Tomis Kapitan: Deliberation and the Presumption of Open Alternatives

-- The Determinism and Freedom Philosophy Website --

What is the point of asking yourself what to do and then thinking hard about
it if all the thinking is settled in advance? What is the point of trying to figure
out how to run your life if determinism governs your every reflection? Do we
not have to suppose that determinism is false if we are to take our own
deliberations seriously? The question has long been taken to bedevil the
doctrine of determinism. It has been supposed that determinists can have no
good answer to it. Well, Professor Kapitan is one good philosopher who thinks
otherwise. He takes  things forward. His piece will repay your close attention.


By deliberation we understand practical reasoning with an end in view of

choosing some course of action. Integral to it is the agent's sense of
alternative possibilities, that is, of two or more courses of action he presumes
are open for him to undertake or not. Such acts may not actually be open in
the sense that the deliberator would do them were he to so intend, but it is
evident that he assumes each to be so. One deliberates only by taking it for
granted that both performing and refraining from any of the acts under
consideration are possible for one, and that which is to be selected is
something entirely up to oneself.

What is it for a course of action to be presumed as open, or for several
courses of action to present themselves as a range of open alternatives?
Answering these questions is essential for an understanding of deliberation
and choice and, indeed, for the entire issue of free will and responsibility.
According to one common view, a deliberator takes the considered options
to be open only by assuming he is free to undertake any of them and,
consequently, that whichever he does undertake is, as yet, a wholly undeter-
mined matter. Built into the structure of deliberation, on this theory, is an
indeterministic bias relative to which any deliberator with deterministic
beliefs is either inconsistent or condemned to a fatalistic limbo. An unmis-
takable challenge is thereby posed: is there an alternative conception of the
presuppositions underlying deliberation more congenial to a deterministic
perspective yet adequate to the data? Convinced that there is, I develop a
partial account of deliberation which, though highly similar to the aforemen-
tioned view, diverges at a critical juncture.

I.  The Postulate of Freedom

That a deliberator presumes himself to be free to undertake any one from

a range of alternatives seems undeniable. While such an attitude might not
involve the agent's knowledge that there are undetermined actions, choices or
deliberations, it is often thought to include his belief to this effect. Perhaps
Kant had this in mind when he set forth his celebrated postulate of freedom:

    It IReason] must regard itself as the author of its principles inde-

    pendent of foreign influences. Consequently, as practical reason or
    as the will of a rational being it must regard itself as free) that is to
    say, the will of such a being cannot be a will of its own except under
    the idea of freedom.1

This passage has been interpreted as implying that agents must adopt an

indeterministic stance with respect to their own practical thinking, or some
portion thereof, that this is essential to the conviction that their choices are
their own.2   If freedom and agency are so mated within practical reason, it
follows that any deliberator who also believes his future acts and choices tc
be (already) determined is ipsofacto inconsistent.

My object is not an exegesis of Kant. Many contemporary philosophers .

advocate this interpretation of the presumption of open alternatives, for
example, Hector-Neri Castaneda, who writes:

One of the fundamental facts about practical thinking is that ii

hinges on the agent's presupposition that he can choose from severa
alternative courses of action open to him. This does not, of course
imply, as Kant firmly stressed, that the agent is free in the sense thai
his acts, or his volitions, are uncaused. Perhaps the presuppositior
is just a dialectical illusion (to use Kant's term) of practical thinking
If it is, the universe is ugly: given the biological and psychologica
primacy of practical over contemplative thinking, we are, thus
condemned to presuppose a falsehood in order to do what we thinl
practically. We must in any case include the presupposition oi
freedom in our analysis of practical thinking ... 3

But what falsehood is it (if the universe is "ugly") that an agent is condemned

to presuppose in order to engage in practical thought - that he can choose
from among open alternatives, that he is free, or that some of his own acts or
volitions are uncaused? In the passage cited these disjuncts are conflated
Castaneda evidently holds that if someone assumes he can choose then he is
committed, qua rational being, to the presupposition that he is free and, thus,
that some of his choices are uncaused and undetermined.

Richard Taylor and others arrive at the same conclusion by focusing on

agency; one who deliberates about what to do must assume that his eventual
undertaking is his to choose, "under his control" or "up to himself". Were he
to suppose that his choice will be the outcome of antecedent conditions over
which he has no control, he could not take his eventual act to be up to
himself. Taylor is insistent, in short, that assuming the latter is to suppose
one's choice alone will determine the undertaking, not some other conditions
existing prior to choice. Consequently, a deliberator must take his choice to
be undetermined.4  With a slightly different emphasis, Nicholas Denyer
argues that since determinism entails the future to be fixed and necessary,
but that one deliberates only about what is taken as contingent, it follows that
"a deliberator cannot then consistently believe that his actions are deter-
mined by events prior to his deliberations."5   Denyer stresses the modality
embedded within the presupposition of freedom; to hold that one is free,
that one both (an perform an action and ran refrain from performing it, is to
assume that one's future undertaking is as yet a contingent matter. This
assumption, he claims, conflicts directly with the belief that one's choices
and actions are already determined by past or present conditions. Reflecting
both approaches, Peter van Inwagen concludes that since we all believe in
our own freedom,

. . . to reject free will is to condemn oneself to a life of perpetual

logical inconsistency. Anyone who rejects free will adopts a general
theory about human beings that he contradicts with every deliberate
word and act.6

That a deliberator does not view himself at the mercy of an indifferent

causal network is, to an extent, unquestimable', his assumption of self-agency,
of his power to choose, is at once a recognition of his partial independence
from the flow of events and of his ability to shape an indeterminate future.
The Kantian postulate of freedom, coordinating agency and contingency, is
well-grounded in the phenomenon of choice, and there is no intent to
oppose it here. Yet, what this presumption of freedom amounts to is not
something which the data unequivocally reveal. The reading so far encoun-
tered, henceforth labeled the "Standard Interpretation", must be measured
against the overt dissent of those who, while deliberating, take their actions
to be caused by their volitions, and these volitions, in turn, to be terminal
points of deliberations whose every phase is determined. To believe in free
will while taking it to be an illusion is not a comfortable position to be in. But
for this very reason, the presence of deliberating determinists, while not
refuting the Standard Interpretation, motivates development of and interest
in a rival account.
II.  The Presumption of Efficacy

To fix intuitions, let us consider an example of a man on a leisurely hike

through the countryside who unexpectedly comes to a fork in the path and
stops to deliberate about which branch to follow. Suppose, as he looks down
each path and weighs the advantages and disadvantages of taking it as
opposed to the other, a companion asks him about what he is thinking. We
can imagine the following exchange:

Companion: Do you feel that you can take either of the two paths?

Hiker. Certainly, I can take either of the paths, depending upon

which one I choose.

Companion: Can you tell, at this stage, which path you will eventually


Hiker. No, not now; I've not yet made up my mind on the matter.

Companion: Arc you aware of anything which will cause you to take,
or to choose, either the path to the right or the one to the left?

