Free Will and Agential Powers
Randolph Clarke
What is free will, or, as it might be called, freedom of the will? Writers with quite different
views on the question often agree that what’s at issue is a power (or powers) of some kind.1
Since will is plausibly seen as itself a power, we would do well to think of freedom of the will
not as a power of the will but rather, as Locke insisted, of agents. The thought fits comfortably
with Hobart’s remark that “in accordance with the genius of language, free will means freedom
of persons in willing, just as ‘free trade’ means freedom of persons (in a certain respect) in
trading” (1934: 8).
Locke’s own treatment of what he calls liberty or freedom illustrates one difficulty in
understanding just what the power in question comes to. His initial offering is a characterization
of freedom with respect to doing what one wills (or might will), not freedom in willing. I am at
liberty to A, Locke says, just in case I have a power to A or not A, according to which of these I
will to do. Plainly, being at liberty to will to A can’t be understood in the same fashion. And
Locke at one point scoffs at the absurdity of asking whether an agent might have the latter
freedom. But the fact that freedom to will can’t be understood in the same manner as is freedom
to do what depends on willing doesn’t imply that there’s no sense at all to be made of it.
It has of course proved difficult to say exactly what free will comes to. I’m not able to
provide such an account; my aim here is more modest. What I hope to do is develop a framework
for thinking about the problem.
I’ll suggest a way of understanding what willing is, one that finds no need to appeal to
volition as a special kind of mental act. I’ll suggest a constraint concerning what an agent’s
freedom with respect to willing might come to. I’ll offer some observations on the conception of
powers that might be employed in an account of free will. And I’ll examine a recent proposal
that construes free will as consisting in dispositions that are fundamentally the same kind of
thing as those, such as fragility or solubility, of inanimate objects. The account runs into
problems that, I’ll argue, characteristically beset efforts of this sort. I’ll offer some suggestions
about how these problems might be dealt with, and I’ll conclude with brief consideration of some
alternative approaches.
1. Willing
If free will is the freedom of an agent in willing, it behooves us to have some idea of what
willing is. I take it that in order for an agent to be free in the relevant respect in doing something
A, it must be the case that A-ing is something that can be done freely. Willing, then, must be the
kind of thing that one can do freely.
If we can do anything freely, we can freely perform intentional actions. I’ll take it that the
freedom of an agent in willing is, at least in the paradigm case, the agent’s freedom in
performing some intentional action. Paradigmatic instances of willing, then, are instances of
performing some intentional action; paradigmatic willings are intentional actions. (I’ll allow, of
course, that not every intentional action is a willing.)
For this reason, it isn’t appropriate, in the present context, to count intending to do a
certain thing as an instance of willing. Intending is a state, not an action. And one can intend to A
without so much as beginning to do what one intends. I might intend today to drive to my office
tomorrow but forget to do so, or change my mind, or not survive the night.
Nor will it do to take willing to be coming to have an intention. True enough, one can
come to have an intention to A by deciding (choosing) to A. And I take it that a decision to A is
itself an intentional action. (I’ll say more about this later.) When one decides to A, one
intentionally forms an intention to A. One can perform this action long before, or without ever,
A-ing. But one can come to have an intention in a way other than by deciding. Deciding settles
practical uncertainty, uncertainty about what to do. But often it’s perfectly clear what to do;
there’s no uncertainty about it. When that’s so, one can come to intend to do a certain thing, and
do it intentionally, without deciding to do it.
When I arrive at my office in the morning, I reach into my pocket, take out my keys, and
unlock the door. I do all of this intentionally, and when I do it I intend to do it. The intention to
unlock the door right away is something I come to have on my arrival. But I don’t make a
decision to unlock the door; when I arrive, I have no uncertainty about what to do that needs to
be settled by making a decision.
An intention can arise from prior rationalizing states in much the same way in which a
belief or desire can. When it does, one’s coming to have that intention isn’t a free action; it isn’t
an action at all. This point applies to present-directed intentions–intentions to do a certain thing
right away–as well as future-directed intentions–intentions to do something at subsequent time.
Still, I think we can fairly count deciding as a kind of willing. We can so count both
deciding to do something right away and deciding to do something at some subsequent time. A
decision is in either case an intentional action, and it’s an action that one performs even if one
never manages to do, or even begins to do, the thing one decides on.
If I decide today to drive to my office tomorrow, I will today to make the drive
tomorrow. I settle on doing that; I actively set my will on doing it. Such an act of willing is
something that can be free, if anything can be.
Another kind of thing we might consider is trying to do a certain thing. I have in mind a
notion of trying on which trying can be effortless. Trying is attempting, and some attempts are
easy. When you intentionally raise your arm, even in normal circumstances, you’ve made a
successful attempt to raise your arm. And a successful attempt is an attempt, or an instance of
trying. You’ve tried to raise your arm.
Attempts are intentional; if you try to A, you intentionally try to A. Intentionally trying to
A, like intentionally doing anything, might require having some intention with relevant content.
But it doesn’t require having an intention to try to A. When you try to raise your arm, you might
intend simply to raise your arm.
As I understand trying, when you try to A and succeed in A-ing, your trying to A is your
A-ing. It is your successful attempt to A.2
On this view, trying isn’t a distinct action type on a par with walking and speaking. It’s
not that each act of walking, speaking, etc. is preceded by or begins with an act of a different
type, a trying. Trying to A is going about or being engaged in the business of A-ing.3 Trying is
thus different from what some writers call volition.4
One is trying to do something when one has begun the execution of an intention to do a
certain thing right away. The process might not get very far. In trying to raise one’s arm, one
might not get one’s arm to budge. The muscles in one’s arm might not contract. One might bring
about no change outside of one’s head. A man whittled down to a brain in a vat might try to raise
his arm (an arm that, in fact, he doesn’t have). But if one is trying to do something, there has
begun the execution of an intention one has to do something right then.
