You are on page 1of 13

Journal of Philosophy, Inc.

Chisholm's Theory of Agency


Author(s): Alan Donagan
Source: The Journal of Philosophy, Vol. 74, No. 11, Seventy-Fourth Annual Meeting American
Philosophical Association, Eastern Division (Nov., 1977), pp. 692-703
Published by: Journal of Philosophy, Inc.
Stable URL: http://www.jstor.org/stable/2025773 .
Accessed: 22/12/2013 07:47
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
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.

Journal of Philosophy, Inc. is collaborating with JSTOR to digitize, preserve and extend access to The Journal
of Philosophy.

http://www.jstor.org

This content downloaded from 130.251.200.3 on Sun, 22 Dec 2013 07:47:23 AM


All use subject to JSTOR Terms and Conditions

692

THE JOURNAL OF PHILOSOPHY

CHISHOLM'S THEORY OF AGENCY


J

N his CarusLectures,Person and Objectt R. M. Chisholm has

expounded his perfected theory of agency so lucidly that to


recapitulate what he has written would not be calculated to
enlighten. I shall therefore discuss a few topics on which my own
opinions have largely been derived from his, but in his treatment
of which I find difficulties.
I begin with his doctrine that what it is for an agent to cause an
event can be understood only in terms of a familiar concept of
event-causation, taken together with the concept of undertaking or
endeavoring, which he explicates in terms of his revival of certain
elements in Aquinas's theory of will. In so connecting agent-causation with will, Chisholm follows a line of thought expressed by
Thomas Reid when he wrote that "Power to produce any effect
implies power not to produce it. We can conceive no way in which
power may be determined to one of these rather than another, in
a being that has no will." I
The fundamental causal concept in Chisholm's theory is that of
causally contributing to. When causation is of one event p by another q, 'p causally contributes to q' can be informally elucidated
as meaning 'p is a member of a set of nonredundant conditions
jointly sufficient, according to the laws of nature, for q'. Although
Chisholm declines to offer this as a definition, on the ground that
no solution to the problem of defining redundancy is at hand (205
fn 9; cf. 60/1), what it is for one event causally to contribute to
another is, I think, clear enough for us to be able to go on. But
what is it for an agent-a substance or continuant, and not an
event-causally to contribute to the occurrence of an event? Plainly,
no analysis of the sort offered in the case of event-causation will do;
for event-causes are parts of sets of jointly sufficient conditions,
and, since Chisholm holds, with Reid, that an agent as agent can
* To be presented in an APA symposium on Chisholm's Person and Object,
December 28, 1977. Anthony Quinton will be co-symposiast, and Roderick
Chisholm will comment. Neither Quinton's paper nor Chisholm's comment is
available at this time.
My research has been supported by the John Simon Guggenheim Memorial
Foundation, by a grant from the National Science Foundation to the Center
for Advanced Study in the Behavioral Sciences, Stanford, and by the University
of Chicago; to all I offer my gratitude.
t La Salle, Ill.: Open Court, 1976. All otherwise unidentified page references
in the text are to this work.
1 Essays on the Active Powers of the Human Mind, Baruch A. Brody, ed.
(Cambridge, Mass.: MIT Press, 1969), p. 35.

