Canadian Journal of Philosophy

The Principle of Alternate Possibilities
Author(s): Phillip Gosselin
Source: Canadian Journal of Philosophy, Vol. 17, No. 1 (Mar., 1987), pp. 91-104
Published by: Canadian Journal of Philosophy
Stable URL: http://www.jstor.org/stable/40231515 .
Accessed: 18/01/2014 13:47

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

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

 .

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

http://www.jstor.org 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

CANADIAN  JOURNAL OF  PHILOSOPHY 
Volume  17,  Number  1,  March  1987,  pp.  91-104 

91 

The Principle of Alternate Possibilities 

PHILLIP GOSSELIN 
Brandon University 
Brandon, Manitoba 
Canada R7A 6A9 

The standard argument for the incompatibility of determinism and 
moral responsibility employs the following two premises: 

1.  A person is morally responsible for what he has done only if 

he could have done otherwise: 

2.  A person could have done otherwise only if his action was not 

causally determined. 

While premise two has been the focus of an enormous amount of con- 
troversy, premise one until recently has remained virtually unchalleng- 
ed.  However,  since  Harry Frankfurt's provocative paper in  1969, 
premise one, which he dubbed the principle of alternate possibilities 
(henceforth referred to as PAP), has begun to attract its share of the 
debate.1 Frankfurt argued that PAP is false and that its falsity under- 
mines the position of those who  assert the incompatibility of deter- 
minism and moral responsibility.2 Two previous papers I wrote were 

1  Harry  Frankfurt,  'Alternate  Possibilities  and  Moral  Responsibility,'  Journal of 
Philosophy 66 (1969) 829-39. The literature which  discusses  Frankfurt's attack on 
PAP  includes  the  following:  David  Blumenfeld, 
'The  Principle  of  Alternate 
Possibilities,'  journal of Philosophy 68 (1971) 339-45; Peter  Van  Inwagen,  'Ability 
and Responsibility,' Philosophical Review 87 2(1978) 201-24; Phillip Gosselin,  Is  There 
a Freedom  Requirement  for Moral Responsibility?'  Dialogue 18, 3 (1979) 289-306; 
Mark Bernstein, 'Moral Responsibility and Free Will,' Southern Journal of Philosophy 
19 (1981) 1-10; Phillip  Gosselin,  'Moral Responsibility  and  the  Possibility  of Do- 
ing Otherwise,'  Philosophy Research Archives 8 (1982) 499-512. There are also brief 
discussions  of Frankfurt's position  in the following:  Donald  Davidson,  'Freedom 
to Act,' in Ted Honderich,  ed.,  Essays on Freedom of Action (Boston, MA: Routledge 
and  Kegan  Paul 1973) 137-56; Robert Audi,  'Moral Responsibility,  Freedom  and 
Compulsion,'  American Philosophical Quarterly 11,  1(1974) 1-14. 

2  In this  paper  incompatibilism will  be  defined  as the  view  that determinism  is  in- 

compatible  with  moral  responsibility. 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

92 

Phillip Gosselin 

devoted in part to showing that Frankfurt's argument is ineffective; 
one of those papers also argued that, while PAP is indeed false as it 
stands (though for reasons entirely different from those advanced by 
Frankfurt), if it is appropriately supplemented, it can continue to serve 
its traditional role in the determinism-responsibility debate.3 

The purpose of this paper is (a) to offer a clarification and refine- 
ment of my supplemented version of PAP (let us call it S-PAP), and 
(b) to reinforce and supplement my earlier defence of S-PAP. However, 
before turning to that clarification and defence, let us consider briefly 
why it is necessary to introduce S-PAP as a surrogate for PAP in the 
incompatibilist's argument. 

I Replacing PAP with  S-PAP: Culpable Inability Does  Not Excuse 

PAP as it stands is false. A drunken driver who can not avoid striking 
a pedestrian may well be morally responsible for the accident. Because 
of his reduced reaction time it was impossible for him to avoid strik- 
ing the man; however, if he had drunk to excess when it was within 
his power to do otherwise, knowing he would have to drive home 
shortly, knowing that he would be dangerously incapacitated as a result 
of his drinking, etc., then he is (to some extent) morally responsible 
for the pedestrian's injury or death. Consideration of this kind of ex- 
ample shows that, stated roughly, even if one is unable to do other 
than a, if one is morally responsible for that inability, it will not serve 
to excuse. In short, culpable inability will not serve to excuse. Accor- 
dingly, PAP must be supplemented as follows: 

S-PAP: a person is morally responsible for what he has done on- 
ly if (1) he could have done otherwise or (2) he is morally respon- 
sible for the fact that he could not. 