Some theorists think of this noncausally: tryings aren’t caused by anything, or they
needn’t be. Or one might think of tryings as causings: my trying to do something is my causing
something. I think of trying as causal in a different way: tryings are events caused by the
intentions of which they are at least partial executions.
If I decided to A, or if I tried to A, I willed to A. Deciding to A is willing to A. It might
not be quite right, however, given my understanding of trying, to say that trying to A is willing to
A.5 As I’ve said, I take it that when one tries to A and succeeds, one’s attempt to A is one’s Aing. But it doesn’t seem to be always the case that doing what one tries to do is willing to do that
thing. If I try to sink a putt and I succeed, it doesn’t seem that my sinking the putt is my willing
to sink the putt.
I’ll take it that in such a case willing to do the thing in question is some early portion of
one’s attempt. It’s a beginning of the execution of one’s intention to do that thing right away. It’s
an initiation of an attempt. (Just how much of one’s attempt consists in the willing is vague, as
vague as is willing.)
Willing might go on just about as long as an action does. Suppose that I slowly wave my
arm in a figure eight. My moving my arm continues for several moments. So does my willing to
wave it in a figure eight, even if my willing isn’t the whole of my action. We can think of the
action as a process, one that begins in my head where my having an intention to so wave my arm
right away begins to excite certain neurons,6 that continues with the excitation of neurons
running to my shoulder, the contraction of certain muscles, and the motion of my arm. There will
be a beginning portion of such a process occurring for several moments, as long as my having
that intention continues to excite the appropriate neurons inside my head.
There might be other kinds of thing that should count as willing. But in what follows I’ll
focus on these, deciding and the initiation of trying.
The willings that I’ve identified share the feature of being at least partial executions of
intentions. When you decide to A, you carry out an intention. The intention might be to make up
your mind what to do, or to make up your mind whether to A, or whether to A or to B. I don’t
suppose that when you decide to A you carry out an intention to decide to A. You nevertheless
execute an intention to do something right away. And when you initiate an attempt to A, your
willing to A is a partial execution of some intention, perhaps an intention to A. I suspect that
everything that should count as a willing shares this feature (though of course not everything
with this feature is a willing).
2. Up to You
It’s a common thought that if you have free will, then at least sometimes when you act, it’s up to
you whether you do the thing you do on that occasion. Applying this idea to willing, we can say
that you’re free with respect to willing to A only if, on the occasion in question, it’s up to you
whether you will to A then.
Where the willing at issue is the making of a certain decision, say, the decision to B, if
you’re free with respect to making that decision, then it’s up to you whether you decide then to
B. Where initiating an attempt to C is in question, you’re free with respect to doing so only if it’s
up to you whether you initiate such an attempt on that occasion.
We often use the expression ‘it’s up to you’ in a way that isn’t concerned with free will. I
might have said to my daughter on some occasion: “You can straighten up your room, or you can
stay home; it’s up to you.” A certain outcome–whether she goes out or stays home–depends on
whether she straightens up her room. Whether she does the latter or not will be the differencemaker. The choice is hers. But all this might be so whether anyone has free will or not.
Similarly, if I have no preference regarding what you do on some occasion, or don’t wish
to express one, or refuse to offer any suggestions or advice or exert any pressure, I might say the
matter is up to you. Make the decision without my direction. That’s something you can do
whether we have free will or not.
But if we have free will, then not only do certain outcomes depend on what we will to do,
but also it’s sometimes up to us whether we will certain things. And its being up to us whether
we will certain things isn’t a matter of no one’s offering us advice or direction. Indeed, it can be
up to you whether you will a certain thing even if someone is offering you advice, expressing his
or her preference about what you do, or exerting pressure on you to do a certain thing.
I take this requirement to be a constraint on construals of powers possession of which by
an agent amounts (perhaps with further conditions) to that agent’s having free will. At least one
such power must be such that, when an agent has it, it can sometimes be up to the agent whether
it’s exercised.
It might not be a requirement of moral responsibility that we have such a power.7 It might
be that I can be responsible for doing something even though it wasn’t up to me whether I did
that thing on that occasion. I take the possibility here to be the possibility that we might not have
to have free will in order to be morally responsible for some of the things we do.
That we have such a power is nevertheless, I submit, part of our ordinary conception of
our agency. When I’m deliberating about whether to A, I take it that it’s up to me whether I
decide to A. I presume not just that my decision will be a difference-maker with regard to what
happens subsequent to it, but that it’s up to me whether I make one decision or another. It is free
will of the sort that is presumed in this kind of case that is my target here.
3. Powers
Powers are a class of properties including dispositions, tendencies, liabilities, capacities, and
abilities.8 I include here both what Locke called active powers and what he called passive
powers. Defining this class of properties might be difficult, but it’s easy enough list some
members. In discussing dispositions, for example, philosophers frequently focus on such things
as fragility, solubility, flexibility, and flammability.
Whether the several terms I used–‘dispositions’, ‘tendencies’, etc.--pick out distinct
species of powers I don’t know. But I do think that some things that are sometimes assumed in
philosophical discussion of dispositions should not be assumed about all powers. And if we wish
to understand free will in terms of agential powers, we’ll need to see whether the sometimes
assumed things apply to the powers in question.
Each of the dispositions mentioned above can be designated, as I’ve done here, using a
single-word term. It’s sometimes taken for granted that each such term can be defined in a
certain standard way, one that specifies a characteristic stimulus and a characteristic response for
the disposition in question. For example, it might be suggested, as a rough first approximation,
that ‘fragility’ can be defined as ‘the disposition to break in response to being struck’. The
template ‘the disposition to R in response to S’ is supposed to provide a canonical form for
definitions of such disposition terms. And defining a disposition term in this manner is often held
to be the first step toward an analysis of attributions of dispositions.9 For example, given the
simple proposal regarding fragility, an analysis of ‘o is fragile’ might take this statement to be
equivalent to ‘o is disposed to break in response to being struck’. Often some further analysis of
this latter sort of claim is then offered.