Hiker. Well, I hadn't thought about that, but now that you ask I

guess that I must say no, I am unaware of any such thing; as far as
I can tell it is entirely up to me which path I take.
Companion: Would you say, then, that you are free to choose either
the path to your right or the one to the left?

Hiker. Indeed, haven't I just told you that I can choose either?

Let us assume the hiker's responses to be typical of what one might expect

from a normal deliberator satisfying at least minimal conditions of rationality,
and so let us exploit the example as a springboard for conjectures about
deliberation. His response to the initial question, for instance, immediately
suggests an underlying attitude; he takes each alternative to be open only
because he feels that he would perform it if he chose to and that, otherwise,
he would refrain from so doing. That is, he assumes his will to be both
necessary and sufficient for the action, viz., that his choice would be
efficacious in bringing about his performance or non-performance of any of
the considered options. Generalizing, we propose a schema attributing what
can be called a presumption of efficacy:

(PE)  an agent presumes that his A-ing is an open alternative for

him only if he presumes that he would A if and only if he were
to choose to A.

A schema of this sort conceals much. A more detailed version would require

temporal indices fixing the times of the presumption, choosing and doing,
and, in many cases, probability qualifiers on the biconditional within the
scope of 'presumes', when the agent does not think his intentional efforts
will be guaranteed success. The term 'choose' may give way to 'decide',
'intend', 'undertake', 'try', etc., though in using 'choose' I assume that choice
is a species of intending to do something or other. In addition, the occurrences
of 'he', 'his' and 'him' within attitudinal scope should be taken to convey the
agent's self or first-person reference, and hence are limited as to their
possible substituends.7 For the present, these refinements can be left

The embedded biconditional in the consequent of (PE) poses no special

problem, whether construed subjunctively or indicatively. Obviously the
assumed linkage between A-ing and choosing to A is not purely logical, but
causal, and thus context-bound. That is to say, the agent takes his A-ing to
be consequent upon his choice gwen circumstances as they are, a qualifier
implicitly within attitudinal scope which could be more precisely exhibited by
a restricted universal quantifier over circumstances. This reading allows the
agent to be mistaken in his presumption of efficacy without saddling him
with suppositions he might recognize to be inconsistent - as would be
permitted on an external reading of the qualifier. An important feature of
(PE), of course, is that the deliberator takes his choosing to be an essential
factor in causal chains leading up to either his doing or refraining. This is
crucial to the sense of agency; that the action is under his control stems
partially from the supposition that he would do it only through his own
conscious effort.

A word about 'presumes'. It would be incorrect to think that a deliberator

is always conscious, via some prepositional attitude, that the alternatives he is
weighing are open to him. More likely, certain dispositional states are
involved, e.g. beliefs. But since 'belief has calcified in the lexicon of some to
imply an ability to articulate the content, perhaps what we want are lower-
level doxastic states - better conveyed by terms like 'feels', 'assumes' or
'takes for granted' - states for which corresponding linguistic abilities may
be lacking. For convenience, 'presumes' shall be used to indicate doxastic
attitudes gencrically, allowing the character of the relevant dispositions to
fluctuate among various doxastic levels.8
III.  The Presumption of Contingency

At first glance, (PE) might be thought to be all that there is to the

presumption of open alternatives. Recalling the conditional analysis of
freedom championed by G. E. Moore and others, why not say that a
deliberator takes a course of action to be open just in case he believes it
possible that he perform it and possible that he refrain, with the modalities
unpacked conditionally as indicated by the consequent of (PE)? Unfortuna-
tely, even if one accepts the equivalence, the conditional analysis no more
provides for the agent's sense of freedom than it does for an account of
freedom itself: to take a course of action as possible in that one would do it if
one chose will not suffice for taking it as open. Nadia, upon entering the local
ice cream shop, might believe that she would eat chocolate ice cream if she
chose, but may also realize that if she did she would break out in a horrible
rash. She might even dislike the taste of chocolate and have formed a belief
that because of this and her fear of a rash she will be caused not to choose
chocolate ice cream. Believing that her not eating chocolate ice cream is
already determined, therefore, she no longer considers it an open alternative
despite her acceptance of the conditional.9

It is tempting to say that a deliberator must also assume that it is possible
for her to clinnue a considered alternative, and it is precisely this that Nadia
lacks. It is evident, however, that applying the conditional analysis to this
sense of possibility would merely postpone the difficulty besides raising
familiar problems about choosing to choose.10  Perhaps such reflections have
led some to suggest that if an agent deliberates about A-ing then he assumes
that his A-ing is still a contingent matter and that, consequently, nothing yet
determines his choice either to A or not to A. Nobody, as Aristotle empha-
sized, deliberates about that which is impossible or necessary. Of course, to
avoid a facile refutation of determinism it is essential to view the modality as
within attitudinal scope, so that we have,
(1) an agent presumes that his A-ing is an open alternative for him
only if he presumes that his A-ing is contingent,
which implies that he also takes his not A-ing to be contingent.11 Alter-
natively, one could speak of his choosing (intending, undertaking, etc.) A as
contingent, and again add that the schema is to be qualified by temporal
parameters, i.e., the agent assumes, while deliberating, that his A-ing at t is,
as of yet, contingent.

The problem now is to give some account of the modality in (1), for it is

certain that not just any sort of contingency will do. Mere logical contingency
is not sufficient, nor, for that matter, any other sort of contingency fixed solely
by reference to a body of laws of nature. Instead, a modality which includes
reference to the actual course of events, to the world of particular objects
and conditions, is required, viz., a relativized, concrete modality. Let us say
that a state of affairs (event, proposition, etc.) P is contingent relative to a set of
conditions S just in case neither P nor not-P is a consequence of S. The
contingency is concrete if S contains particular facts or conditions, and
unqualified with respect to time t if S contains all conditions existing prior to
and at t.12

For Taylor and Castaneda, the contingency in (1) is, at least, causal in that

the agent assumes that there do not exist, nor have existed, conditions
causally sufficient for his (t)-ing at the time in question. One could, alter-
natively, drop mention of causation, as van Inwagen does, and say that the
agent assumes his A-ing is not a consequence of any set of conditions (plus
laws of nature) antecedent to and including the time of deliberation. In
either case, the agent takes the contingency to be fixed with respect to all
standing conditions, past and present. Denyer, even more strongly, opts for a
type of absolute contingency; in no sense is the agent's A-ing taken as
necessary or impossible, or, in other words, it is not a consequence of any set
of truths. A deliberator must, he contends, assume that neither the proposi-
tion that he will A nor the proposition that he will not A is already true, so
that no truth about what happens in the future entails a proposition to the
effect that he will A or that he will not A.

Each of these construals of (1) is a variant of what I have previously called

the Standard Interpretation. In the present context its claim is that one who
takes his A-ing to be open assumes it to be contingent relative to all
conditions (facts, events, propositions) existing (obtaining, occurring, being
true) prior to and including the time at which the assumption is held.
This unqualified modality requires the deliberator to consider his A-ing to
be, as yet, undetermined by those same conditions, hence, undetermined

To minimize complexities, define determinism broadly as the doctrine

that each state of the world is fully determined by antecedent states, where P
is determined by Q just in case the existence (obtaining, occurrence, truth) of
Q is sufficient for the existence of P. Following Denyer and van Inwagen,
determinism implies that at any instance there is just one possible future--in the
unqualified or causal sense of 'possible'. Of importance is the fact that
(2) a determinist assumes that whatever he will do (choose, under-
take, etc.) is already determined.