This content downloaded from 130.251.200.3 on Sun, 22 Dec 2013 07:47:23 AM


All use subject to JSTOR Terms and Conditions

CHISHOLM S THEORY OF AGENCY

693

causally contribute to an event only if it is in his power not to


contribute to it, agent-causes cannot be parts of sets of jointly sufficient conditions.
Chisholm's solution divides into three parts: (1) that every human action is an agent's causally contributing to the occurrence of
an event which I shall describe as "internal" to the action and
other than it; (2) that an agent's causally contributing to that event
is his causally contributing to a further event of a special kind,
namely, an undertaking or endeavor, which causally contributes to
the internal event in the familiar event-causal way; and (3) that
agents causally contribute directly to the occurrence of their undertakings, as also to the occurrence of their deliberate omissions to
undertake. It follows that causally contributing directly to an
undertaking, or to an omission to undertake, is not a full human
action, although every full human action has as an element the
making of such a direct causal contribution. It also follows that
agent-causation, in its pure form, is exhibited in making causal
contributions, not to actions as such, but to undertakings to act.
Let us look more closely at each of the three points in this analysis.
The first is that every human action is a complex event, consisting in the making of a causal contribution to an event internal to
it. What Chisholm has in mind is readily seen by examining representative sentences which in his view designate actions; for example, 'S runs at t', 'S raises his arm at t'. Each is to be considered
identical in sense with a sentence in which S is said causally to
contribute to an event of whatever kind is internal to the action
designated by the original sentence. In general, for every sentence
'S Fs at t', in which an upper case 'F' stands for a kind of action,
a sentence of the form 'S causally contributes to an f-ing at t', in
which a lower case 'f' stands for the kind of event that is internal
to the action of F-ing, will have the same sense. In the case of S's
raising his arm, it is easy to say what the internal event is, namely,
that S's arm goes up. In the case of S's running, the description
would be much more complicated. But with sufficient ingenuity it
would be possible to describe, for any kind of action recognized by
human beings, the kind of internal event in the making of a causal
contribution to which an action of that kind consists.
The second point in Chisholm's analysis is that an agent causally
contributes to the occurrence of an event internal to an action by
causally -contributing to an undertaking or endeavoring. Undertaking or endeavoring is an undefined propositional attitude (cf. 22).
In elucidating -what that propositional attitude is, Chisholm re-

This content downloaded from 130.251.200.3 on Sun, 22 Dec 2013 07:47:23 AM


All use subject to JSTOR Terms and Conditions

694

THE JOURNAL OF PHILOSOPHY

marked that the "least misleading" way of expressing it might be


by the verb phrase 'acts with the intention of' (74). However, it
appears that 'acts with the intention of' should itself be understood
as 'intends here and now to': that is, as standing for an intending
and not, as Chisholm's preferred locution suggests, for a doing.
For, on the very next page, Chisholm describes a point about acting with the intention of bringing something about as "a point
about intending" (75); and, later still, he points out that undertakings are not acts (85).
There is, however, a puzzling passage in which Chisholm describes an undertaking as the object of what Aquinas called an
elicited act of will (actus voluntatis elicitus). A voluntary action
itself was called by Aquinas a commanded act of will (actus voluntatis imperatus), and the willing of whatever is directly willed in
performing such an action-of what I have called the event internal to it-was what Aquinas meant by an 'elicited act' (66/7).
Hence, if an undertaking is the object of an elicited act, it must be,
not a propositional attitude, but rather the proposition that is the
object of that attitude. With trepidation, I shall therefore treat as
a slip the identification of an undertaking with the object of an
elicited act, and shall take Chisholm to identify undertakings with
elicited acts themselves. Here, however, a terminological confusion
must be avoided. An elicited act, as Aquinas understands it, is not
a full human action, but only an element in one. When Chisholm
denies that undertakings are acts, he means that they are not full
human actions, not that they are not acts in the wider Thomistic
usage according to which any actualization of a potentiality is an act.
The third point in Chisholm's analysis is that agents causally
contribute directly to the occurrence of their undertakings. However, he says little about the concept of causally contributing to
the occurrence of p directly, which he defines as (i) contributing
causally to p at t, and (ii) there being no q such that (a) one contributes causally to q, and (b) q contributes causally to p (85;
D. II. 20).
Many philosophers profess to find unintelligible the concept of
an individual thing's causally contributing to an event, as distinct
from an event's doing so; either on the question-begging ground
that causation is nomological, or a matter of sufficient conditionship, and hence that no form of causation that is not reducible to
event-causation is conceivable; or, on Donald Davidson's ground,
that, since the concept of agent-causation is indistinguishable from

This content downloaded from 130.251.200.3 on Sun, 22 Dec 2013 07:47:23 AM


All use subject to JSTOR Terms and Conditions

CHISHOLM S THEORY OF AGENCY

695

that of agency, the word 'causation' in the expression 'agent-causation' is vacuous.2