I don’t think it should be assumed, and I doubt that it’s true, that all powers are amenable
to this kind of treatment. In particular, I doubt that every term expressing a power can be defined
in the canonical fashion supposed in the first step of the process just described; I doubt, further,
that for every power there is some familiar term designating that power that can be so defined. It
might be that all terms expressing dispositions can be, and if that’s so then what I suspect is that
some powers aren’t dispositions; dispositions would be a species of power, differing in this way
from some others.
I accept that all powers are powers to do something--to produce or undergo or inhibit or
resist some change, or to remain in or sustain some state.10 Each has some (perhaps more than
one) characteristic manifestation. What I doubt is that every power has some characteristic
stimulus that can be identified by semantic analysis of a familiar name for that power.
Free will aside, there seem to be examples of powers that do not. Consider certain
tendencies or susceptibilities that might be thought of as passive powers. People with narcolepsy
tend to fall asleep. Episodes of sleep that manifest narcolepsy might well have characteristic
triggers, and it might be that these triggers are co-causes, with the narcolepsy, of these
manifestations. But it would take empirical investigation to find out whether this is so and, if it
is, what the characteristic triggers are. It isn’t a matter of semantic analysis of ‘narcolepsy’;
linguistic competence and analytic subtlety won’t reveal it. ‘Narcolepsy’ doesn’t lend itself
readily to the kind canonical definition just described.
Other powers might have characteristic stimuli that are identifiable by semantic analysis,
but lack semantically identifiable stimuli that, given possession of the powers, are bound to
produce their characteristic manifestations. The issue here doesn’t concern indeterminism. It
might be that every manifestation of such a power is fully determined. It might nevertheless be
that there’s no stimulus condition identifiable by semantic analysis of any familiar name of the
power that, together with the power, determines that the manifestation will occur. What, if
anything, determines whether the manifestation occurs, given possession of the power and
occurrence of the stimulus, might be discoverable only by empirical investigation.
Philosopher sometimes speak of “sure-fire” dispositions, sometimes taking these to be the
same as dispositions whose manifestations are determined by their stimuli. The powers I have in
mind here aren’t good candidates for sure-fire dispositions, even if determinism is true. For
analysis of familiar names referring to them might fail to reveal what determines whether or not
they’re manifested. I suspect that very many of our familiar names for powers, such as
‘irritability’, ‘diligence’, and even ‘fragility’ are of this sort. It isn’t semantic analysis that shows
us exactly what kind of striking, or what in addition to striking, suffices to cause fragile things to
break.
Finally, there appear to be powers that simply don’t have any relevant stimulus
conditions. Everything with rest mass has, in virtue of having rest mass, a power to curve spacetime. That power is manifested constantly, so long as the thing retains rest mass. Beyond
possession of this property, there seems to be no stimulus needed or relevant to whether the
power is manifested. Other powers manifest spontaneously. The instability of some particles or
elements is manifested in spontaneous decay. There seems to be no stimulus needed and none
relevant to whether the kind of decay in question occurs.11
When it comes to the kind of power that Locke focused on–the freedom of an agent with
respect to doing what the agent wills or might will–there seems to be an obvious relevant
stimulus: the willing of that thing.12 I think it doubtful that such a power is sure-fire, even if
determinism is true. One can have a power to do something in response to willing to do it and yet
sometimes fail in one’s attempt to do it.
It seems even less likely that the freedom of an agent with respect to willing a certain
thing is a sure-fire power. Indeed, it isn’t so obvious what, if anything, might be the relevant kind
of stimulus when it comes to such a power. We might consider desiring to do that thing, desiring
most strongly to do it, preferring to do it, or judging it best; I’ll suggest later a certain kind of
intending. But previous attempts to construe free will in this way aren’t encouraging.
If having free will requires that it’s up to you whether you will this or that, there might
stem from this requirement a familiar difficulty faced by any such effort. As standardly
understood, the manifestation of a disposition that has a characteristic stimulus is dependent on
the occurrence of that stimulus. A glass that’s fragile might break in the absence of any
characteristic stimulus of fragility, but if it does, its breaking doesn’t manifest its disposition to
break in response to being struck.
And now, it’s hard to see how it can be up to me whether I now will to A if whether I
manifest my power to so will depends on whether some stimulus, the occurrence of which isn’t
itself be up to me, occurs. And to try to make it out that it’s up to me whether the stimulus in
question occurs is to begin a regress, one that certainly looks to be vicious.
I suspect that it is for reasons of this kind that some writers on free will (e.g., McCann
1998: 174) describe its exercise as a kind of spontaneity. Of course, it won’t do to see the
exercise of free will as just like the untriggered decay of an unstable atom. It isn’t up to anyone
or anything whether the atom decays at some particular time. And it’s hard to see how a
spontaneous manifestation of a power to will could be up to the agent in question. But the
problem of free will is hard.
Most of what I say in this paper will concern efforts to construe free will in terms of
powers that are susceptible to analysis in the canonical way described above, powers that have
characteristic stimuli on which their manifestations depend, and whose manifestations are caused
by, among other things, those stimuli. (Though, as I said, I won’t take these powers to be surefire in the sense explained above.) Toward the end I’ll briefly consider alternatives that construe
free will in terms of different sorts of power.
Reid (2001) maintained that the active powers of intelligent agents are utterly different in
kind from the powers of inanimate objects.13 Indeed, he took only the former to be powers in the
proper sense of the word. I don’t agree. But I do take seriously the possibility that no powers of
the kind possessed by inanimate things can bestow us with freedom of the will of the sort that we
commonly take ourselves to have. Indeed, it might be that we don’t and couldn’t have any such
thing. But we should explore the matter thoroughly before accepting any such pessimistic
conclusion.