Again, stated roughly, moral responsibility for the action in question 
requires ability (to do otherwise) or culpable inability (to do otherwise); 
that is,  nonculpable inability excuses completely.4 

3  Gosselin,  Is There a Freedom Requirement for Moral Responsibility?' and Gosselin, 
'Moral Responsibility  and the Possibility  of Doing  Otherwise/  The second  paper 
includes  some  discussion  of Blumenfeld,  Van Inwagen  and  Bernstein  as well  as 
Frankfurt. 

4  For a fuller statement  of this argument  see  'Freedom Requirement/  301-4. Notice 
that PAP and S-PAP are concerned  with  moral responsibility for an action or failure 
to act. As  I argue  in  'Freedom  Requirement/  responsibility  for action  or failure 
to act is basic in any  determination  of  moral responsibility.  One  can be  morally 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

The Principle of Alternate Possibilities  93 

It is easy to see that S-PAP can take PAP's place in the standard argu- 
ment for incompatibilism. It is true that S-PAP does not require that 
the agent could have done otherwise at the time of the act for which 
he is being censured, but it does require that the agent either could 
have done otherwise at the time or is to blame for the fact that he could 
not. And his inability will be blamable only if it is the product of an 
earlier act which is blamable; that act in turn will be blamable only if 
the agent either could have done otherwise or is at fault for the fact 
that he could not. And so on, until we come at last to an act of which 
it is true to say 'The agent could have done otherwise.' In other words, 
if one  is to be morally responsible for a there must have been the 
possibility of alternate action either at the time of a or at some place 
along the line of a's causal ancestors -  and the incompatibilist, of 
course, argues that if determinism is true such alternate possibilities 
of action never obtain. S-PAP, then, can function as PAP's surrogate 
in the standard argument for incompatibilism. However, S-PAP re- 
quires some classification and refinement. 

II Refining S-PAP: The Ambiguity of Could 

S-PAP  requires that one could have done otherwise (with the qualifica- 
tion indicated), but could have done otherwise has two very different 
senses (both of which, incidentally, can be accepted by compatibilists 
and incompatibilists alike). Those senses can be set out best if we begin 
by considering the statement P could not have done otherwise. Compare 
the following statements: 1. 1 can't lift the car off him; it's just too heavy! 
2. You can't go near that (burning) vehicle; it may explode at any mo- 
ment! The first statement asserts that for me under the circumstances 
it is categorically impossible to lift the car. I simply lack the strength 
necessary to accomplish the feat; no amount of trying will do the trick. 
The second  statement asserts something  quite  different: it  is  not 
categorically impossible to approach the burning vehicle, but it is foolish 
to do so; prudence forbids such action. The impossibility asserted by 
the can't is, in this case, a conditional impossibility; one cannot approach 
the car if one is to avoid serious risk to life and limb. 

The latter kind of can't may be termed 'normative' for it asserts that 
the agent cannot perform a certain action if he is to act in accordance with 

responsible  for something  other  than  an action  or failure to act, but only  if that 
something  is the  product  of one's  earlier action  or failure to act. Thus,  you  may 
be  morally  responsible  for a state  of  affairs such  as  a death,  or a character trait 
such  as your child's truculence,  but only  if the death  or the truculence  is the pro- 
duct  of  your  own  earlier action  or inaction. 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

94  Phillip Gosselin 

a certain norm: it says that a certain norm forbids a certain kind of ac- 
tion. The norm in question may be any of several. It may be a norm 
of prudence as in the preceding example. Or it may be a moral norm 
as in 'I can't play badminton tonight: I promised my wife we would 
go out to dinner/ It is not categorically impossible for me to play bad- 
minton tonight; but it is impossible if I am to keep my promise. The 
norm in question may be aesthetic as in 'You can't plant a tree there; 
it would destroy the symmetry of your landscape.' It may be a legal 
or a religious norm; it may be a norm of custom or of fashion. But in 
any of these cases the impossibility expressed is merely conditional; 
it says that you cannot act in a certain way if you are to satisfy a cer- 
tain (usually implicit) norm. 

