional action. The term
been guiding his hands and fingers as they moved. "intentional action" may be used, or rather mis-used,
In the case, on the other we find to that an action is necessarily a
epileptic's hand, simply convey
it unlikely that a person would have created such an movement whose course is under an
agent's guid?
incoherently complicated pattern if he had been ance. When it is used in this way, the term is pleo?
guiding his body through its movements. A person's nastic. In a more usage, it refers to
appropriate
movements, as Pears notes, actions which are undertaken more or less deliber?
simple generally suggest
neither an action nor a mere This is or
happening. ately self-consciously?that is, to actions which
because their do not strike us as the agent intends to In this sense, actions
patterns ordinarily perform.
being in themselves either or incoherent. are not intentional.
meaningful necessarily
They do not present us on their faces with any When a
person intends to
perform
an action,
indication of whether or not they are being guided what he intends is that certain intentional move?
the person as occur. ments of his should occur. When these move?
by they body
of movement action ments do occur, the is performing an inten?
Complexity body suggests person

only when it leads us to think that the body, during tional action. Itmight be said that he is then guiding
the course of its movement, is under the agent's
the movements of his body in a certain way (thus,
guidance. The performance of an action is accord? he is acting), and that in doing so he is guided by and
ingly a complex event, which is comprised by a fulfilling his intention to do just that (thus, he is
bodily movement and by whatever state of affairs acting intentionally). There appears to be nothing
or
activity constitutes the agent's guidance of it. in the notion of an intentional movement which
Given a
bodily movement which occurs under a implies that its occurrence must be intended by the
person's guidance, the person is performing
an agent, either by way of forethought or by way of self
action of what features of his prior causal conscious assent. If this is correct, then actions
regardless (i.e.,
history account for the fact that this is occurring. He intentional movements) may be performed either
is performing an action even if its occurrence is due or not.
intentionally
to chance. And he is not an action if the Since action is intentional movement, or be?
performing
movements are not under his as havior whose course is under the of an
guidance they guidance
proceed, even if he himself provided the antecedent agent, an
explication of the nature of action must
causes?in the form of beliefs, desires, intentions, deal with two distinct problems. One is to explain
decisions, volitions, or whatever?from which the the notion of guided behavior. The other is to
movement has resulted. specify when the guidance of behavior is attribut?
able to an agent and not as when a
simply, person's

Ill pupils dilate because the light fades, to some local


process going on within the agent's body. The first
When we act, our movements are This concerns the conditions under which
purposive. problem
is merely another way of that their course is behavior is while the second concerns the
saying purposive,
guided. Many instances of
purposive
movement are conditions under which purposive behavior is
not, of course, instances of action. The dilation of the intentional.

pupils of a person's eyes when the light fades, for The driver of an automobile guides the movement
example, is a
purposive movement; there are of his vehicle by acting :he turns the steering wheel,
mechanisms which guide its course. But the occur he depresses the accelerator, he the brakes,
applies
l60 AMERICAN PHILOSOPHICAL QUARTERLY

and so on. Our of our movements, while the course of a movement never have
guidance bodily may
we are does not that we occasion to do for no feedback of the
acting, similarly require so; negative
perform various actions. We are not at the controls sort that would trigger their compensatory activity
of our bodies in the way a driver is at the controls of may occur. The behavior is purposive not because it
his automobile. Otherwise action could not be con? results from causes of a certain but because it
kind,
ceived, upon of generating an infinite would be affected certain causes if the accom?
pain regress, by
as a matter of the occurrence of movements which
plishment of its course were to be jeopardised.
are under an The fact that our
agent's guidance.
movements when we are are is
acting purposive IV
not the effect of we do. It is a character?
something
istic of the operation at that time of the systems we Since the fact that certain causes
originate
an
are. action is distinct from the considerations in virtue of
Behavior is purposive when its course is subject which it is an action, there is no reason in principle
to adjustments which compensate for the effects of why
a person may not be caused in a
variety of
forces which would otherwise interfere with the course different ways to
perform the same action. This is
of the behavior, and when the occurrence of these
important in the analysis of freedom. It is widely
adjustments is not explainable by what explains the accepted that a person acts freely only if he could
state of affairs that elicits them. The behavior is in have acted otherwise. Apparent counterexamples
that case under the guidance of an independent to this
principle?"the principle of alternate possi?
causal mechanism, whose readiness to about bilities"?are cases that
bring provided, however, by
compensatory tends to ensure that the involve a certain kind of overdetermination. In these
adjustments
behavior is accomplished.3 The activity of such a cases a person
performs
an action entirely for his own
mechanism is normally not, of course, us. reasons, which inclines us to him as
guided by regard having
Rather it is, when we are
performing
an action, our
performed it freely; but he would otherwise have
guidance of our behavior. Our sense of our own been caused to perform it by forces alien to his will,
agency when we act is nothing more than the way so that he cannot avoid as he does.4
actually acting
it feels to us when we are somehow in touch with the Thus, suppose
a man takes heroin because he