4. A Recent Proposal
Might we find a way to understand having free will as a matter of having certain dispositions,
dispositions that are fundamentally the same as those of the inanimate objects around us,
dispositions such as fragility and solubility? Kadri Vihvelin has recently offered a detailed
account of this sort. I think it falls short. We might nevertheless benefit from attending to the
difficulties it faces.
Vihvelin distinguishes between what she calls narrow abilities and broad abilities.
Having a narrow ability to do a certain thing is a matter of having the skills, know-how, and so
forth needed to do it, and being sufficiently wakeful, healthy, strong, etc. to do that thing. It’s
having what it takes to do a thing of that kind. One might lack a narrow ability to ride a bike if
one has never learned how, or if one is asleep, or if one has a broken leg.
Having a wide ability to do a certain thing is having the narrow ability to do it and being
in circumstances that are friendly to the exercise of that narrow ability–circumstances that
provide any necessary means and that lack insurmountable obstacles or impediments. You might
lack a broad ability to ride a bike, despite having a narrow ability to do so, if there’s no bike
available, or if there’s one present but someone would snatch it away should you try to get to it.
What’s added to a narrow ability by a wide ability includes what is sometimes called
opportunity.
Having the free will we think we have, Vihvelin says, consists of having both narrow and
wide abilities to act. The former are required by the latter. I’ll focus here on what she says about
narrow abilities to act.
Our narrow abilities, Vihvelin contends, are intrinsic dispositions or bundles of these,
each of which is fundamentally the same kind of thing as familiar examples of dispositions. Each
has a characteristic manifestation and a characteristic stimulus. In the case of abilities to act, the
characteristic stimulus will be some intrinsic change of the agent. But so it is in the case of an
alarm clock’s disposition to sound the alarm at the set hour. Many inanimate objects have
dispositions of this general sort.
When it comes to narrow abilities to perform intentional actions, Vihvelin proposes:
(Proposal) To have the narrow ability to do X is to have an intrinsic disposition
to do X in response to the stimulus of one’s trying to do X. (Ch. 6, p. 11)
The proposal is meant to cover all cases in which doing X is performing an intentional
action; I’ll take it that what is meant is intentionally X-ing. Hence I’ll take it that the proposal
concerns what it is to have a narrow ability to intentionally X. To have such an ability, it’s said,
is to have a certain disposition that has a certain characteristic manifestation and a certain
characteristic stimulus.
Is the manifestation supposed to be intentionally X-ing? I’ll take it that this is what is
meant. (Something that I note in my next paragraph supports this interpretation.) However, since
trying to A isn’t generally trying to intentionally A, I’ll take it that, on the proposal, the stimulus
is meant to be just trying to X (though, as I’ve said, trying to A is intentionally trying to A).
Vihvelin’s proposal, then, says that to have a narrow ability to intentionally do a certain
thing is to have a certain intrinsic disposition, one to intentionally do that thing in response to
trying to do it. She then suggests a view that further spells out what having such an ability comes
to, one that combines features of analyses of dispositions advanced by David Lewis and by
David Manley and Ryan Wasserman. The suggestion is:
(LPA) S has the narrow ability at time t to act in way R in response to the
stimulus of S’s trying to do R iff, for some intrinsic property B that S has at t, and
for some time t’ after t, and for some suitable proportion of test-cases, if S were
in a test-case at t and retained property B until time t’, S’s trying to do R and S’s
having of B would be an S-complete cause of S’s acting in way R. (Ch. 6, p. 25)14
In later (p. 26) restating the suggestion, Vihvelin speaks of “the narrow ability at time t to
intentionally do R,” and expresses the manifestation as “S’s intentionally doing R.” Hence I’ll
take it that, like Proposal, LPA is to be understood in this way.
An S-complete cause, following Lewis, is said to be a cause complete insofar as what’s
intrinsic to S is concerned. The test cases that we’re to consider are restricted to possible
situations in which the laws of nature are the actual ones. And they’re possible cases in which the
characteristic stimulus, S’s trying to do R, occurs, and in which this stimulus occurs “in
surroundings that count as providing a test for whether the person has the narrow ability; that is,
in surroundings where the extrinsic enablers for the ability are in place and where there are no
masks (extrinsic impediments) to the exercise of the ability” (26). The idea is that the agent has
the ability to A just in case, in enough such test cases, when she tries to A she A-s (and her A-ing
has the indicated agent-complete cause).
The analysans of LPA employs the notion of a test case; and this notion, in turn, is
explicated in terms of the very ability to act with which the analysandum is concerned. Hence the
suggestion doesn’t yet provide us with an analysis. We’ll have that only when we get a
characterization of test cases that doesn’t employ the notion of the ability that’s being analyzed.15
I’ll set this issue aside. What I want to focus on are difficulties that the approach faces
given that it’s meant to offer an account of free will.
5. Some Problems
A decision to A, I’ve said, is itself an intentional action. When one is uncertain about whether to
A and one makes up one’s mind to A, one actively forms an intention to A. What one does
commonly has a phenomenal quality characteristic of action. And one intentionally does that
thing; one means to be doing what one does then. There’s good reason why discussion of free
will has focused heavily on freedom in making decisions (or, what I take to be the same thing,
choices).
Likewise, as I’ve said, trying to A is an intentional action. Hence, the ability to decide to
A and the ability to try to A are abilities to perform certain intentional actions, to intentionally do
certain things.
Since Proposal is meant to cover to all narrow abilities to perform intentional actions–to
intentionally do things--and since LPA is meant to analyze such abilities, both should apply to
abilities to decide and abilities to try. But it’s doubtful that they apply to the former and certain
that they don’t apply to the latter.
When to do X is to decide to A, Proposal would have it that to have the narrow ability to
decide to A is to have an intrinsic disposition to decide to A in response to the stimulus of one’s
trying to decide to A. But it isn’t clear that when one decides to A, one tries to decide to A. One
might well try to make up one’s mind whether to A, but that’s another matter.