These objections uncannily reverse the objections Reid made in
the eighteenth century against event-causation. Reid pointed out
that the necessity possessed by the laws of Newtonian physics was
so mysterious that Newton himself indignantly denied that he
thought it causal, whereas for centuries the model of causality had
been that of the relation of an agent to his actions, and, above all,
that of the divine agent to his creative actions. As J. L. Austin once
remarked, "'Causing', I suppose, was a notion taken from a man's
own experience of doing simple actions, and by primitive man
every event was construed in terms of this model: every event has
a cause, that is, every event is an action done by somebody-if not
by a man, then by a quasi-man, a spirit." 8 In terms of this, the
original notion of cause, Reid complained that the post-Newtonian
concept of event-causation was improper and obscure. Confronted
with the Davidsonian objection that the concept of agent-causation
is vacuous, I think he would have demurred: agreeing that indeed
the word 'causation' here stands for nothing over and above the
relation of an agent to his actions, but contending that that relation is what the word 'cause' properly stands for, and that we
understandwhat it iS.4
The point of describing the connection between an agent and
his undertakings or elicited acts of will as causal is that whether or
not a given undertaking is actualized or elicited is wholly up to the
agent: ascribing it to him implies that it was in his power not to
have actualized it. This will dissatisfy us if our minds are filled
with the nomological conception of causation. Perturbed by the
objection that the existence of an agent cannot be the cause of his
acts of will, since the longer he lives, the longer his existence has
failed to cause some of them, Baruch Brody, Reid's most recent
editor, was moved to pronounce the answer to the question, What
causes the act of willing to take place when it does? to be the "missing link" in Reid's theory.5 Unable to get out of his head the
notion that a cause must be a sufficient condition or a nonredundant part of one, Brody failed to see that Reid held that the agent
himself-an individual thing-is the cause of his acts of will, and
2 "Agency,"in R. Binkley and others, Agent, Action and Reason (Oxford:
Blackwell, 1971),p. 15.
8 "A Plea for Excuses," in Philosophical Papers (Oxford: Clarendon Press,
2nd ed. 1970),p. 202.
4 Reid, op. cit., pp. 16-21, 36-39, 41-47.
5 Baruch A. Brody, in Reid, op. cit., p. xix.

This content downloaded from 130.251.200.3 on Sun, 22 Dec 2013 07:47:23 AM


All use subject to JSTOR Terms and Conditions

696

THE JOURNAL OF PHILOSOPHY

not the existence of the agent-an event. What caused Julius


Caesar's act of willing that his army cross the Rubicon to take
place when it did? Reid's and Chisholm's answer is: Julius Caesar.
To dismiss this answer on the ground that it furnishes no sufficient
condition for that act of will to have occurred when it did presupposes what Reid and Chisholm deny: that Caesar could not
have chosen not to cross the Rubicon.
Undertakings, like everything to which an agent directly causally
contributes, are internal changes. They may, more scholastico, be
said to be immanently caused because, as states of the agent himself, they "remain within the agent" (85). But it is not within the
province of action theory to rule on H. A. Prichard's doctrine that
whatever is directly caused in action, if anything, must be some
change in the agent's brain.6 What is within the province of action
theory is that "whether or not our undertakings are themselves
changes in our brains, it would seem that they contribute to certain
changes in our brains that we know next to nothing about" (86).
So far, I have deliberately abstained from attempting to identify
the concept of event with which Chisholm is working in the analysis of human action I have set out; for I am persuaded that there
is a concept of event in terms of which all three points in that
analysis are sound. However, I suspect that, in terms of the concept
of event with which Chisholm himself works, his analysis gives rise to
great and avoidable difficulties, two of which I shall try to explain.
As Chisholm conceives it, an event is a state of affairs of a certain
kind, namely, one that occurs at a place and time, that is not a
proposition, and that entails the exemplification of nonessential
properties that may not be rooted outside the present (126-128).
All these conditions are expressed in defined technical terms, the
definitions of which it would be superfluous in the present inquiry
to rehearse: they should convey something even to those unfamiliar
with those definitions. The genus state of affairs, in terms of which
the species event is defined, is itself defined as follows:
D.IV.1 p is a state of affairs=
who acceptsp (117).