The other sense of can and cannot I shall call the 'power sense/ It was 
not within my power to lift the car with my bare hands. It is not that 
some norm forbade that I lift it (quite the contrary); rather, I simply 
lacked the physical ability to do it. The lack of some intellectual abili- 
ty, likewise, may render an act quite impossible: I cannot, for exam- 
ple, memorize the entire Oxford English Dictionary. Lack of know-how, 
also, will render certain actions impossible: someone unacquainted with 
the inner workings of an automobile cannot (power sense) do a valve 
job. Finally, circumstances of time or place may render an act impossi- 
ble: I cannot while in my office help my wife prepare lunch at home; 
I do not have the opportunity to help her. It seems perfectly clear that 
the  lack of  opportunity, the  lack of  know-how  or the  lack of  the 
necessary  physical  or  mental  abilities will  render certain actions 
categorically impossible. It is less clear, but perhaps true, that the lack 
of what might be called 'volitional ability' also renders certain actions 
categorically impossible. 
'Volitional ability' refers  to  the  ability to  bring  oneself  to  do 
something, i.e., the ability to perform an act that is 'volitionally effort- 
ful.' By a 'volitionally effortful' act, I simply mean an act that is oppos- 
ed by a desire/fear/aversion that is keenly felt at the moment of action. 
What is involved, typically, is the sacrifice of some pleasure or the 
tolerance of  some  unpleasant experience that is  both  intense  and 
relatively immediate for the  sake of  a less  immediate good  (or the 
avoidance of a less immediate ill). The desire/fear/aversion in ques- 
tion may be pathological as in phobias, compulsions and addictions; 
but, it need not be, of course: rational fears and normal desires, if suf- 
ficiently acute at the moment of choice, may also result in volitionally 
effortful action -  what we sometimes refer to as 'action requiring an 
effort of will.' Now, if the desire/fear/ aversion in question were so in- 
tense that it was not only difficult but impossible for the agent to act 
contrary to it, then the agent involved would lack the volitional ability 
to perform the action in question. 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

The Principle of Alternate Possibilities  95 

Whether this is ever the case is an interesting question -  partly em- 
pirical and partly philosophical. It is partly philosophical because a 
careful analysis of the concept of someone being unable to bring himself 
to do something is necessary in order to know what conditions would 
make it true. I shall not offer such an analysis here; however, I will 
state a necessary condition for volitional inability which will show that 
it is probably extremely rare. It must surely be conceded that no mat- 
ter how difficult it is for a person to bring himself to perform a certain 
kind of act, it is not impossible for him to do so if there is some incen- 
tive  that,  if  operative,  would  induce  him,  constituted  as  he  is, 
deliberately to perform the act in question. If, constituted as he is (with 
fears, aversions, etc. intact), he would deliberately do it for reason r 
then he can do it -  even though reason r is not operative in this situa- 
tion. Consider, for example, a claustrophobic. There may in fact be 
no reason for him to enter a certain narrow tunnel, or there may be 
a reason -  say the retrieval of his son's baseball -  but none sufficient 
to  move him; nonetheless,  while  it would  be volitionally difficult, 
perhaps excruciatingly difficult, for him to enter the tunnel, it is not 
volitionally impossible provided we can say that he would do it if he 
were given such and such a reason -  say the saving of his son's life.5 
Thus, while there may exist cases of genuine volitional impossibility, 
such cases are probably very uncommon. 

Notice that there is a use of the phrase P could not bring himself to 
do x which seems to conflict with the foregoing. If I say that I could 
not bring myself to pass up a second piece of fresh apple pie, normal- 
ly I do not mean that it was categorically impossible for me to do so; 
I mean something like 'I tried to pass it up and I failed.'6 However, 
it is important to note that this loose and rather misleading sense of 

5  Notice,  incidentally,  that the difficulty of doing  something  is sometimes  sufficient 
to excuse  one's  failure to do it -  impossibility  is not required.  Whether  it is suffi- 
cient  in any  given  case  will  depend  on  the  gravity  of  the  failing  in question.  To 
put  it another  way,  the  greater the  obligation  one  has  to  meet,  the  greater the 
volitional  obstacles  one  can  be  expected  to  overcome  in  the  effort  to  meet  it. 

6  Some  may be inclined  to question  the meaningfulness  of the phrase  trying to pass 
up a piece of pie. How  does  one  go  about  'trying' to do  such  a thing?  The answer 
is that one  goes  about  it in  essentially  the  same  way  one  tries to  overcome  any 
temptation: by attempting to put the forbidden object out of one's mind, by rehear- 
sing the good  reasons  for turning away from it, by reflecting on the consequences 
of  not  doing  so,  etc.  One  must  not  suppose  that all trying  involves  physical  ac- 
tivity.  Of course,  a statement  such  as 'I couldn't  bring  myself  to  refuse  the  pie' 
may  well  exaggerate  the  actual extent  of  the  agent's  effort  to  resist  temptation. 
In fact, there might well be no effort at all -  but, if there was  capitulation without 
any resistance,  strictly speaking,  it is inaccurate to talk of failing to luring oneself' 
to  do  it. 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

96 

Phillip Gosselin 

could not is not the sense that absolves from moral responsibility; and, 
accordingly, it is not the sense that concerns us in our discussion of 
PAP and S-PAP. 