When to do X is to try to A, Proposal is even worse off. It would have it that to have the
narrow ability to try to A is to have an intrinsic disposition to try to A in response to the stimulus
of one’s trying to try to A. We don’t try to try. One might try to bring it about that one tries to do
a certain thing, but again, that’s another matter.
It seems, then, that Proposal’s applicability to abilities to decide is in doubt, and that it
certainly doesn’t apply to abilities to try.16 And since LPA appeals to the same kind of stimulus
in its account of narrow abilities to act, it runs into the same trouble.
Inapplicability to abilities to decide and abilities to try is no small thing. For freedom
with respect to deciding and initiating attempts is at the heart of free will. As Vihvelin observes,
“it’s possible for someone to be unfree due to her inability to try” (ch. 6, p. 43).
There’s a second difficulty for LPA that arises with regard to trying. I’ve taken trying to
be attempting. A successful attempt to A is an A-ing. In the case of such a success, the trying to
A is the A-ing. The trying can’t then be a cause of the A-ing.
Vihvelin doesn’t endorse precisely this view of trying. But she does say (ch. 6, p. 12) that
sometimes trying to A is just the beginning of one’s successful A-ing. In such a case, when one
succeeds in A-ing, the trying to A would seem to be a part of the A-ing (as the beginning portion
of a baseball game is a part of that game). A part of an A-ing wouldn’t be a cause of that A-ing.
So, on Vihvelin’s view as well, trying to A isn’t a cause of A-ing, even when one’s attempt to A
succeeds.
LPA takes trying to do R to be the characteristic stimulus of S’s narrow ability to R. But
such a stimulus wouldn’t be a cause of the agent’s R-ing, when this ability is manifested. A
disposition to R in response to trying to R wouldn’t then be a causal disposition. It would be
fundamentally unlike familiar examples of dispositions such as fragility, fundamentally unlike
even the alarm clock’s disposition to sound the alarm at the set time. For these are causal
dispositions. When they’re manifested, their characteristic stimuli are causes of their
manifestations.
One response to this difficulty would be to revise LPA so that it takes the stimulus to be
S’s beginning to try to do R and ends with, “S’s beginning to try to do R and S’s having of B
would be an S-complete cause of S’s completion of doing R.” But this way of formulating the
analysis underscores a point I made earlier. As it might be put here, our freedom with respect to
beginning to try to do what we do is at the heart of free will. And it seems that neither Proposal
nor LPA are applicable to an ability to begin to try. To begin to try to R is to initiate an attempt
to R. And an initiation of an attempt to R wouldn’t have the beginning of an initiation of an
attempt to R among its causes.
6. The Ability to Make Choices
Vihvelin takes the narrow ability to make choices on the basis of our values, reasons, and
reasoning to be not a single intrinsic disposition but a bundle of these. Some of the dispositions
that make up the bundle, such as a disposition to form and revise beliefs in response to evidence
and argument, aren’t abilities to act intentionally. At least generally, coming to have a belief isn’t
an intentional action.
One disposition in the bundle, Vihvelin suggests, is the disposition to form intentions in
response to desires and beliefs. Forming an intention in this way is at least sometimes making a
choice, or as I’ve called it, a decision. Making a decision is actively setting one’s will. When we
actively set our wills in this way, we do so intentionally. The ability to do this is an ability to
intentionally do a certain thing. Its sometimes being up to an agent whether she sets her will on a
certain occasion to do a certain thing is something required for that agent to be free with respect
to setting her will. This is partly constitutive of having free will.
But as I’ve said, it’s doubtful that Proposal and LPA are applicable to an ability to
decide. The latter takes the characteristic stimulus of acting in way R to be trying to do R. When
R-ing is deciding to A, the characteristic stimulus for the ability to perform this action would be
trying to decide to A. And it isn’t clear that when one decides to A, one tries to decide to A.
Suppose one does. In that case, the trying in question is itself an intentional action.
Vihvelin has offered no account of what an ability to do that kind of thing comes to. Since
there’s no trying to try, Proposal and LPA are inapplicable. Yet the ability to initiate attempts–
to begin trying–lies at the very heart of the free will problem.
7. A Power to Initiate an Attempt
Imagine a young child who sees a shiny object across the room, crawls over to the object, and
grasps it. The child, we might suppose, lacks the capacity to make choices for reasons that would
be needed for her to have free will. But there’s no reason to deny altogether that she’s an agent.
She has, and on this occasion exercises, many of the abilities that are distinctive of agency. As
Vihvelin says about the case, “The child acts intentionally; she tried to do something, and she
succeeded” (ch. 6, p. 12).
The child, it seems, exercises an ability to try to crawl over and get the shiny object.17
And likewise she exercises an ability to initiate an attempt to crawl and get the object. In this
case, the child can have the latter ability even if it isn’t up to her whether she exercises it on this
occasion; we’re allowing that the child lacks free will. One might have a power to will without
having free will, without having a power to freely will.
We might manage an account of an ability to initiate an attempt to do a certain thing if we
set our sights lower than free will, on something that the young child has despite lacking free
will. When one initiates an attempt to A, one begins to execute a present-directed intention, an
intention to do something right away. To be engaged in initiating an attempt is to be in the
process of implementing a present-directed intention.
We might, then, consider a construal of an ability to initiate an attempt to A as, at least in
part, a disposition to initiate an attempt to A in response to coming to have a present-directed
intention with relevant content. This approach would allow us to see an ability to initiate an
attempt as, at least in part, a causal disposition. We might think this an advantage, since, we
might think, we have a good understanding of causal dispositions. Or we might think that our
account of agency must see present-directed intentions as causes anyway, for we might think that
that’s the way to understand the implementation of an intention.