To locate an inconsistency within the beliefs of a deliberating determinist

now seems easy; for as a deliberator, by (1), he takes his future act to be yet
undetermined, but as a determinist, by (2) he assumes the very opposite, that
it is already determined.

But matters are not so simple. To say that a determinist who deliberates

about a range of actions A1,...,An supposes that whatever he will do is
already determined is not to imply that he takes his Ai-ing to be determined,
1   i   n.  The quantifier 'whatever' in (2) falls within the scope of
his assumption, so that he need not believe of any specific action that it is
already determined. We cannot, then, automatically attribute to the deter-
minist who deliberates about whether to A the bald inconsistency of both
believing that his A-ing is determined and that it is not.

One could argue from the claim that it is impossible to deliberate about

what one knorns one will do.13  If one knows one will A then there is no point
in deliberating about whether to A; the issue is already settled and A-ing is
no longer open but closed. Indeed, i/this is so it seems fair enough to
generalize to belief as follows:

(3) an agent presumes that his A-ing is an open alternative for him

       only ifhe does not yet believe that he will A.

Now it is implausible that the consequent of (3) be satisfied if one believes

one's A-ing is determined, that is, for minimally rational agents, (3) yields:

(4) an agent presumes that his A-ing is an open alternative for him

      only if he does not yet believe that there are conditions sufficient
      for his A-ing.

So, the argument goes, satisfying the consequent of (4) renders inconsistent

any determinist who believes he will undertake at least one of the alternatives
about which he deliberates. But this reasoning is also deceptive. Schema (3)
is plausible only if negation has larger scope than the attitude within the
consequent and, if so, it ascribes no belief at all to a deliberator. (4), however,
generates the inconsistency only when negation has smaller scope, in which
case it derives no support from (3). The confusion stems from the fact that
expressions of the form 'he does not believe' are used to express both
disbelief and nonbelief - an unfortunate ambiguity, but devastating for the
argument at hand.

At the same time, this argument indicates where the inconsistency is to be

found, if the deliberator is minimally rational and believes he will undertake
one of the alternatives. For, by (1), he assumes of each alternative that his
undertaking it is contingent and, thus, that there is, or will be, a future
undertaking which is, as yet, undetermined. This consequence, nn the
Standard Interpretation, involves a belief which does conflict with that
ascribed in (2), and the ascription of an inconsistency to deliberating deter-
minists is secured.

A showdown with the Standard Interpretation over (1) is inescapable.

That deliberation is wedded to a sense of contingency is manifest in our
example of the hiker. But examine his response to the companion's third
question. Taken literally, the words 'as far as I can tell' suggest an interpre-
tation of (1) in terms of epistemic contingency; a deliberator assumes his A-
ing to be contingent relative to what he knows. However, more seems
involved. I may, for instance, believe I will not fly to Copenhagen tomorrow
and thus I do not deliberate about so doing, yet I may not. know what I
believe (perhaps some unforeseen emergency will call me to Copenhagen).
The action is impossible relative to what I believe and so does not appear
open to me, though it is contingent with respect to what I actually know. The
words 'as far as I can tell', in fact, point to a broader construal of the modality
in terms of doxastic contingency so that (1) would give way to something like

(5)  an agent presumes that his A-ing is an open alternative for him

      only if he presumes that his A-ing is contingent relative to what
      he then believes,
where 'then', occurring before 'believes', refers to the time of presumption.

(5) says, simply, that the agent takes no set of his beliefs to be sufficient for

his A-ing or for his not A-ing.

   As a necessary condition on deliberation (5) appears uncontroversial, but

before judging whether it captures the full flavour of (1) a further issue must
be addressed. Given that the modality falls within attitudinal scope is the
same to be said for the qualifier 'relative to what he then believes'? If it has
an external occurrence then the modality would be fixed by the entire body of
the agent's beliefs with the consequent of (5) reading: for every subset S of
x's beliefs, x presumes that his A-ing is contingent relative to S. The
problem here is that since no one consciously rehearses all his beliefs while
deliberating he may overlook what they entail or even what he believes they
entail. Suppose at 10 a.m. Mr. Hawkins, having decided to take his son
bowling at 3 p.m., acquires the bclieFthat he will take his son bowling then.
At 2 p.m., temporarily overlooking his earlier resolve, he deliberates about
playing golf at 3 p.m.. Given all that he believes (dispositionally) at 2 p.m. it
is not true that he assumes it possible that he play golf at 3 p.m. and, so, (5)
would fail to formulate even a necessary condition. Weakening the conse-
quent to refer to only some subsets of x's beliefs would saddle the condition
with the same insufficiency that affected epistemic contingency. An external
occurrence of the qualifier, in short, renders (5) unsuitable.

A solution is to insist upon an internal occurrence. This allows us to take

the hiker's response at face value; by using the words 'as far as I can tell' he
relativizes the modality to what he then take! himself to believe. As such, the
occurrences of both 'he' and 'then' in the qualifier function in just the way
that 'his' does within the scope of 'presumes', namely, as devices for
attributing self-reference to the agent (see note 7).

A residue of ambiguity lingers. There are questions whether the scope of

'what he then believes' is to include that of the modal operator and whether
'what' indicates a quantifier occurring outside or inside the scope of 'pre-
sumes'. The first, I think, can be answered affirmatively since the qualifier
specifies the character of the modality. The second turns on a choice
between, roughly, (i) x presumes that if S is any set of his beliefs then his <j)-
ing is contingent relative to S, and (ii) there is a set S such that x presumes
that S is the set of his beliefs and his d)-ing is contingent relative to S. (i)
bears a structural accord to the Standard Interpretation where quantifiers
implicitly occur within attitudinal scope; its satisfaction is a minimal require-
ment. (ii), on the other hand, would seem to imply that a deliberator
consciously reviews all that he takes himself to believe whenever the dis-
positional presumption ascribed in the consequent of (5) is activated.
Though (ii) is perhaps not to be ruled out, (i) is a more cautious reading. We
arrive, thus, at a version of (5) which can be labeled the presumption of

(PC) an agent presumes that his A-ing is an open alternative for

        him only if he presumes that if S is any set of his beliefs then
        his A-ing is contingent relative to S.14

It follows immediately from (PC) together with (PE) that anyone who takes

his A-ing as contingent relative to his beliefs thereby takes his choosing to A to
be similarly contingent, assuming, once again, minimal rationality.