Thus to say that one could not have done otherwise in the power 
sense (hereafter referred to as could not have done otherwise [p]) is to say 
that it was categorically impossible to do otherwise due to the lack of 
opportunity, the lack of know-how, the lack of some physical or in- 
tellectual ability or (perhaps, in rare cases) the lack of volitional abili- 
ty. On the other hand, to say that one could not have done otherwise 
in the normative sense (hereafter referred to as could not have done other- 
wise [n]) is not to assert a categorical impossibility at all; rather it is 
to say that one could not do otherwise if one were to do the right thing 
or the prudent thing, etc. That is, in the normative sense / could not 
do x is always  elliptical for / could not do x if I were todoy  -  which was 
required by norm n. 

This distinction explains the hesitation that is sometimes experienc- 
ed in describing a victim of coercion: could he have done otherwise 
or not? There is a tendency to say Tes'  and a tendency to say 'No.' 
The truth is that in many, probably most, cases of coercion, while the 
agent cannot do otherwise (n), he can do otherwise (p). Typically cases 
of coercion (or alleged coercion) involve some kind of threat, explicit 
or otherwise. However, unless the threat excites a fear so terrible that 
nothing would induce the coercee to resist it, we must say that he could 
resist it in the power sense -  though he chooses not to. On the other 
hand, the cost of resistance may be sufficiently great that compliance 
is justified. In such a case we will say that he cannot do otherwise in 
the normative sense  -  meaning that resistance would be foolish. 
If, then, there are two very different senses of he could have done other- 
wise, which sense does moral responsibility require? To put it another 
way, which sense of he could not have done otherwise serves to exculpate? 
Consider first the normative sense. In saying that the agent could not 
have done otherwise one is saying that some norm required that he 
do what he did. But, of -course, whether that will serve to exculpate 
will depend on what the norm is. If morality required that he do what 
he did, then of course the agent is not to blame since he did nothing 
wrong. What we might call the normative could not of morality necessarily 
exculpates. Of course, some normative could nots usually would not 
be appealed to in the context of moral assessment -  for example, those 
of fashion or aesthetics. However, there is one other normative could 
not that frequently is appealed to when determining moral responsibili- 
ty and that is the normative could not of prudence. (He  could not do other- 
wise if he was to avoid such and such very serious consequences to 
himself.') Often it does exculpate, but not necessarily so: prudence may 
dictate what morality forbids. To have done otherwise might have 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

The Principle of Alternate Possibilities 

97 

resulted in some harm/cost to the agent, but morality might have re- 
quired that he make the sacrifice; it would depend on the size of the 
sacrifice as well as the moral gravity of the act being contemplated. 
Accordingly, the question of whether the normative could not of prudence 
exculpates depends on the details of the case. 

In any case, when one discusses the question of whether a subject 
of moral assessment could have done otherwise (n) one is concerned 
with  the  question of  whether his  action was  morally justified. Did 
morality require that he do what he did? Did self-interest require it 
-  and, if so, was the cost of alternative action great enough to render 
alternative action supererogatory? That is, was his action morally per- 
missible? 

The inability to do otherwise (p), on the other hand, is concerned 
not with justification but with excuse. If someone asks why P failed 
to help Q in his moment of distress and the reply comes that some 
physical deficiency rendered it impossible for P to do so,  the reply 
serves not to justify P's inaction but to excuse it; the reply does not 
say that inaction was morally permissible in this situation, but rather 
that it is excusable in P's case. Furthermore the common sense convic- 
tion that nonculpable inability excuses completely seems perfectly ac- 
ceptable: it does not seem right to hold someone morally responsible 
for doing something if through no fault of his own it was categorically 
impossible for him to do anything else. Our moral intuitions are sure- 
ly very strong on this point and if the Frankfurt counterexamples to 
it, which we shall consider shortly, are to be effective, they must sure- 
ly be unequivocal. 