Df

It is possible that there is someone

To this definition Chisholm supplies this gloss: "Frege used the


term 'thought' (Gedanke) as we are using 'state of affairs' and observed: 'The being of a thought may also be taken to lie in the
possibility of thinkers' grasping the thought as one and the same
thought.' (For the present, we will interpret many of the observa8

Moral Obligation (Oxford: Clarendon Press, 1949), p. 193.

This content downloaded from 130.251.200.3 on Sun, 22 Dec 2013 07:47:23 AM


All use subject to JSTOR Terms and Conditions

CHISHOLM'S THEORY OF AGENCY

697

tions philosophers have made about propositions as being applicable generally to states of affairs. Subsequently, we will attempt to
single out propositions as constituting a subspecies of states of
affairs)" (1 17/8).
The criterion of identity for a Fregean Gedanke is less than
clear. Chisholm, however, offers a criterion of identity for a state
of affairs in terms of the following very strict concept of entailment:
D.IV.2 p entails q = Df p is necessarilysuch that (a) if it obtains
then q obtains and (b) whoever accepts it accepts q (118).
That criterion is: "if a state of affairs p is identical with a state of
affairs q, then p entails q and q entails p" (118). The definition of
entailment on which this criterion depends makes use of a notion
that Frege would presumably have rejected: that there can be entailments between things that are neither true nor false. Yet Chisholm's states of affairs are plainly very like propositions.
According to the criterion given for the identity of states of affairs, Grant accepting Lee's surrender is a different state of affairs
from the accepting of Lee's surrender by the general then commanding the armies of the Republic (for it is not necessarily the
case that Grant was commanding the armies of the Republic when
Lee surrendered to him). Again, that 2 + 3 = 5 is a different state
of affairs from that the sum of the smallest even number and 3 is 5
(for somebody might accept that 2 + 3 = 5 and yet not know what
an even number is). The proposition-like character of states of
affairs, and their extremely strict criterion of identity generate the
first difficulty in Chisholm's analysis of action to which I wish to
draw attention: that it objectionably multiplies actions, and particularly unintentional actions.
Although this difficulty is most acute with respect to what is
unintentionally done in acting successfully, it arises even with respect to what is unintentionally done in acting unsuccessfully-the
variety of unintentional actions Chisholm most often discusses.
Consider a simple example: on a walk, S sees a puddle in his path,
lengthens his stride in order to step over it, but, misjudging how
much his stride must be lengthened, steps into it; and, his shoe not
being waterproof, consequently gets his foot wet. Here S obviously
acts unintentionally as a by-product of acting unsuccessfully. The
primary component in unsuccessful action, as Chisholm conceives
it, is unsuccessful endeavor (or undertaking or actus voluntatis
elicitus); and unsuccessful endeavor is to be understood in terms of
successful, which is defined as follows: "a man is successful in his