What, then, can be said to favour this doxastic interpretation over the

Standard Interpretation of (1)? The issue is largely empirical, and a full-
fledged defence of (PC) must await the presentation of additional proposals
which as a body are to be measured against the data (see section VI). But
three minor considerations merit attention here. First, (PC) does provide a
sense of contingency useful for explaining some cases of non-deliberation,
e.g., that of Nadia and the chocolate ice cream. Second, one must avoid
defending the Standard Interpretation by appealing to (3) and its supposed
derivative (4), even if these are conditions on deliberation. The derivative
guarantees only that a deliberator does not believe his A-ing to be determined,
not that he assumes it to be undetermined, and a confusion over the scope of
negation in (3) and (4), I suspect, is one reason for the initial appeal of the
Standard Interpretation. Finally, the very existence of deliberating deter-
minists who deny holding indeterministic beliefs constitutes some evidence
that they do not. Of course, this observation must be tempered by the
notorious difficulty of establishing non-belief, particularly in this manner,
but as inconsistencies are not to be lightly ascribed, it shifts the burden of
proof to the opposition.15
IV.  The Analysis

Both (PE) and (PC) formulate necessary conditions for a course of action

to be presumed as open by an agent; jointly, they are sufficient. With
temporal parameters implicit once again, we have:

(PO) an agent presumes that his A-ing is an open course of action

        for him if and only if (i) he presumes that he would A if and
        only if he were to choose to A, and (ii) he presumes that if S is
        any set of his beliefs then his A-ing is contingent relative to S.

In yet other words, an agent takes his A-ing as open just in case he assumes

that his will is efficacious and that he both can A and can refrain from A-ing.
From this basic analysis the other properties of the presumption of open
alternatives can be derived (see sections V and VI below).

Nothing has been said about the underlying action theory that a principle

like (PO) might require, specifically, about what a course of action is. The
schematic letter 'A' is intended to have expressions designating what are
often called "action-types" as substituends, whether simple or compound.
However, for a theory admitting compound courses of action though not
compound action-types, (PO) is limited, and any attempt to extend or adapt
the proposals would require more groundwork. An appraisal of (PO) must
hear this in mind, but two points can be made here. First, if we view x's not
A-ing at t as the complement of x's A-ing at t then it is not difficult to see that
(PO) yields the desired result that a course of action is presumed as open by
an agent if and only if its complement is as well. Second, it is plain that
deliberation can also be hypothetical, as when one contemplates what to do if
some condition P holds, e.g., whether to complain if one loses.16 Courses of
action deliberated about on the supposition that P holds may be said to be
open-relative-to-P. It is easy enough to construct an analysis of this notion,
in turn, along the lines of (PO) with the obvious adjustments in both of the
clauses (i) and (ii). Once accomplished, and we acknowledge conditional
intentions, then the following should be a targeted theorem: x presumes that
his A-ing is open-relative-to-P just in case x presumes that "his A-ing if P"
is open for him.

Where (A1, .. ., An) is a set ofn distinct courses of action, then the central

principle on the presumption of open alternatives is no surprise:
(POA) an agent presumes that (A1, .. ., An) is a range of open
            alternatives for him if and only if (i) for each Ai , 1 i  n,
            he presumes that Ai is open for him, and (ii) he presumes
            that not all of (A1, .. ., An) are conjointly realizable.

Concerning (ii), it is allowed that an agent takes some of the members of the

range to be conjointly realizable, e.g. one might debate whether to go to the
butcher's, go to the baker's, or stay home while believing the first two to be
compossibic. Reference to the totality of the elements in the range is
presupposed. If by deliberative content we understand a set of courses of
action about which one deliberates, then a main assumption throughout has
been: a set of courses of action is a deliberative content for an agent only if
he presumes it to be a range of open alternatives for him. The converse does
not hold; a sense of the relative significance of the included items seems
required to secure a place in any deliberative content. That is, the presump-
tion of open alternatives is only a necessary condition for deliberation.

Some fine points can be touched upon. For one thing, it may be erroneous

to speak of the deliberative content if an agent can carry on several deliber-
ations simultaneously. Also, adjustments concerning temporal parameters
are needed to cover cases where a course of action comes to be dropped
from deliberative content during deliberation. Content can fluctuate and
what appears open at the onset of a deliberation may lose this character as
the process unfolds (or vice-versa). The failure of the main assumption
mentioned in the previous paragraph shows that inclusion of a course of
action in deliberative content does not guarantee inclusion of its comple-
ment. That is, one can deliberate about two "positive" acts, say, whether to
study French or Arabic, without consciously considering the complements of
either. (PO) demands only that if a course of action appears open then so
does its complement, not that if it is deliberated about then so is its
IV.  Indecision and Uncertainty

With (PO) and (POA) we have an analysis of a deliberator's presumption

of open alternatives. The similarity of this account to the Standard Interpre-
tation is apparent, but there is a fundamental divergence in the way each
handles a deliberator's sense of contingency. It remains to be seen whether
(PO) and (POA) can be used to explain other features of deliberation,
specifically, a deliberator's state of uncertainty and his sense that he is free to
choose. First, we consider the former.

Taylor, Ginct and others have argued that one cannot deliberate about

doing something if one already knows one will do it (see note 13). Our hiker,
for example, does not deliberate about the disjunctive act of taking the path
to the right or the one to the left if this is something He has already decided
upon and takes for granted he will do. His denial that he can tell which path
he will take and his words I've not yet made up my mind' point not only to
his ignorance or lack of belief about which alternative he will undertake but
also to his state of indecision. More directly, there is a connection between
deciding and believing what one will do which indicates that (3), if accept-
able, should be accompanied by:

(6) an agent presumes that his A-ing is an open alternative for him

     only if he has not already decided to A.

Initially, states of ignorance and indecision appear obvious as antecedents

to decision and, thus, as ingredients in the presumption of open alternatives.
However, objections have been raised against a requirement of ignorance
and, mutatis mutandis, against proposals like (3) and (6).17  For example, it
might he thought that a person could decide upon a given course of action,
believe he will succeed in his endeavour, and yet deliberate about it. The
hiker, having made up his mind to go left, may continue to reflect upon his
choice by considering likely benefits of going right or by attempting to locate
justificatory grounds for his preference. Though still engaged in practical
reasoning, he is no longer deliberating about whether to go left; taking the left
path, by supposition, has already been settled. On the other hand, his
subsequent thought may cause him to doubt the wisdom of his choice,
deliberate anew about the action, or even abandon his previous decision.
This possibility shows, at least, that (3) and (6) cannot stand in the form
given. Modified versions might insist that one cannot take (j)-ing as open
while at the same time intending to A and believing one will A. But even these
amendments face difficulties. Take the case of Mr. Hawkins who at 10 a.m.
not only decides to take his son bowling at 3 p.m. and acquires a belief that
he will do so but also instructs his secretary to remind him of this at 2:45
p.m. At 2:44 p.m., preoccupied with the day's business and having tempor-
arily overlooked his earlier resolve, he suddenly deliberates about whether to
play golf or to treat his son to a few games of bowling at 3 p.m. Has he
abandoned his previous decision? Not necessarily; that he sustains his
intention is evidenced by his ready acceptance of his secretary's reminder at
2:45 p.m., which reveals his existing dispositions not only to affirm that he
will take his son bowling but to have a volition to do so. Plainly, the contrast
of occurrent with dispositional states applies to intentions as much as to
beliefs and, when coupled with the fact that agents can overlook or forget
what they have previously accepted, this renders (3) and (6) open to such
counterexamples.18  Schema (4) falls prey to these as well insofar as Hawkins,
by satisfying (PE), views his decision as a determining factor, and, with
further modifications, the example casts doubt upon the more restricted
ignorance requirement.