One final point before turning to some objections to S-PAP. Both 
senses of could have done otherwise seem to be relevant in assessments 
of  moral  responsibility,  but  are  both  senses  operative  in  the 
determinism-responsibility debate? That is, is the could have done other- 
wise that appears in the standard argument for the incompatibility of 
determinism and moral responsibility (see the first paragraph of this 
paper) the normative could or the power could? Clearly it is the power 
could and only the power could. If determinism is true, says the incom- 
patibilist, no one could ever have done other than he did and hence 
no one is ever morally responsible for what he did. Now if could were 
here taken in the normative sense, the incompatibilist would be say- 
ing that, if determinism is true, then no one could ever have done other- 
wise  if he was to satisfy certain norms. That is,  the incompatibilist 
would be saying that, if determinism is true, then it is true of absolutely 
every act that some norm or other required that that act be performed 
precisely as it was.  But that is absurd. 
To begin with, it is perfectly clear that many actions we perform are 
not required by any norms -  whether determinism is true or not. But 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

98 

Phillip Gosselin 

in any case how would the truth of determinism imply anything about 
what certain norms required or forbade? Determinism would, of course, 
imply that what norms are operative among us were completely deter- 
mined to be what they are; but it would not imply that those norms 
required this conduct rather than that -  for example, that they required 
the conduct that actually obtained.7 
It is the power sense of could, then, that is operative (or should be 
operative) in discussions of determinism and moral responsibility. If 
an action is determined, then under the totality of conditions that ob- 
tained at the time, it was categorically impossible for the agent to do 
anything other than just what he did -  or, at least, so says the incom- 
patibilist. Thus, while both the normative and the power coulds are 
relevant in moral assessments, only the power could is operative in the 
determinism-responsibility debate. Failure to recognize this point has 
led  to  considerable confusion.  For example, in  the  context of  the 
determinism-responsibility question cases of coercion have frequent- 
ly been taken as paradigm cases of the inability to do otherwise -  yet 
the inability arising out of coercion is almost invariably a normative 
inability, hence completely different from the inability supposedly aris- 
ing from determinism. An example of how failure to appreciate this 
difference has led to bad arguments will be examined below. 

Let us turn now from a clarification of PAP to its defence. We shall 
examine several arguments that have been brought against it recently 
-  and which with minor modifications might be brought against S- 
PAP as well. 

Ill  Defending  S-PAP: Frankfurt's Three Objections 

Harry Frankfurt has presented three different kinds of counterexam- 
ple  to  PAP each of  which  requires a  separate and quite different 
defense. Two of those counterexamples were, I think, effectively crippl- 
ed  in  a previous paper and I shall provide merely a recap of  the 
arguments with some minor modifications. My earlier reply to the third 
counterexample was, I think, less effective than it might have been, 
and I shall offer a substantially different critique based on the distinc- 
tion set out in the previous part of this paper. 

In what  I  shall call the  Evil Overseer counterexample Frankfurt 
describes a situation in which Jones commits some misdeed, say an 

7  Indeed,  some  argue that the truth of determinism  would  render all norms mean- 
ingless  (since  ought  implies  can  [p]  and  determinism  is  incompatible  with  can 
[p]). It is not, of course, necessary for me to rest my case on this controversial point. 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

The Principle of Alternate Possibilities 

99 

embezzlement, unaware that, if he had reneged on his  decision to 
embezzle, Black would have intervened to ensure that he do it (by coer- 
cing Jones or by somehow inducing in him an irrestible urge to do it, 
or by whatever means you please guaranteeing the performance to the 
act in question). In such a case, says Frankfurt, Jones having done the 
deed for his own reasons is morally responsible for it; yet because of 
Black he was bound to perform it, he could not have not performed it. 
Two comments are in order. To begin with, the crucial question is 
whether it is true that Jones could not have done otherwise at the time 
he performed the act (t): the question is not whether he was bound to 
perform it sooner or later. When we bear this in mind, we  see that 
Frankfurt's counterexample, as stated, is ineffective. Jones could have 
changed his mind -  even at the last moment. It is true that the Evil 
Overseer would then have intervened and Jones eventually would have 
embezzled after all; but in that case Jones would not have been moral- 
ly responsible for the embezzlement. Now,  the counterexample can 
be modified to meet this problem, but it will fail on another count. 
Suppose Black is an evil genuis who is able to forsee what Jones will 
do and who would have intervened if necessary to guarantee that Jones 
perform the embezzlement at t, i.e.  at the time he actually performs 
it (for his own reasons). In that case it may seem that Jones is both 
morally responsible for the embezzlement at t and is bound to embez- 
zle at t . However, the problem that still faces Frankfurt is that the sense 
in which Jones is 'bound' is a problematic one. What 'binds' Jones is 
neither coercion nor compulsion nor any factor that compatibilists have 
traditionally admitted as rendering one powerless to do otherwise. 
What 'binds' Jones is simply the fact that the evil genius's readiness 
to intervene guaranteed that Jones would embezzle at t, that is, the 
fact that there were conditions antecedent to the embezzlement that 
guaranteed its occurrence at t . But if determinism is true there are con- 
ditions antecedent to every act that guarantee its occurrence in a similar 
sense  -  yet it is notoriously controversial whether such a guarantee 
renders the agent powerless to do otherwise. 
The point here is that Jones is no more powerless to do otherwise 
in virtue of the evil genius's resolve to see him embezzle than I would 
be powerless to do other than write this paper if determinism were 
true. Whether I would be is a controversial question. Furthermore, it 
would be no less controversial a question to ask whether I was or was 
not able to do otherwise in a universe determined and prearranged by God. 
The fact that my act was determined by God would not make me any 
less able to do otherwise than if it was determined by impersonal forces. 
Compare this now to Jones' case: his embezzlement was not actually 
prearranged by anyone -  though it would have been if the evil genius 
had seen the necessity of doing some arranging. Is Jones any more 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