If we’re on the right track so far, then it appears that having abilities to initiate attempts to
do various things requires having a host of other powers. One must, for one thing, have a power
to come to have present-directed intentions. A being lacking this latter power doesn’t have a
power to try to do things, or to initiate attempts to act, no matter what it would do were it to
acquire a present-directed intention to act.
Intentions are motivated in light of beliefs. Hence, it seems, in order to have powers to
initiate attempts to act, one must have powers to come to have motivational states and beliefs.
A power to initiate an attempt to do some specific thing would require, it seems, having
some specific motivational and conceptual capacities. An agent who lacks a power to become
motivated to A, or to try to A, or to have any other relevant motivation, would seem to lack an
ability to initiate an attempt to A. Likewise for an agent who lacks a power to come to think of
doing a certain thing. A medieval knight would lack an ability to initiate an attempt to find an
iPhone, in part because he lacks a power to come to have a present-directed intention to search
for an iPhone. And he lacks the latter power because he lacks a power to think of doing such a
thing. It may well be that were he to intend to begin searching right away for an iPhone, he’s
straightaway initiate such an attempt. He still lacks the power to do so.
Even setting aside the problem of free will, then, a power to will includes a variety of
powers, many of them not powers to do things intentionally. Even relatively simple agents
capable of intentional agency are sophisticated beings.
8. Up to the Agent Whether She Wills
Our powers to will are, even if those of the young child aren’t, rational powers. And if we’re
eventually to have a conception of free will in terms of agential powers, it will have to include a
conception of rational powers. No agent with free will entirely lacks the latter.
If these are powers to do various things in response to various stimuli, the stimuli might
be a kind of seeing-as, or taking there to be reasons to do certain things. The powers to come to
have present-directed intentions that concern us might be understood as powers to do so in
response to taking there to be certain practical reasons. Likewise, it seems, for the powers to
become motivated to do certain things. The powers to come to believe might be understood as
powers to acquire beliefs in response to taking there to be evidence for those beliefs or
arguments in their support.18
I don’t pretend to be able to say precisely what any of these powers comes to. I’ve
suggested that they might be causal dispositions, with characteristic manifestations and stimuli. I
haven’t offered, and I’m not able to offer, an analysis of dispositions of this sort.
Even setting this point aside, the suggestions face a big problem. Supposing it granted
that we can construe in this way a power to rationally will, how do we construe a power to freely
will? For that requires not just rational powers; it requires that it’s up to the agents who have free
will whether they exercise some of these powers in certain ways on certain occasions. If the
exercise of our powers to will depends on things not themselves up to us, it’s hard to see how it
can be up to us whether we exercise these powers.
9. Opposing Powers
One thing that might help deal with this recurring problem is to accept that being free to will to
A requires having competing or opposing powers: if it’s up to me on a certain occasion whether I
will to A then, I must have a power to will to A and also some power to do something
incompatible with my willing to A. One such power would be a power to will not to A. Another
would be the sort of power that Locke appealed to when he finally (Essay Bk. 2, Ch. 21.48) took
the notion of free will seriously: a power to suspend the execution of one’s motivational states
while one evaluates their objects, considering whether they’re worthy of pursuit. This latter
might be called a power to reflect rather than willing, though there might be some willing
involved in reflecting. Reflection sometimes itself involves the active direction of attention and
pursuit of certain lines of thought.
Consider being free to decide to A. To have such freedom, it must be up to me whether I
decide to A. Suppose that I have a power to decide to A in response to coming to intend to make
up my mind whether to A, and also a power to decide not to A in response to that same stimulus.
It might be that having both of these powers, or two or more similarly opposing powers, is
required for being free to decide to A.
Consider being free to initiate an attempt to A. To have such freedom, it must be up to me
whether I initiate such an attempt. Suppose I have a power to initiate such an attempt in response
to coming to intend to A right away, and also a power to suspend execution of such an intention
in response to the same stimulus. It might be that having both of these powers, or two or more
similarly opposing powers, is required for being free to initiate an attempt to A.
If indeed having free will requires sometimes having such opposing powers, some writers
might reject the idea that the powers in question are dispositions. For some (e.g., Choi
forthcoming) maintain that no object can simultaneously possess opposing dispositions. But what
would the difference come to, such that powers of some sort admit of this opposition while
dispositions don’t? I’ll leave this questions to those who defend this view of dispositions.
10. Stimulus Presence
Suppose I have both of the powers to decide just mentioned: a power to decide to A in response
to coming to intend to make up my mind right away whether to A, and a power to decide not to
A in response to this same stimulus. But suppose I don’t now intend to so make up my mind, I’m
not going to so intend, and it isn’t up to me whether I so intend. How can it then be up to me
whether I decide to A?
Such circumstances commonly render it not up to an agent whether she does a certain
thing. Imagine that Sue’s standing on a tall ladder is necessary for her changing a certain light
bulb in a chandelier. Suppose that she isn’t standing on a tall ladder, she won’t be, and it isn’t up
to her whether she comes to stand on a tall ladder. It then seems that it isn’t up to Sue whether
she changes the light bulb.
It might contribute to conditions in which it’s up to me whether I decide to A if I now
intend to make up my mind right away whether to A. We might consider whether the presence of
this stimulus together with my having both the power to decide to A and the power to decide not
to A in response to it will suffice.19
The presence of such a stimulus surely isn’t necessary for its being up to me whether I
decide to A. For it might be that, although the stimulus is absent, it’s up to me whether it’s
present. I observed earlier that requiring such a thing would begin a regress that appeared
vicious. The suggestion here isn’t that it’s required, but rather that together with other conditions
it might be sufficient.
Suppose that I’ve just now taken there to be good reason to A. I’ve not yet become
motivated to A, but I have a power to become so motivated in response to taking there to be good
reason to A. And I have further powers: to come to intend to make up my mind right away
whether to A in response to coming to be motivated to A, to decide to A in response to coming to
have such an intention, and to decide not to A in response to this same stimulus.