How, then, are we to interpret the hiker's response to the second ques-

tion? That a decision terminates a period of indecision seems beyond doubt
and lends immediate credence to Ginet's claim that decision involves change
from a state of uncertainty into a kind of knowledge.19  Restricting (3), (4) and
(6) to occurrent beliefs and intendings might appear the best that can be
hoped for. However, a different sort of problem follows upon this sugges-
tion. The consequents of the conditions so modified still embody negation
with larger scope than the (occurrent) attitude. Ascribing no positive attitude
to deliberators, therefore, they add nothing to the content of the agent's
sense of openness and, consequently, are of no assistance in analyzing the
hiker's awareness that he has not yet made up his mind, i.e., his feeling of
indecision. To capture the latter we need, not (6), but a more complex
presumption of indecision:

(PI) an agent presumes that his A-ing is an open alternative for him

       only if he presumes that he has not yet decided whether or not
       to A.20

This condition does not do full justice to the hiker's admission of ignor-

ance. Being undecided falls short of a more encompassing state of uncer-
tainty, for it is conceivable that a person might predict his own future
undertaking without having yet decided upon it. Recalling our previous
observations, a deliberator's prediction cannot be ruled out when construed
dispositionally, and the mere exclusion of an occurrent attitude contributes
little in analyzing the attitudes identified with a state of uncertainty. Feeling
uncertain, while extending beyond a state of indecision, is not simply a
condition of ignorance, and, for that reason, (3) is deficient. A more suitable
means of accommodating the hiker's second response is a presumption of

(PU) an agent presumes that his A-ing is an open alternative for

        him only if he presumes that he does not yet believe whether
        or not he will A.

So, we can avoid the difficulties attending (3) and its suggested modifica-

tions, yet provide an immediate account of the hiker's professed ignorance.
What is essential to realize is that the consequents of (PU) and (PI) describe
a state of uncertainty which, being sensitive to cases like those of Mr.
Hawkins, preserves the core of Ginet's insight about decision.

These proposals are not unrelated. A minimally rational agent under-

stands that if something is or will be caused to occur then it will occur and,
thus that if he does not yet believe that it will occur then he does not believe
that it is or will be caused to occur. That is, on the rationality proviso, (PU)

(PS) an agent presumes that his A-ing is an open alternative for

        him only if he presumes that he does not yet believe that there
        are conditions sufficient for his A-ing or for his not A-ing.

By satisfying the consequents of both (PS) and (PE), in addition, a delibera-

tor realizes he is not committed to the existence of conditions sufficient for
his choice to A or not to A. If, by (PU), he presumes he does not believe
(know) that he will A, then, by (PS) and (PE), he presumes he has not yet
decided to A. That is, on the proviso, not only is (PS) a consequence of (PU),
but (PI) follows from (PU) together with (PE).

What about (PU) itself; can it be established on the basis of (PO)? The

answer, I think, is yes, assuming, once again, minimal rationality. Suppose
that a deliberator satisfies (PC); if he takes his A-ing not to be contingent
with respect to a set S then he will not regard S as a set of his own beliefs.
Since he is rational, he does not think his A-ing to be contingent with respect
to the set consisting solely of the proposition that he As at the time in
question. So, he does not view the latter as a set of his own beliefs, that is, he
believes that he does not believe he will A at that time. Therefore, he
satisfies (PU), and the sense of uncertainty emerges as a dimension of the
contingency assumption.

What we have with (PU) and its derivatives (PI) and (PS), in sum, is not a

deliberator's ignorance or indecision but, more cautiously, his disposition to
affirm his own non-commitment to a specific alternative. This in no way
implies that one who is conscious of being committed to A-ing feels compelled
to A; he may correctly assume that he is able to refrain from A-ing in that he
would refrain were he to so choose. But while his A-ing may be seen as a
possible alternative in this conditional sense, he still might not take it as open.
To see this, we need only to remind ourselves of the frequent claim that one
cannot do a certain thing because one has already decided to do something
else, and not because one's will would not be efficacious as regards that act.

Finally, a new light is cast upon the hiker's statement that he is unaware of

anything causing him to choose one way or the other, indeed, that he is
"free" to undertake either; his presumption of freedom includes recognition
of both his own uncertainty and lack of intentional resolve. As such, (PU)
codifies an additional feature of a deliberator's assumption of an open future
and causal independency from the past -- without the more sweeping
imputations of ignorance and indetenninistic beliefs. That it is a conse-
quence of (PO), given a modest assumption of rationality, is an indication of
the latter's strength.
VI.  Freedom to Choose

The largest hurdle remains; does the foregoing do justice to Kant's insight

that a deliberator's presumption of open alternatives is an assumption of
freedom? Much depends upon what precisely is meant by 'freedom' here, but
(PO) and (POA) do embody elements central to any reliable account of
practical freedom, specifically, (i) of ability to act, (ii) of contingency of the
eventual undertaking, and (iii) of non-commitment to a particular alternative. It
is a deliberator's sense of (i)-(iii) -- within a context fixed by what he himself
takes to be the case - that is his presumption of freedom and self-agency. In
this way, the above account presents a genuine philosophical contrast to the
Standard Interpretation which construes freedom in terms of undetermined
choice. Nevertheless, more is needed to show that the agent thereby takes
himself as free to choose from among the alternatives before him.

It has already been remarked that (PO) explains certain cases of non-

deliberation, e.g., why Nadia does not deliberate about eating chocolate ice
cream, or, why a compulsive truth-teller -- aware of his irresistible desires --
does not consider lying to his best friend. In both, the agent takes himself as
unable to choose, and it is failure to satisfy (PC) and (PS) that makes the
difference. In other cases both (PC) and (PS) might be satisfied but (PE) not:
consider a man in a room with a single door he believes to be locked but who
knows that at 11 a.m. it will either be flung open or remain locked; he cannot
deliberate about whether to open the door at 11 a.m. since he does not
envision his will as efficacious in the matter. Van Inwagen describes a man in
a room with two doors, one of which he believes to be locked and the other
unlocked though he does not know which, suggesting that he cannot deliber-
ate about which door to leave by. Failure to satisfy (PE), once again, can
account for this, though it need not prevent the man from deliberating about
which door to try. On the other hand, if deliberating about trying to A is, by
that very fact, deliberating about A -- thereby disputing the example -- then
the probability qualifiers implicit in (PE) lend it a desired flexibility.