This content downloaded from 130.251.200.3 on Sun, 22 Dec 2013 07:47:23 AM


All use subject to JSTOR Terms and Conditions

698

THE JOURNAL OF PHILOSOPHY

endeavour to bring about a state of affairs p provided that (i) he


does endeavour to bring about p, and (ii) his endeavour to bring
about p contributes causally to the occurrence of p" (83). Unsuccessful endeavor is that which does not causally contribute to what
is endeavored, and unsuccessful action is action in which a state of
affairs to which an endeavor causally contributes is incompatible
with what is endeavored.
Since, according to Chisholm's analysis, an action is a state of
affairs, at least three distinct actions can be distinguished as byproducts of the action (1): S lengthening his stride (in order to step
over the puddle). They are: (2) S lengthening his stride insufficiently to clear the puddle; (3) S stepping into the puddle; and
(4) S getting his foot wet. Here, since (1) does not entail (2), although (2) entails (1), (1) and (2) are distinct states of affairs. And
since in none of the pairs formed by the set (2), (3), and (4) does
either member entail the other, (2), (3), and (4) are each distinct
from the others.
Does this unacceptably multiply S's actions? There is a persuasive argument for thinking that it does: namely, that having
lengthened his stride as he did-that is, insufficiently to clear the
puddle-S had no more to do in order to perform the unintentional actions (2), (3), and (4).7 First, any lengthening of a stride
has a magnitude: there is no such thing as lengthening it, and also
lengthening it by a definite amount. That such a lengthening may
be described either with or without specifying its magnitude does
not convert it into two lengthenings. Secondly, any man who
lengthens his stride insufficiently to clear a puddle has no more to
do in order to step into it: that he steps into it is simply a matter
of the length of his step and of something in no way a matter of
what he does here and now: namely, of the width of the puddle.
And finally, given that his shoe is not waterproof, a man who steps
into a puddle has no more to do in order to get his foot wet; and
that his shoe is not waterproof is not a matter of anything he does
here and now. What we appear to have, in short, are not descriptions of four different things that S does, but four different descrip7 The line of thought in this and the following paragraph is taken from
Davidson, op. cit.; but its expression is also indebted to Irving Thalberg,
"Singling Out Actions, Their Properties and Components,"this JOURNAL, LXVI,
21 (Nov. 4, 1971): 781-787, p. 786. It is not accidental that (2), (3), and (4)
illustrate, respectively, the augmentation generation, simple generation, and
causal generation of Alvin I. Goldman's A Theory of Human Action (Englewood Cliffs, N.J.: Prentice-Hall, 1970), pp. 22-30.

This content downloaded from 130.251.200.3 on Sun, 22 Dec 2013 07:47:23 AM


All use subject to JSTOR Terms and Conditions

CHISHOLM S THEORY OF AGENCY

699

tions of one thing S does, namely, his lengthening of one of his


steps while walking.
In discussing what is unintentionally done in acting successfully,
the example Chisholm considers is the familiar classical one:
Oedipus intended to kill an insolent stranger on the road to
Phocis, and succeeded; but since, unknown to him, the insolent
stranger was his father, he unintentionally and disastrously killed
his father. Here it is unmistakably evident that, having killed the
insolent stranger, Oedipus had nothing more to do in order to kill
his father: doing one, he did the other.
Yet if Oedipus's killing the insolent stranger and Oedipus's killing his father were one and the same action, would it not follow
that one and the same action had the contradictory properties of
being both intentional and unintentional? By recognizing this as a
problem, Chishoim allows prima facie force to the notion that only
one action of Oedipus is in question. But, instead of dismissing
that notion as merely prima facie, as his analysis demands, and
getting rid of the apparent contradiction by insisting that really
two actions are involved, he solves the problem as follows:
'Oedipus'sact was both intentional and nonintentional' is elliptical
for 'Oedipus'sact was intentional with respect to one thing and nonintentional with respect to another thing'. What Oedipus did was
intentional with respect to the killing of the unfriendly traveller,and
nonintentionalwith respect to killing his father (87).
Here, if I understand him, Chisholm abandons his analysis of actions as states of affairs, concedes that only one action of Oedipus's
is in question, and resolves the apparent contradiction in the same
way as his adversaries: by treating the predicate 'is (un)intentional'
as covertly propositional.
This solution is not, as has sometimes been alleged, an ad hoc
device for avoiding difficulties in action theory. In Frege's theory
of reference (the only one we have that is at all well worked out)
a cardinal doctrine is that predicates applying to things only as
they are spoken or thought about, and not as they are in themselves, constitute sentential contexts that are oblique; and that, as
a result, expressions occurring in those contexts have oblique and
not direct reference-that is, refer to their usual senses, and not
to their usual referents. If intending is a propositional attitude, as
Chisholm himself holds, it is a natural inference that the predicate
'is (un)intentional' should constitute such an oblique context. And
in that case the sense of the sentence 'Oedipus's killing of his father