We might decline to say that it’s up to me whether I become motivated to A. For coming
to have a certain motivation isn’t typically something we do intentionally, and we might think
that only in the case of things done intentionally can it be up to us whether we do them. Still, we
might consider whether the circumstances just described suffice for its being up to me whether I
decide to A. The characteristic stimulus for my power to decide to A isn’t present, but the
stimulus for some prior power is present, and I have that prior power, as well as powers linking
its manifestation to the presence of a characteristic stimulus of my power to decide to A.
These suggestions are sketchy, and I’m not able to determine whether, even if filled in,
they might give us a satisfactory account of free will. They do seem to me worth serious
consideration.
11. Non-Causal Powers
Some philosophers will think that we took a wrong turn early on, when the question of rational
powers was broached. The manifestation of a rational power is something done in the light of
reason, something done for a reason. And some philosophers maintain that nothing can be both
done for a reason and caused. A rational power, they say, can’t be a causal power. It can’t be one
that has a characteristic stimulus that, when the power is manifested, is a cause of the
manifestation.
Why must something done for a reason be uncaused? Something done for a reason is
responsive to the normativity of reasons. But causal processes, it’s sometimes said, “bring about
their effects with complete indifference to the question of whether those effects have cogent
considerations in their favour” (Lowe 1998: 156). No causal process, then, can result in
something that is done for a reason.
I don’t think this argument is correct. Taking there to be a reason for something can cause
some outcome. And there can be causal outcomes that are sensitive to whether earlier stages of
the processes leading to them consist of someone’s taking there to be reasons of various sorts.
Further, taking there to be a reason can be responsive to there actually being something that has a
certain normative force, just as taking there to be a horse can be responsive to there being
something with horsiness. We can have causal powers to recognize reasons as such just as we
have causal powers to recognize horses as such. One might doubt this if one doubts that there
really is any such thing as normativity, but the proponent of the argument sketched above
certainly doesn’t doubt that.
When one’s taking there to be a good enough reason to A is caused in an appropriate way
by there being a good enough reason to A, and one’s A-ing is caused in the right way by one’s
taking there to be a good enough reason to A, one can have A-ed for a reason. Coming to desire,
believe, and intend, and deciding and trying can be things that we do for reasons and also things
that are always caused. Powers to do these things can be causal dispositions. Rational powers can
be causal dispositions.
This is not yet to say that free will can be. For there remains the problem of
understanding how it can be up to me whether I do a certain thing if my power to do that thing is
a causal disposition of the sort we’ve been considering. If it is, then the manifestation of that
disposition depends on the occurrence of some stimulus, and it won’t typically be up to me
whether that stimulus occurs.
However, it’s hard to see how appeal to a non-causal or spontaneous power is going to
help here. It isn’t up to an unstable atom, or to anything else, whether that atom’s spontaneous
power to decay is manifested at a given moment; how can it be up to me whether some
spontaneous power I have is manifested? Nor does it seem to help to observe that the powers in
question are rational powers. As I argued above, there’s no good reason to think that such
powers must be non-causal. On the contrary, it is hard to believe that a power to come to believe,
desire, or intend for reasons could be anything but causal, with its manifestation caused by the
stimulus of one’s taking there to be a reason to do, or evidence in favor of, the thing in question.
12. Agent-Causal Powers
There’s a rather different understanding of causal powers that’s sometimes appealed to in
discussions of free will. When an agent freely makes a certain decision, it’s said, the agent
causes something, such as her coming to have a certain intention, and the agent’s causing that
thing isn’t causation by any occurrence or state. It’s causation by an enduring substance, which,
on this view, is what a rational agent is.20 A power to freely decide, then, is a causal power, but
its manifestation is some event caused by the agent (or perhaps it is the agent’s causing some
event). Such a power might have a characteristic stimulus, but, it’s usually said, the stimulus
isn’t a cause of the manifestation, nor (if the manifestation is taken to be the agent’s causing
something) of the thing that the agent causes.
Sometimes it’s said that there exists this kind of substance causation only in the case of
exercises of free will. All other causation, it’s said, is causation by events or states. I find it rather
incredible that causation might split in this way. Why would the events in rational agents become
so impotent? And why couldn’t any other substances cause things?
Other theorists (e.g., Lowe 2008, chs. 6 & 7) hold that all causation is, fundamentally,
causation by objects or substances.21 All causal powers, then, are powers of substances
themselves to cause certain occurrences. Free will might differ from many other powers in that
it’s a power that only a rational substance can have, and it might differ in that it’s a power of a
middle-sized substance to cause things, where its causing those things doesn’t consist in its
components causing things. But it has the same general form as all other causal powers: it’s a
power of something to be a substance-cause.
Substance-causal powers of this sort might have characteristic stimuli. Indeed, it’s
sometimes held that they must. An object causes something, it’s said, always by doing
something, or by undergoing some change (Lowe 2008: 146). It’s just that the cause in that case
isn’t the thing that the object does; it’s the object.
This kind of view of agential powers requires only minor alteration of the dispositional
view suggested earlier. A power to initiate an attempt will be a power of an agent to agent-cause
the initiation of an attempt in response to coming to have a present-directed intention. Powers to
intend, believe, and desire will similarly be powers of rational beings to substance-cause their
coming to have certain mental states in response to their taking there to be reasons or certain
evidence.
But now, precisely because this construal of agential powers involves such a minor
reformulation of the dispositional view, it’s hard to see that it constitutes any advance over that
view. If that dispositional view has trouble capturing the idea that when someone freely wills,
it’s up to that agent whether she so wills on that occasion, then it seems that a view of free will as
an agent-causal power will equally have trouble capturing this idea.22
13. Indeterminism
Some readers will think that I’ve failed to raise the key, or one of the key, requirements for free
will, namely, that its exercises be undetermined by what precedes them. If we’ve had trouble
capturing the idea that when you freely will something, it’s up to you whether you so will on that
occasion, that’s because we haven’t recognized that free will is incompatible with determinism.