How do we characterize the deliberator's presumption that he is free to

choose, that he has both the ability to choose to A and the ability to choose
not to A (or, to refrain from choosing A)? Consider Nadia who, having ruled
out chocolate, is still faced with a decision about which of the remaining 32
flavours to order. Suppose that she satisfies both (PE) and (PC) as regards
each alternative; does she, therefore, find herself free to choose? If at all
adequate, our analysis must sustain an affirmative response. But here comes
a challenge. Imagine that Nadia consciously believes what the local astro-
loger told her, namely, there is a certain flavour such that it is already
determined she will not choose it, though she has no idea which flavour this
is. Assume, moreover, that she is not so irrational as to also believe her not
choosing this flavour to be undetermined. Can she deliberate about which
kind of ice cream to order? We come to a critical parting of ways; Nadia's
deliberation is permitted as far as (PO) is concerned, but the Standard
Interpretation must rule it out. That is, since Nadia now fails to believe of
each alternative that not choosing it is undetermined, she does not take it
as unqualifiedly contingent, and so, by the Standard Interpretation, her
deliberation would be pointless', realizing that her choice is not entirely
"under her control", Nadia must remain without ice cream.

This latter assessment seems unreasonable. Why shouldn't the following

thoughts convince Nadia, and ourselves, that deliberation here does have a
point? Look, I am hungry for ice cream and I want to select a kind that is
both tasty and filling. I have definite likes and dislikes and I know I will order
a given flavour just in case I choose to do so. Moreover, I will choose a
flavour only through a conscious effort on my part, even if it is already deter-
mined, by the stars or whatever, that I will not choose one of them,
whichever it might be. As far as I can tell the matter is entirely under my
control: I can choose any one of the 32 flavours even though, at this stage, I
am undecided as to which. I must, in any case, try something and it is only
through deliberation that I will make the best choice.

In attributing to Nadia the belief that she is already determined not to

choose one of the flavours we are not supposing that she will be, or believes
she will be, prevented from so doing, i.e., that she would fail were she to
somehow try to choose one of the flavours. At the same time, in analyzing
her claim that she takes herself as able to choose we cannot simply ascribe a
belief that there is no obstacle preventing her from carrying out her will, and
in this, ability to choose differs from ability to do (which does involve such a
belief). But certain analogies persist. That X can A implies that it is
possible that X As, and it is essential that what is said to be possible is an
action of  X's, as distinct from a mere bodily movement where agency is not so
implied. Similarly, to say X can choose to A implies that it is possible that X
chooses to A and, with this, X continues to be viewed as an agent, a maker of
choices, and not merely a passive object in some event or state-of-affairs.

The substantive claim concerns not agency but modality; when Nadia

assumes it possible that she chooses any of the 32 flavours, the doxastic
interpretation suffices to unpack the modality. Notice that her situation is not
akin to one who feels he "has no choice" in the usual sense that his will
would not be efficacious, e.g., one who does not deliberate about hovering
unaided above the floor. Nor is it that of the compulsive truth-teller who
finds himself unable to lie to a friend; thinking that he cannot choose to lie
because of his own internal condition he fails to satisfy the consequent ol
(PC). Nadia, on the contrary, retains a sense of an open future to be partially
completed by actions resultant upon choices which she yet takes as contin-
gent given circumstances as she undentands them. Her sense of an ability to
choose consists in her presumptions that (i) her choosing to order any of the
32 flavours is contingent relative to what she then believes; (ii) she does not
believe of any flavour that there are conditions sufficient for her choosing it
or for her not choosing it; and (iii) her choosing is a conscious effort of her
own. Both (i) and (ii) follow directly from (PO), while (iii) is a result of her
having a concept of what a choice (an intention) is. More briefly, Nadia's
sense of the contingency of what she is to choose, together with her
conception of a choice as a conscious effort on her part, is her presumption
of an ability to choose. Coupling this with her belief in the efficacy of her will
and her desires to have the future completed in this way rather than that. we
have all that is needed to give deliberation a "point".

This conclusion remains in force even if Nadia assumes that her choosing

precisely one of the flavours, whatever it might be, is already determined,
viz., even if Nadia is a full-blown determinist. At this point, no doubt, we
arrive at a vivid clash of intuitions, perhaps, to a conflict that can only be
settled by appeal to experimental psychology. But, at this level, the leap to
the unqualified sense of 'can', or absolute contingency, has been premature.
The doxastic characterization of the modality embedded in a dcliberator's
sense of ability cannot be disqualified if it has yet to be articulated and
subjected to proper test.21

Principles (PO) and (POA) can also be used to explain a deliberator's

awareness or feeling of freedom. Of course, this feeling is not an invariant
companion of deliberation; it emerges only when a measure of contemplative
thinking overlays the process of practical reasoning, of deciding what to do,
or how, or when to do it. This does not negate the fact that the feeling is of
something that permeates deliberation all along. Of what? Not the act of
choosing, for this is precisely what terminates awareness of indecision,
contingency and an open future.22  Instead, the agent's focus is now upon his
ability to think and act within a context fixed by his own doxastic and
intentional states; the feeling is the activation of the dispositions ascribed in
(PO) and (POA).

Like its competitor, the doxastic interpretation preserves the indeter-

minacy of the future. More than this; it comprehends a factor that the
Standard Interpretation cannot. A person who deliberates about whether to
eat an apple, an orange or a peach may claim to be conscious that nothing
causally necessitates his choice. Is there not some sense in which he is
epistemically justified, by his experience, in saying that he is free to eat any of
the fruit? An affirmative response not only provides grounds for distinguish-
ing between the experience and the mere rehearsal of belief, it nicely
explains the universality and conviction with which the assumption of free-
dom is held. This is no brute endorsement ofindeterminism; the deliberator
who says he is free as far as he can tell may very well be justified in so doing,
for the contingency is there, detectable within his experiential content. But
there is no reason to suppose that he can similarly be justified -- by his
experience -- in claiming that he is free with respect to all past and present
conditions, even if this latter claim is true. It is the doxastic interpretation
that is on firmer footing here, and the deliberator's choice is determined,
then it is this view which avoids the uncomfortable conclusion that his
experience, not just his belief, is purely illusory.23
IV.  Concluding Remarks

Although the preceding discussion has centered on deliberation, it is likely

that the proposals culminating in (PO) and (POA) have a wider applicability.
For one thing, they seem to pertain to all choice, even that which does not
emerge from conscious deliberation, insofar as decision involves a selection
among presumed alternatives. Perhaps they govern all intention as well; what
is the point of intending something which is not taken as open at some time
before intending it? If so, then each intention is a choice, minimally, between
a course of action and its complement, and we can appreciate anew Kant's
insistence that a presupposition of freedom underlies all practical thought.
Additionally, the proposals imply that an omniscient being cannot deliberate,
choose, or perhaps, intend - a consequence of no small theological impor-
tance if creativity, perfection, or omnipotence necessitate such abilities.24 It
remains to be seen what relevance they have for the overall free will
controversy, though there is every reason to suspect a firm and fruitful