100 

Phillip Gosselin 

txnind', is his possibility of alternative action any less than mine would 
be now if this universe were determined by God? Surely not. Accor- 
dingly, if the possibility of alternative action is controversial in the 
universe-determined-by God case, it must be at least as controversial 
in the evil-genius case. But Frankfurt needs a noncontroversial case; 
he needs a case where the impossibility of alternative action is clear. 
I conclude, then, that even the evil genius version of the Evil Overseer 
counterexample fails.8 

The second Frankfurt counterexample I shall consider is that of the 
Willing Addict. In this argument we are to imagine a case of addiction 
so extreme that it is not merely difficult but impossible for the agent 
to refrain from taking the drug to which he is addicted when the op- 
portunity to do so presents itself. (The counterexample can be made 
to present a challenge to S-PAP as well as to PAP by supposing that 
the agent has become addicted through no fault of his own.) However, 
the addict does not perceive his condition as a burden; indeed, he is 
positively delighted with it. If somehow magically he could remove 
the addiction, he would not do so. Such a one is morally responsible, 
says Frankfurt, in spite of the fact that he is unable to do otherwise. 
Such a one may indeed be morally responsible -  but for what? Is 
it quite clear he is to blame for an act that at the time he could not 
avoid performing? Through no fault of his own it is impossible for him 
to refrain from taking the drug, but is it clear that he is morally respon- 
sible for that? I think not. He is a legitimate object of contempt, but 
for what? For his attitude toward his addiction. But he could adopt a dif- 
ferent attitude: he could try to quit, he could take steps to remove the 
addiction, but he does not. Perhaps he cannot even do that, you say. 
Furthermore, perhaps his being the kind of person who would revel 
in such a condition is in no way the result of any previous voluntary 
choices on his part; it is a condition thrust upon him like a disease. 
A desert island case surely -  but in such a case our attitude must 
change; it must no longer be one of moral condemnation though it 
may remain one of distaste. Again, I submit, Frankfurt's counterex- 
ample fails.9 

8  Frankfurt's statement  of the Evil Overseer  counterexample  appears in 'Alternate 
Possibilities/  835-6.  A  fuller  statement  of  my  argument  against  this  counterex- 
ample  is to be found  in 'Freedom Requirement/  290-2. 1 also consider  a modified 
version  of the Evil Overseer  argument  (advanced  by David Blumenfeld)  in 'Moral 
Responsibility  and  the  Possibility  of  Doing  Otherwise/  500-3. 

9  The  case  of  the  Willing  Addict  was  first set  out  in  Frankfurt's 'Freedom  of  the 
Will and the Concept of a Person/  Journal of Philosophy 68, 1 (1971) 18-20. My argu- 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

The Principle of Alternate Possibilities 

101 

Frankfurt describes his third counterexample, which I shall call the 

case of the Willing Coercee, as follows: 

Jones decides for reasons of his own to do something, then someone threatens 
him with a very harsh penalty (so harsh that any reasonable person would sub- 
mit to the threat) unless he does precisely that, and Jones does it. ...  [Suppose 
now that] Jones was neither stampeded by the threat nor indifferent to it. The 
threat impressed him, as it would impress any reasonable man, and he would 
have submitted to it wholeheartedly if he had not already made a decision that 
coincided with the one demanded of him. In fact, however, he performed the 
action in question on the basis of the decision he had made before the threat was 
issued. When he acted, he was not actually motivated by the threat but solely 
by the considerations that had originally commended the action to him. It was 
not the threat that led him to act, though it would have done so if he had not 
already provided himself with a sufficient motive for performing the act in ques- 
tion. ... Then I think we would be justified in regarding his moral responsibility 
for what he did as unaffected by the threat, even though, since he would in any 
case have submitted to the threat, he could not have avoided doing what he did.10 