operation of mechanisms of this kind, by which our enjoy its effects and considers them to be beneficial.
movements are
guided and their course
guaranteed. But suppose further that he is unknowingly addicted
Explaining purposive behavior in terms of causal to the drug, and hence that he will be driven to take
mechanisms is not tantamount to a it in any even if he is not led to do so own
propounding event, by his
causal of action. For one the pertinent beliefs and attitudes. Then it seems that he takes the
theory thing,
activity of these mechanisms is not to but con? that he could not have done otherwise
prior drug freely,
current with the movements they guide. But in any than to take it, and that the principle of alternate
case it is not essential to the purposiveness of a
possibilities is therefore false.
movement that it actually be causally affected by Donald Davidson argues to the contrary that
the mechanism under whose guidance the move? whereas a person does what he does
intentionally
ment A driver whose automobile is for his own reasons, he does not do
proceeds. intentionally
coasting downhill in virtue of gravitational forces what alien forces cause him to do. While the move?
alone may be entirely satisfied with its speed and ments of his body may be the same in both cases,
direction, and so he may never intervene to adjust Davidson maintains that the is not
person per?
its movement in any way. This would not show that an action when the movements occur
forming apart
the movement of the automobile did not occur under from pertinent attitudes and beliefs. Someone who
his guidance. What counts is that he was prepared has acted freely might have done the same thing
to intervene if necessary, and that he was in a even if he had not been moved on his own to do it,
position to do so more or less effectively. Similarly, but only in the sense that his body might have made
the causal mechanisms which stand to affect the same movements: "he would not have acted
ready

3
A useful discussion of this way of understanding behavior is provided
purposive by Ernest Nagel, "Goal-directed Processes in
Biology," The Journal of Philosophy, vol. 74 (1977), pp. 27 iff. The details of the mechanisms in virtue of which some item of behavior
is purposive can be discovered, of course, only by empirical But specifying the conditions which any such mechanism
investigation.
must meet is a philosophical problem, belonging to the analysis of the notion of purposive behavior.
4
Cf. my "Alternate Possibilities and Moral Responsibilities," The Journal of Philosophy, vol. 66 (1969), pp. 829-839; and "Freedom
of the Will and the Concept of a Person," The Journal of Philosophy, vol. 68 (1971), pp. 5-20.
THE PROBLEM OF ACTION l6l

intentionally had the attitudinal conditions been attitudinal conditions of a


person's action may
absent." Even in the "overdetermined" cases, then, themselves be alien to him. There is no reason to