The spectacle of a determinist who deliberates is at first perplexing. What

is  the point of deliberating if whatever  one  chooses  and  does  is  already
determined? What difference can one's own deliberations possibly make?
Faced with such questions, some conclude that we are, by our very nature as
rational agents, indeterminists - an idea which can only disturb the deter-
minist who takes his actions and volitions to be the outcome of antecedent
factors while retaining a passion for consistency. Agreeing that an agent has
a sense of the contingency of his own future, I have urged that the modality
is  indexed  to what he  himself assumes  to  be  the  case;  it need  not  be  a
presumption of the non-existence of any determining conditions whatever.
No more is required to give deliberation a point than the agent's ends, his
belief that those ends will not be realized except through his own intentional
activity, and his sense of freedom based, in part, upon his incomplete grasp
of the future. If forgetfulness, as Nietzsche once wrote, is a precondition of
action, an imperfect conception of what will be is no less essential. Practi-
cally-minded determinists, haunted by the spectres of inconsistency and
fatalism, can be encouraged by this account of the matter.25

1. Immanuel Kant, Fundamental Principles of the Metaphysic of Morals (Bobbs-Menrill, 1949)
p. 65, translation by T, K. Abbott of Grundlegung zur Metaphysik der Sitten, originally published in 1785.

2.This interpretation is strongly suggested by what Kant says elsewhere in the same work, for
example, pp. 63, 69-79, 73-78, and various commentators have urged this reading, e.g., H. G. Paton in the introduction to his translation of the Grundlegung (New York, 1964), pp. 46-48.

3.  Hector-Neri Castaneda, Thinking and Doing (Dordrecht. 1975), pp. 134-5.  Again
on p. 312 of this work, he writes that ". . . to the consciousness of an agent making deliberations: (i) he appears free to choose from alternative courses of action; (ii) his choices appear uncaused. . . ."

wood Cliffs, 1966), pp. 178-182 and in Metaphysics (Englewood  Cliffs, 1974) 2nd edition, pp. 53-55. See also his "Deliberation and Foreknowledge",  American Philosophical Quarterly I (1964), pp. 73-80. Similar views are espoused by Carl Ginet,  "Might We Have No Choice?", in Keith l.ehrer, ed.. Freedom and Determinism (New York,  1966), pp. 87-104; J. M. Boyle, G. Grisez and 0. Tollefsen, Free Choice (Notre Dame 1976);  andJ. W. Lamb, "On A Proof of Incompatibilism", Philosophical Review 86 (1977).

5.  Nicholas Denyer, Time, Action & Necessity: a proof of free will (London, 1981), p, 5, and see
also pp. 39-42 and 65-6. Central to his position is a denial  of true future contingents, so that even if one does A at time t (when this is a result of his choice) it is not true beforehand that he will A at t. Cf. my review of Denyer's book in Nous 18 (1984).      

6. Peter van Inwagen, An Essay on Free Will (Oxford, 1983), p. 160.

7.  I am assuming, thus, that the presumptions are to be taken in what is sometimes called a de
se sense, see David Lewis, "Attitudes De Dido and De Re",  in his Philosophical Papers (Oxford 1981), pp. 133-59. A view that I find congenial is Castaneda's  where the latter occurrences of 'he' in (PE) are quasi-indicators, that is, devices we have for attributing indexical reference to others. See his "He: A Study in the Logic of Self-Consciousness",  Ratio 8 (1966), pp. 130-57; "Indicators and Quasi-indicators", American Philnsophical  Quarterly 4 (1967), pp. 85-100; and
sophical Association 53 (1980), pp. 763-822.  He   has argued in these and other papers that quasi-indicators cannot be replaced by third-person designations  and, thus, that first-person reference is irreducible to third-person reference.

8.  Compare Castaneda's illuminating discussion of the locution 'feels that' in "Philosophical
Method and Direct Awareness of the Self", Grazer Philosophische  Studien 7/8 (1979), pp. 1-58.

9.  The literature on the conditional analysis of freedom is copious. Besides Moore's classic
work, Ethics (Oxford, 1912), ch. 6. echoing long standing  views of John I,ocke, Jonathan Edwards, et. al.  there is also, J. L. Austin,  "Ifs and Cans", Proceedings of the British Academy 42 (1956), pp.!09-132; R. Chisholm, "J. L. Austin's Philosophical  Papers", Mind 73 (1964); K. Lehrer, "An Empirical Disproof of Determinism", in Lehrer, ed..  Freedom and Determinism, K. Lehrer,
"Cans Without Ifs", Analysis 29 (1968), pp. 29-32; D. Davidson, "Freedom to Act", in his Essays on Actions (Oxford, 1980), pp. 63-82; and A. E.  Falk, "Some Modal Confusions conceded that Moore's attempt to construe 'I can' in terms  of 'I shall, if I choose' fails, though it is disputed what this means for the larger questions of determinism,  compatibilism and freedom.

10.  See Wilfred Sellars, "Thought and Action", in Lehrer, ed. Freedom and Determinism, who
mentions not only the threat of a regress that such an analysis engenders but also that it mistakenly construes volitions as actions to be brought  about by yet further acts of will. See, however, Lehrer's treatment of the regress in "Preferences,  Conditionals and Freedom", in van Inwagen, ed., Time and Cause (Dordrecht, 1980), pp. 187-201,  as well as Krister Segerberg's discussion of Lehrer in "Could Have But Did Not", Pacific  Philosophical Quarterly 64 (1983),
pp. 230-41.

11.  Aristotle, Nicomachean Ethics 1140a, and see Richard Sorabji's endorsement of this reading
in Necessity, Cause, and Blame: Perspectives on Aristotle's  Theory (London, 1980), pp. 228 and 245. Compare Denyer, of. cil., pp. 30, 40-2, and R. Burton, "Choice", Philosophy and Phenomenological Research 42 (1982) pp. 581-6. In speaking of P as contingent  I mean, throughout, the conjunction of the possibility of P with the possibility  of not-P.

12. This notion of a relativized modality must be handled with some care to avoid unnecessary
confusion. My preference is to construe the relation of  consequence, employed in the definiens, in a generic sense, not to be restricted to the narrower concept of logical consequence unless otherwise specified. This has a great deal to do with whether the set S includes laws or nomological propositions or laws are principles underlying the consequence  relation.  I refer the reader to my
"On the Concept of Material Consequence", History and Philosophy of Logic 3 (1982), pp. 193-211, for an extended discussion of extra-logical  consequence.  For more about rela- tivized modality, see Hans Reichenbach, Elements of Symbolic  Logic (New York, 1975), p. 396; T. Smiley, "Relative Necessity", Journal of Symbolic Logic  28 (1963), pp. 113-34; and I.L. Humberstone, "Relative Necessity Revisited", Reports on  Mathematical Logic 13 (1981), pp. 33-42. J. W. Lamb, op. cit., and others have used the term 'categorical' instead of 'unqualified' in discussing a deliberator's assumption  of freedom, though with much the same meaning.