At the time that this argument was advanced Frankfurt was vague- 
ly aware of the ambiguity of the phrase could have done otherwise and 
suspected that this ambiguity might be used to launch an objection 
to his counterexample; however, he did not attempt to spell out the 
ambiguity, nor did he attempt to determine whether it could be used 
to undermine his case.11 1 have set out the two senses of could in the 
second part of this paper and I have shown that it is the power sense 
that is operative in the determinism-responsibility debate. It remains 
only to point out that when we say the Jones of this counterexample 
could not have done otherwise we  are using the phrase in the nor- 
mative sense: Jones was a victim of coercion; in face of such a harsh 
threat it would have been foolish to do otherwise, but there is no reason 
to suppose it was categorically impossible for him to do otherwise (the 
way it would have been had he lacked, for example, the physical ability 
to do so.) Accordingly, as stated, the counterexample has no relevance 
to  the version of PAP (or S-PAP) that figures in the  determinism- 
responsibility debate. 

In order to make the counterexample relevant to that debate we must 
suppose that the threat has the effect, not only of rendering alternative 
action foolhardy, but of rendering it categorically impossible. Thus, 
we must suppose (contrary to what Frankfurt supposes in his version 

ment against this counterexample is stated more fully in 'Moral Responsibility 
and the Possibility of Doing Otherwise/ 507-10. 

10 Frankfurt, 'Alternate Possibilities/ 831-2 
11 Frankfurt, 'Alternate Possibilities/ 834 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

102 

Phillip Gosselin 

of this counterexample) that the threat has induced a fear so terrible 
that at the moment of action it renders defiance not merely volitional- 
ly difficult but impossible. It is now true that the agent could not have 
done otherwise in the relevant sense,  but is it quite clear that he is 
morally responsible  for  the  act? In  the  original counterexample 
Frankfurt was  at  pains  to  point  out  that Jones 'was  not  actually 
motivated by  the  threat but  solely by  the  considerations that had 
originally commended the action to him'; that, Frankfurt believes, made 
Jones morally responsible for what he did. Two points can be made 
here. First, it is extremely doubtful that the modified counterexample 
will allow us to say that Jones was not motivated by the threat: he was, 
at the moment of action, possessed of an insurmountable fear of the 
consequences of doing otherwise. But, in any case, given that this fear 
stood as an insuperable obstacle to alternative action, I submit that it 
is not at all clear that Jones is morally responsible for what he did. He 
may be contemptible for his earlier decision to proceed, or for being 
the kind of person who would have done it even without the threat, 
but it is far from clear that he is morally responsible for the act itself. 
In other words we  have ample ground for heaping scorn on Jones 
without having to reject PAP or S-PAP. Notice that Jones' situation 
in this modified version of the counterexample is precisely the same - 
in all relevant respects -  as that of the willing addict. And as in the 
case of the willing addict we must bear firmly in mind that ex hypothesi 
it was just as impossible for him to do otherwise as it is impossible 
for me to lift my house with my bare hands. 

As was pointed out earlier, S-PAP is a principle that is very firmly 
entrenched in our moral thinking; if Frankfurt is to prove it false, his 
counterexamples must be unequivocal. I submit that they are not. 

IV Michael Zimmerman's Discussion  of  PAP 

In conclusion I wish  to comment on a recent discussion of PAP by 
Michael Zimmerman.12 Zimmerman describes a counterexample to PAP 
similar in its essentials to Frankfurt's Evil Overseer argument and raises 
the question of its effect on the determinism-responsibility debate. He 
then attempts not to resolve the issue but to clarify it by describing 
in some detail what he believes are the five main positions that might 
be taken on it. The emergence of so many alternative positions results 
mainly from Zimmerman's contention that PAP requires the  two 
following supporting premises: 

12  Michael Zimmerman,  'Moral Responsibility,  Freedom, and Alternate Possibilities/ 

Pacific Philosophical Quarterly 63  (1982) 243-54 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

The Principle of Alternate Possibilities 

103 

SP1: A person is morally responsible for what he has done only 
if he did it freely. 