This content downloaded from 130.251.200.3 on Sun, 22 Dec 2013 07:47:23 AM


All use subject to JSTOR Terms and Conditions

700

THE JOURNAL OF PHILOSOPHY

was unintentional' (grammar forbids Chisholm's gerundive locution 'Oedipus killing his father was unintentional') is the same as
the overtly propositional 'That Oedipus killed his father was unintentional'. That sense is not that a certain action, no matter how
it may be truly described, has a certain property, but that in doing
that action, its agent did not have a certain propositional attitude
toward it. Hence both the property of being intentional and that
of being unintentional are true of actions only as they are described by descriptions of them that are implicit in true statements
about propositional attitudes taken toward them by their agents.
And so it may well be that the same action is intentional "under"
one description (or with respect to it), and unintentional "under"
another.
A second difficulty that arises in Chisholm's theory of action because of his doctrine that events are states of affairs has to do with
causation. Chisholm accepts the familiar view that every event has
a cause. It follows that, if events are states of affairs, then every
state of affairs satisfying the conditions for eventhood must have
a cause. And, despite the formidable appearance of Chisholm's
definition of an event, those conditions are liberal:
D.IV.7 p is an event = Df p is a state of affairswhich is such that:
(i) it occurs;(ii) it is not a proposition; and (iii) it entails a property
G which is such that (a) only individual things can exemplify G,
(b) that it is possible that no individual things exemplify G, and
(c) G is such that it may not be rooted outside the times at which
it is had (128).
It follows from this definition that, if p and q are events, and if it
is true that p causally contributes to q, then p causally contributing to q is also an event. Ex hypothesis p causally contributing to
q is not only a state of affairs, but is one that occurs (118/9). Is it
a proposition? Undoubtedly, it has a counterpart that is a proposition, namely that p causally contributes to q. But so do all events
(cf. 124). And p causally contributing to q, unlike the undoubted
proposition that is its counterpart, does seem to be the sort of thing
that Chisholm would say can occur at some times and not at others.
If so, it is not a proposition (123). Finally, since ex hypothesi p and
q are events, they each entail some property G that satisfies conditions (iiia), (iiib), and (iiic) in Chisholm's definition of an event.
And, since any complex state of affairs containing other states of
affairs as components must entail whatever properties they entail,
it follows that p causally contributing to q also entails some prop-

This content downloaded from 130.251.200.3 on Sun, 22 Dec 2013 07:47:23 AM


All use subject to JSTOR Terms and Conditions

CHISHOLM S THEORY OF AGENCY

701

erty G that satisfies conditions (iiia), (iiib), and (iiic). But in that
case, it satisfies all the conditions laid down for eventhood.
Chisholm himself accepts one proposition which implies that
causal contributings are events: namely, that some of them have
causes; for his analysis of purposely undertaking p in order to
bring about q entails that such undertakings involve undertakings
causally to contribute to the undertaking p contributing causally
to q (76). If causal contributings are events, then indeed they must
have causes, on a supposition which Chisholm appears not to question, that every event has a cause. But in that case, every ordinary
event of the form p causally contributing to q is the first member
of an infinite series that continues: for some r, r causally contributing to p causally contributing to q; for some r', r' causally contributing to r causally contributing to p causally contributing to q;
and so on. Quite apart from the objection that generating such a
series is vicious (for it is a series of grounds, not of consequents),
there is a direct objection from the theory of causation: that if p
causally contributes to q, then the set of conditions, of which p is
a nonredundant member, that are jointly sufficient for q are so by
virtue of the nature of things, not by virtue of some further event.
Chisholm's doctrine that causal contributings are events, which
itself derives from his analysis of events as a subspecies of states
of affairs, impairs his theory of action chiefly in his analysis of
purposive activity. His fundamental definition of undertaking with
a purpose is as follows:
D.II.13 S undertakesp and does so for the purpose of bringing about
q = Df S undertakesto bring about (i) p and (ii) his-undertaking-pcontributingcausallyto q.
Chisholm expounds some of his reasons for this definition historically. In an earlier paper he had proposed a simpler definition
with the following definiens:
S undertakesto bring about that state of affairswhich is: (i) p and
(ii) p contributingcausallyto q (77; cf. 207/8).
There are, however, two objections to it.
The first is that it does not allow for cases of a kind to which
Descartes drew attention: that, in order to bring about an event
that is the cause of another event, it may be necessary for an agent
to direct his attention to the effect, and not to the cause. For example, if he wants to bring about certain cerebral processes, which
in fact are those that cause his arm to go up when he raises it in