something
rests with the agent: "not . . . what he assume that an addict who succumbs unwillingly to
does (when described in a way that leaves open his craving finally adopts as his own the desire he has
whether it was intentional), but whether he does it tried to resist. He may in the end merely submit to it
intentionally."5 with resignation, like a man who knows he is beaten
The issue here is not, as Davidson at one and who therefore the con?
suggests despairingly accepts
point, whether a
person's
action can be intentional sequences defeat must bring him, rather than like
when alien forces rather than his own attitudes someone who decides to join with or to incorporate
account for what he does. It is whether his behavior forces which he had formerly opposed. There are
can be intentional in those circumstances. Now the also obsessional and delusional "If
beliefs?e.g.,
behavior of the unknowing addict is plainly as I step on a crack it will break my mother's back"?
intentional when he is caused to take the drug by the which a person may know to be false but whose
compulsive force of his addiction, as it is when he influence he cannot escape. So even if it were true
takes it as a matter of free choice. His movements
(which it is not) that every action necessarily has
are not mere when he takes the attitudinal conditions its antecedent causes,
happenings, drug among
because he cannot help himself. He is then per? it might nonetheless be alien forces alone which
the very same action that he would have it about that a person an action.
forming bring performs
performed had he taken the drug freely and with the The assertion that someone has performed
an
illusion that he might have done otherwise. action entails that his movements occurred under
This example is not designed to show that his guidance, but not that he was able to
keep
him?
Davidson is mistaken in insisting that there can be self from guiding his movements as he did. There are
no action without or in the absence occasions when we act or of
intentionality, against independently
of pertinent attitudinal conditions. Even when the our wills. On other occasions, the guiding principle
addict is driven to do what he does, after all, his of our movements is one to which we are not merely
behavior is presumably affected both by his craving resigned; rather, we have embraced it as our own.
for the drug and by his belief that the procedure he In such cases, we will ordinarily have a reason for
follows in taking it will bring him relief. His move? embracing it.
Perhaps,
as certain philosophers
ments, as he sticks the into his arm and would our a reason for
syringe claim, having acting may
the are intentional. sometimes cause it to be the case that movements of
pushes plunger, certainly
However, the relevant is not whether an our bodies are us in a manner which
problem guided by
action can occur from attitudinal conditions. reflects that reason. It is indisputable that a person's
apart
It is whether it is possible that an action should be beliefs and attitudes often have an
important
caused by alien forces alone.
bearing upon how what he is doing is to be inter?
This will seem to be impossible only if it is thought preted and understood ; and it may be that they also
that an action must have attitudinal conditions at times in the causal of his
figure explanations
among its causes. But it is not essential to an action actions. The facts that we are rational and self
that it have an antecedent causal of any conscious affect the character of our
history substantially
kind. Even if there can be no action in the behavior and the ways in which our actions are
particular
absence of certain attitudinal conditions, therefore, into our lives.
integrated
it is not as causes that these conditions are
prior
V
essential. The bears upon the that is
example point
at issue, how an action The to our actions of states and events
actually by illustrating significance
(including, of course, any requisite attitudinal which depend upon the exercise of our higher
may have no causes other than non should not lead us, to exagger?
constituents) capacities however,
attitudinal or alien ones. Thus it confirms the ate the peculiarity of what human beings do. We are
falsity of the principle of alternate possibilities, by far from being unique either in the purposiveness of
showing
a person
that may be caused by alien forces our behavior or in its intentionality. There is a
alone to an action which he might also to discuss the nature
perform tendency among philosophers
on his own. of action as character?
perform though agency presupposes
The example also suggests, by the way, that the istics which cannot plausibly be attributed to

5
Op. Cit., pp. 149-150.
l62 AMERICAN PHILOSOPHICAL QUARTERLY

members of species
other than our own. But in fact Each contrasts instances in which purposive be?
the contrast between actions and mere havior is attributable to a creature as agent and
happenings
can readily be discerned elsewhere than in the lives instances in which this is not the case.

of people. There are numerous agents besides our? This generic contrast cannot be explicated in
selves, who may be active as well as
passive with terms of any of the distinctive higher faculities which
to the movements of their bodies. come into when a acts.
respect characteristically play person
Consider the difference between what goes on The conditions for attributing the guidance of
when a spider moves its legs inmaking itsway along bodily movements to a whole creature, rather than
the ground, and what goes on when its legs move in only
to some local mechanism within a creature,
similar patterns and with similar effect because they evidently obtain outside of human life. Hence
they
are manipulated by a boy who has managed to tie cannot be satisfactorily understood by relying upon
to them. In the first case the movements are concepts which are inapplicable to spiders and
strings
not simply purposive, as the spider's digestive their ilk. This does not mean that it must be
doubtless are. are also attributable for an of human agency to
processes They illegitimate analysis
to the
spider,
who makes them. In the second case invoke concepts of more limited scope. While the
the same movements occur but are not made conditions of agency are unclear, it may well
they general
by the spider, to whom they merely happen. be that the satisfaction of these conditions by human
This contrast between two sorts of events in the the occurrence of events or
beings depends upon
lives of which can be observed in the states which do not occur in the histories of other
spiders,
histories of creatures even more creatures. But we must be careful that the ways in
benighted, parallels
the more familiar contrast between the sort of event which we construe agency and define its nature do
that occurs when a raises his arm and the sort not conceal a bias, which causes us to
person parochial
that occurs when his arm without his the extent to which the concept of human
goes up neglect
it. Indeed, the two contrasts are the same. action is no more than a case of another
raising special
The differences they distinguish are whose is much wider.
respectively concept range
and have, as it were, the same
alike; they point.

Yale University Received August iy, igyy

You might also like