SP2: A person has freely done what he has done only if he could 
have done otherwise.13 

Zimmerman explains the need for these two supporting premises as 
follows: 

The dispute  between  compatibilists  and  incompatibilists  is primarily concerned 
with the relation between  freedom of human  action and determinism.  Why,  then, 
has  [PAP] played  such  a dominant  role  in  this  dispute,  given  that  it mentions 
neither freedom  nor determinism  but appears to concern  itself,  instead,  with  the 
matter of moral responsibility? The answer  is four-fold.  First, the debate between 
compatibilists  and incompatibilists  is fueled  by a common  concern  with  accoun- 
ting  for the  possibility  of  persons  being  morally  responsible  for certain  of  their 
actions  ....  Second,  both  parties  to the  debate  assume,  roughly,  that a person's 
being  morally responsible  for an action requires that he acted freely.  Third, both 
parties assume,  roughly,  that, to act freely, a person  must be able to do something 
other  than  that which  he  in fact does.  And  fourth  ...  the  incompatibilist  asserts 
[and the compatibilist denies],  roughly  that a person  can do something  other than 
that  which  he  in  fact does  only  if determinism  is  false.14 

Now, if Zimmerman is right about this, not only is the debate made 
more complex, but PAP is rendered more vulnerable. PAP (and hence 
S-PAP) could be attacked via an assault on either SP1 or SP2 -  and 
there is reason to suspect the soundness of both these principles. 
SP1 says that moral responsibility requires that the agent acted free- 
ly. But consider an alcoholic whose incessant drinking wreaks havoc 
on his family. He may well be morally responsible for continuing to 
drink, but does he act freely in doing so? Some will say 'No: he is a 
slave to his passions/  Others will say 'Yes, for it is possible for him 
to quit, however difficult.' I am not sure a clear answer to this ques- 
tion is possible without stipulating a definition of acted freely. It seems 
that our linguistic conventions do not permit an unequivocal answer 
here. 
SP2 says that acting freely requires being able to do otherwise. Now 
it seems fairly clear that acting freely does not require that one could 
have done otherwise in the normative sense; and it is questionable 
whether it requires that one could have done otherwise in the power 
sense.  Suppose a person breaks an inconsequential engagement in 

13  Zimmerman,  244 

14  Ibid. 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

104  Phillip Gosselin 

order to help a friend in need. He could not do otherwise in the nor- 
mative sense  for morality required that he  act as he  did.  But if he 
wanted to help his friend, if he was glad to do so, we may say that 
he acted freely, that he did it of his own free will. It is even questionable 
whether one who is unable to do otherwise in the power sense necessari- 
ly acts unfreely. Does Frankfurt's willing addict act unfreely in taking 
the drug? Perhaps most addicts do. But this fellow is an unusual case: 
it is impossible for him to resist his craving, but he is quite happy to 
have it. In taking the drug he is doing what he wants to do. Is an unam- 
biguous answer possible here?15 

Now  it may well be that there is a plausible sense of free action in 
which both SP1 and SP2 (an SP2 using the power sense of could) are 
true; but does the acceptability of PAP (and S-PAP) depend on these 
problematic premises? Zimmerman assumes that it does; he provides 
no support whatsoever for the claim that it does. Yet this claim requires 
support -  for, prima facie, S-PAP stands on its own. Essentially it says 
that one is not to blame for doing something (or failing to do something) 
if through no fault of one's own it was impossible to do anything else; 
it says that it is wrong, not fitting from the moral point of view,  to 
fault someone for failing to do the impossible (unless he is to blame 
for its being impossible). Do we require a demonstration of the truth 
of this principle? Must it be derived from premises more evident than 
itself? I submit that it is more probably true than either SP1 or SP2 
taken singly -  let alone taken in conjunction. I conclude that by fail- 
ing to recognize this Zimmerman not only introduces an unnecessary 
complexity into the debate initiated by Frankfurt, but represents PAP 
(and hence S-PAP) as vulnerable on fronts that do not exist.16 

Received June, 1985 
Revised October, 1985 

15  A recent discussion  which  demonstrates  once  again the difficulty of formulating 
a reportive definition  of free action is Three  Concepts  of Free Action' by Don Locke 
and Harry Frankfurt, Aristotelian Society Supplementary  Volume  49 (1975) 95-125. 
16  I wish  to  thank  an  anonymous  referee  for the  helpful  comments  on  an  earlier 

draft of  this  paper. 

This content downloaded from 69.163.225.153 on Sat, 18 Jan 2014 13:47:04 PM

All use subject to JSTOR Terms and Conditions