13.   See Section IV below.  R. T'aylor in "Deliberation and Foreknowledge" and again in Action
and Purpose, pp. 174-6, has contended that one cannot know, while deliberating, which course of action he will eventually undertake, a claim also endorsed  in C. Ginet, "Can the Will Be Caused?", Philosophical Review 71 (1962), pp.49-55; A.  N. Prior, Papers on Time and Tense (Oxford, 1968), pp. 47-8; A. Goldman,  Theory of Human Action (Englewood Cliffs, 1970), p. 195; and Denyer, op. cit.,  p. 48.  Taylor's  argumentation, in particular, supports the stronger
claim that a deliberator cannot have a belief that he will perform this or that alternative, as he himself realizes in "Deliberation and Foreknowledge",  p. 77.

14.  Peter van Inwagen has suggested that the variable 'S' in a principle like (PC) be restricted
to sets of beliefs the agent takes to be consistent "since  not everyone will be willing to assume that his own beliefs are consistent, and since, presumably,  no proposition is contingent relative to an inconsistent set of beliefs," (in comments at the  American Philosophical Association meetings (Western Division) April 1984). I have two reservations  about accepting this qualifica- tion: (1) While it may be that there is someone who is not willing to assume his own beliefs to be consistent, it docs not follow that he takes them to be  inconsistent. He may simply be in a state of suspending judgment on the consistency of his own beliefs  since few, if any, can rest content with a recognized inconsistency. (2) I do not rule out considerations of relevance in a proper account of logical consequence, thus, am reluctant to accept  the view that every proposition is a consequence of an inconsistent set.

15. Casteneda, Denyer and van Inwagen, has, to my knowledge, seriously  considered or, at least, directly discussed, an alternative explication of the modality involved in a deliberator's sense of an open alternative.

16. See, for example, van Inwagen, op. cit., p. 155. The existence of hypothetical deliberation
suggests that intentions and, thus, courses of action,  can be conditional in form, a point that has long been urged by Castaneda. See, for instance, his Thinking  and Doing, pp. 160 ff., and also his "Reply to Sellars", in Agent, Language, and the Structure  of the World  (New York, 1983), ed., J. E.Tomberlin, pp. 419-33, and compare D. Davidson, op.  cit., pp. 92-4. This underscores the previous assertion about the limited nature of (PO) in  the form given. I might add that, according to the way (PO) is stated, the very item that  is presumed open and deliberated about seems to be the same as that which is said to be contingent  in clause (ii). I do not wish to be terminology, wherein the thing deliberated about is a practition  and the thing viewed as contingent is a proposition. See Castaneda's Thinking and Doing, passim.       

17.  Ginct's advocacy oflhe ignorance condition in "Can The Will Be Caused?", for example,
has spawned a number of critics including J. Canfield,  "Knowing About Future Decisions", Analysis 22 (1962), pp. 127-9; J. W. R. Cox, "Can I Know  Beforehand What I Am Going to Decide?" Philosophical Review 72 (1963), pp. 88-92; and  M. Stocker, "Knowledge, Causation and Decision", Nous 2 (1968), pp. 65-73. Richard La Croix  has also advocated the ignorance condition in "Omniprescience and Divine Determinism", Religious Studies (1976), pp. 365-81,
but Phillip Quinn has argued to the contrary in "Divine  Foreknowledge and Divine Freedom", International Journal for Philosophy of Religion 9 (1978),  pp. 219-40. I have also discussed the issue in "Can God Make Up His Mind?",  15 International  Journal for Philosophy of Religion 15 (1984), pp. 37-47, particularly as it bears on the La Croix/Quinn debate.

18. In this context a distinction must be drawn between formulating a decision, i.e., making up
one's mind, and rehearsing that decision, viz., consciously  affirming an intention already held. Castaneda's work on intentions is especially relevant here,  see Thinking and Doing, pp. 275-8.

19.  Ginet, "Can The Will Be Caused?", p. 51. Brian O'Shaughnessy in The Will: A Dual Aspect
Theory (Cambridge, 1980), p. 297, also endorses a claim  of this sort saying that a necessary condition of decision is that it resolves a state of uncertainty  about what to do.

20. I take it as obvious that the phrase 'whether or not' occurring in (PI), and also in (PU)
below, indicates a conjunction of denials of belief falling  within the scope of 'presumes'.

21.  It might be charged that I have dealt unfairly with the Standard Interpretation inasmuch as
its adherents do offer arguments in its favour. Arguments  yes, but arguments whose inadequacy is due to a failure to explore alternative explications  of 'possible' and 'determined'. In "Might We Have No Choice?" Ginet, for instance, has argued in the following manner: a deliberator must believe that his eventual choice is effeclive, i.e.,  that he "has" a choice (pp. 92-93); determinism, however, entails that our choices are always  ineffective (pp. 90, 93): hence, anyone who believes in determinism either cannot choose at all  or else is aware that he is constantly
deluded - an "implausible" if not "impossible" condition  to be in (pp. 93, 104). What is the meaning of 'effective' and 'has a choice' in this argument?  From his examples of the prisoner and the child in an amusement park, one's choice to A is  ineffective if one's choice is not an essential factor in determining whether one will A. But  a determinist need not believe his choice is ineffective in thif sense; if he satisfies the efficacy  assumption then he obviously thinks that his intentions are essential components in causal chains leading  up to his actions, and this is compatible with the belief that his intentions and actions  are already determined. On the other
hand, if an effective choice is, by definition, an undtternincd  choice, as the discussion on pp. 90-92 suggests, then it is the initial premise of Ginet's  argument that demands further defence. In either case, the argument as it stands poses  no obstacles to the proposed theory. On the contrary, since the position to which the deliberating  determinist is forced by the Standard Interpretation is indeed implausible, it is a virtue of  our proposals that they rescue the determinist from this doxastic quagmire.

22.  Here I go against the suggestion offered by Douglas Browning in "The Feeling of
Freedom", Review of Metaphysics 18 (1964), pp. 123-46,  who writes: "The long sought feeling of freedom is no other than the experience of the act of choice itself as it is performed, as it must he performed, within the practical stance" (p. 145). Compare,  Boyle, Griscz and Tollefsen, op. cit., pp.  18-20.

23.  Boyle, Grisez and Tollefsen, ap. eil., pp. 20-3, are also careful to distinguish the experience
of freedom from the judgment that one is free. That an awareness of freedom would be virtually impossible if freedom is analyzed in terms of unqualified  contingency has been emphasized by J. W. Corman and K. Lehrer in Philosophical Problems and  Arguments, (New York, 1968), pp. 131-47.

24.  I refer the reader to the papers by La Croix, Quinn, and myself listed in note 17 above.

25.  I am indebted to Hector-Neri Castaneda for the valuable comments and criticisms
provided during the development of this paper, to J. Christopher  Maloney who years first kindled my interest in deliberation, and to Robert Audi,  Robert Good, Hugh Harcourt, Steven Lee, Al Mele, Ron Miller, George Nakhniliian, Mark Pastin,  Lynn Stephens, Eric Stiffler, and Leslie Stevenson for their helpful comments on earlier versions.


This article comes from The Philosophical Quarterly Vol. 36 No. 14 (1986), 230-251

to Det & Free Website front page

BACK to T.H. front page