This content downloaded from 130.251.200.3 on Sun, 22 Dec 2013 07:47:23 AM


All use subject to JSTOR Terms and Conditions

702

THE JOURNAL OF PHILOSOPHY

the ordinary way, he had better undertake that his arm go up, and
not that those processes occur (77, 86/7). The second objection,
pointed out by Annette Baier,8 is that one's purpose in undertaking something may not be to bring about anything either caused
by it or causing it, but rather something constituted by it. A conductor's purpose in undertaking certain movements of his arms and
body that would be extraordinary except in front of an orchestra
is that they constitute conducting an orchestra, not that they cause
it or are caused by it (77/8).
Chisholm's more complex definition meets both of these difficulties. For in undertaking his arm's going up in order to bring about
certain cerebral processes, an agent's undertaking does causally contribute to the occurrence of those processes, even though his arm's
going up does not. And in undertaking certain movements of his
arms and body for the purpose of bringing about that an orchestra
is conducted, a conductor's undertaking does causally contribute to
an orchestra's being conducted, even though the movements of his
arms and body do not. These successes have, I suspect, distracted
Chisholm's attention from the deeper objection already mentioned:
that whether or not a set of nonredundant conditions, of which p
is one, are jointly sufficient for q, depends on the nature of things;
it is not something to which any other event whatever can causally contribute.
However, it does not follow that there are no circumstances in
which an event p can causally contribute both to q, and to p's
causally contributing to q; for it can make a direct causal contribution to q, and also to some other of the conditions that, together
with itself, are jointly sufficient for q. Heating certain morsels of
dough in an oven causally contributes to certain buns being hot;
but it also causally contributes to its causally contributing to those
buns being hot by cooking those morsels into buns. If it had not
done that, there would have been no buns for it to make hot. In
his definition of taking a preliminary step, Chisholm makes use of
this legitimate sense in which p may causally contribute, not only
to q, but also to its own making of a causal contribution to q
(cf. 81).
Unfortunately, his definition of purposive activity is not confined
to cases in which S's undertaking something in order to bring
about q causally contributes not only to q, but also to some other
member of the set of nonredundant conditions jointly sufficient
8 "Act and

Intent," this JOURNAL, LXVII, 19 (Oct. 8, 1970): 648-658, pp. 648/9.

This content downloaded from 130.251.200.3 on Sun, 22 Dec 2013 07:47:23 AM


All use subject to JSTOR Terms and Conditions

703

CHISHOLM S THEORY OF AGENCY

for q. But, except in those cases, it is objectionable for the reason


given. Here a simple example is best, although Chisholm's own
problem cases of purposive undertaking would also serve. When
S undertook to pull a trigger in order to shoot T, his undertaking
causally contributed, by way of that trigger's being pulled, to T's
being shot, but not to that undertaking's causally contributing to
T's being shot. If the gun whose trigger was pulled was in working
order, loaded, and properly aimed (matters to which S's undertaking to pull the trigger did not causally contribute), then it was a
matter of nature that S's undertaking, by way of the trigger's being
pulled, causally contributed to T's being shot. But in making that
causal contribution, it did not causally contribute to its making of
it. That was up to nature.
If Chisholm's definition of purposive undertaking is objectionable, what can be put in its place? The objections made suggest
something along the following lines:
S undertakesp and does so for the purpose of bringing about q= Df
S undertakescausallyto contribute to q, and S believes that his causally contributingto p would causallycontribute to q, and because of
that undertakingand that belief, S undertakescausally to contribute
to p.
There are difficulties in a definition along these lines; but I have
no time to explore them, or to correct the one given in the light
of them.
I have argued that the three points identified at the beginning
of this paper as cardinal to Chisholm's theory of agency are made
harder to accept if they are interpreted in accordance with Chis-.
holm's ontology of states of affairs. Even if the arguments I have
presented have force (and they may well be invalidated by a variety
of misunderstandings) Chisholm's theory interpreted according to
Chisholm's ontology does deal with a diverse array of difficult philosophical questions (cf. 72-78, 115-117); and I have offered no
more than hints of alternative ways of dealing with those questions.
ALAN DONAGAN

University of Chicago

This content downloaded from 130.251.200.3 on Sun, 22 Dec 2013 07:47:23 AM


All use subject to JSTOR Terms and Conditions

You might also like