I acknowledge that thinking of agential powers within determinism, it’s hard to see how
it can be up to agents whether these powers are manifested when and as they are. But I doubt that
thinking of their manifestations as caused but not determined makes the problem any easier.
Maybe in some way it does; but I’ll leave it to proponents of this idea to explain how.
32
Notes
1. Locke (Essay, Bk. 2, Ch. 21.10) holds that freedom (or as he sometimes calls it, liberty) is a
power of the agent. Hume takes it to be “a power of acting or not acting, according to the
determinations of the will” (Enquiry, Sec.8, Pt. 1). Van Inwagen (1981: 8) maintains that if an
agent has free will, then on some occasions there are two or more mutually incompatible courses
of action each of which is such that the agent has it within his power to carry it out. And Lowe
takes will to be a rational power, one “that is characteristically exercised in the light of reason”
(2008: 155).
2. Adams and Mele (1992: 326) take this view of trying.
3. I agree with some of McCann’s remarks on this matter. He says: “even though it is fair to say
trying attends all overt doing, trying should not be taken as a species of action equivalent to
volition. Rather, trying is a general term for the business of going about the performance of an
action, and an agent’s attempt consists in as much as he accomplishes, or could reasonably have
been expected to accomplish, toward the action at issue” (1998: 6). However, unlike McCann, I
see no need to posit volitions in addition to intentions and attempts.
4. Ginet (1990: 9-14) takes trying to be a mental action that, in the case of successful bodily
action, causes some bodily exertion. The mental action in question, which Ginet calls a volition,
need not have any cause, and it lacks internal causal structure; it doesn’t consist of one thing’s
causing another. (In contrast, although McCann posits volitions, he doesn’t identify them with
tryings.)
5. Stephen Kearns brought this point to my attention.
6. Brand (1984: 20) suggests this view of where and when actions begin.
7. We might formulate the conclusion of Frankfurt (1969) this way.
8. Some philosophers hold that all (genuine) properties are powers. I don’t rule this out here,
though nothing I say is committed to it, either.
9. Lewis (1997) suggests this procedure, and Choi (2008) follows it.
10. Molnar (2003: 60) calls this feature of powers “directedness.”
11. The examples are from Molnar (2003: 85-87).
12. However, I’ll point out later that such a “stimulus” is importantly different from, for
example, the striking that is a characteristic stimulus for fragility.
13. Reid maintained that power, in the proper sense of the word, implies will and is incompatible
with necessity.
14. Vihvelin calls this proposed analysis LCA-PROP-Ability. I’ve renamed it LPA for the sake
of brevity.
15. One might also worry about how to make sense of some proportion of infinitely many things.
Presumably for a given disposition there are infinitely many test cases, and infinitely many yield
the characteristic manifestation, while infinitely many don’t. Manley and Wasserman offer a
solution to this problem; I’m not sure what to make of it.
16. The same limitations afflict Fara’s Dispositional Analysis of an ability to act: “An agent has
the ability to A in circumstances C if and only if she has the disposition to A when, in
circumstances C, she tries to A” (2008: 848). Fara acknowledges the limitations when he says
that the analysis is meant to apply only to “actions that one can in principle try to perform”
33
(849).
17. Vihvelin remarks that the child “has no ability to think before she acts, to conjure up and
contemplate alternatives, to weigh ends, to deliberate for the purpose of deciding what to do, and
so on” (ch. 6, p. 12). But what’s at issue aren’t abilities of these kinds. Trying to A isn’t thinking
or deliberating before one A’s.
Something that we have, Vihvelin says, and the child lacks, is “the ability to make
ourselves try to do things, and the related ability to prevent ourselves from trying to do things”
(ch. 6, p. 13). It’s by having the ability to make choices for reasons, she says, that we have these
further abilities, and the child lacks them because she lacks the ability to make choices for
reasons.
But, again, the ability at issue is an ability to try to do a certain thing, not–if this is
something different–an ability to make oneself try to do that thing. Even lacking the ability to
make choices for reasons, the child has abilities to try to do various things. And she exercises
such an ability in the imagined case.
18. Pettit and Smith (1998) and Smith (2003) offer accounts of freedom in terms of capacities to
believe and desire rationally. Free will includes such capacities, but it includes more; let’s not
forget the action!
19. In work in progress, Thomas Reed proposes that having an ability to A consists in (i) having
a certain disposition, (ii) the obtaining of the stimulus conditions for that disposition, and (iii) the
absence of all extrinsic and certain kinds of intrinsic finks and masks for that disposition. It was
his work that suggested the idea in my text here to me.
Note to readers: Reed isn’t the eighteenth century Scottish philosopher Reid to whom I
refer in section 3 of this paper.
20. For views of this sort, see Chisholm (1966), Clarke (1993), O’Connor (2000), Pereboom
(2001, ch. 2), and Taylor (1966).
21. Note that Lowe doesn’t hold that decisions (or willings generally) are agent-caused; he takes
them to be uncaused manifestations of spontaneous powers. On his view, it’s by deciding (or,
more generally, willing) that a rational agent causes various things. Thus, it isn’t Lowe’s view of
free will that I’m discussing here, but rather one that construes agent causation as Lowe does and
takes willings to be agent-caused.
22. Jacobs and O’Connor (forthcoming) suggest a somewhat different view on which all causes
are substances, one that rejects Lowe’s claim that always a substance causes something by
undergoing some change. Still, they want to hang onto the idea that effects can be explained by
prior events. The events that do the explaining would seem to be somehow involved in
substances causing the effects in question, as that in virtue of which the substances cause those
outcomes. (Otherwise, how do they explain?) The difference from Lowe’s view thus seems to
vanish.