Responsibility and Motivation*

Could an agent be morally responsible for an action most of whose causes predate his birth? The hard determinist would answer in the negative, given that such an action is the result of a chain of events over most of whose 'links' the agent had no control. Harry Frankfurt's much discussed case, she would say, may show that a person can justifiably be held accountable for an action even though his performance of it was unavoidable. But it does not refute her principle that an agent is morally responsible for his action only if he could have "produced, altered, or prevented" each event in its causal history, where an event's causal history is made up of its mediate and immediate causes (Pereboom, 26). Given that no human agent exercises such control over his conduct, we are faced with the possibility that all of our common sense ascriptions of moral responsibility are false.

Call the notion upon which this possibility is based "ultimate moral responsibility" (UMR). It might then be asked, is there some other concept backing our everyday attributions of moral responsibility? How is it, in other words, that some agents are for all practical purposes morally responsible whereas others are not?

The following account answers this question. It is based upon the idea that 'being morally responsible' is ambiguous, meaning either being the controller of each one of an action's causes or being an action's author in a sense to be explained.1 In making everyday attributions of moral accountability, I argue, we do not rely on UMR. It is instead the notion of authorship that legitimizes what Peter Strawson calls our "reactive attitudes," that is, the practices of praising, blaming, resenting, forgiving, thanking, and respecting that are associated with assignments of moral responsibility (Strawson, 48-9).

I

Let us focus on the following "Frankfurt" case:2 Black, a partisan neurosurgeon, wants Jones to vote Republican. Jones is known by Black to be leaning toward the GOP candidate. Black, however, does not wish to take any chances. Thus, he contrives a medical ruse that allows him to implant a device in Jones' brain the operating of which will cause him to make the desired choice. The device will be activated if Jones decides against the GOP candidate, rendering irresistible the desire to vote Republican.

It seems, then, that Jones is not able to avoid voting Republican. If he decides in favor of the GOP candidate, that aspirant will receive his vote; if he decides against her, the GOP candidate will still get his vote. Yet, when Jones votes Republican 'on his own', i.e., without being caused to do so by the activation of Black's device, it appears to be something for which he is responsible- even though it is the only course of conduct open to him.3 It is cases like this that led Frankfurt to reject the principle that morally responsible agency requires "alternative possibilities" of action (PAP).

Why, though, would we hold someone like Jones morally accountable for his action? Jones is practically speaking morally responsible (PMR) for voting Republican, I want to say, because he failed to perform the mental act that would have activated Black's device, viz., trying to avoid voting Republican. The device did not render him incapable of performing this deed, although it guaranteed that it would have an unintended consequence. That is to say, Jones could have "countermanded" his desire to vote Republican, even if Black's mechanism would have made his effort futile. That Jones instead made this desire the vote's motive, i.e., gave it over to the control of his will rather than attempting to resist it, makes him the author of the above deed, and, as such, (practically speaking) morally responsible for it. Responsibility, in this sense, requires neither alternative courses of non-mental action nor control over each one of an action's approximate causes. (Presently, I shall have more to say on whether or not this sense of responsibility entails at least a modification of PAP.)

From a Frankfurt case, then, the following Kantian account of moral responsibility may be culled:

An agent is morally responsible for an action A iff A's motive is a desire to which she has "assented" (PMR).

According to PMR, moral responsibility stems from the forming of motives out of desires.

In assenting to a desire, one regulates one's will, determining its content by making one desire effective rather than another. It is this process, involving the exercise of what Thomas Reid calls the "active powers," over which the responsible agent has control. These powers enable one to act autonomously, that is, activate the desires of one’s choosing. They do so by making possible an understanding of the comparative strength of competing urges as well as the likely consequences of securing their objects. The exercise of one's active powers produces self-determined action, as opposed to automation. That a desire’s effectiveness is the result of one's exercise of this faculty is what makes one morally responsible for the ensuing action.4 (I shall speak below to the issue of responsibility for one's assentings.)

Thus, to answer the question posed above, in ordinary- I am tempted to say non-philosophical- contexts the judgment that someone is morally responsible for doing something is based upon the notion of authoring an action. An agent motivated by a desire to which she has not assented, on the other hand, is not typically held morally accountable for her conduct. Such an agent is usually thought of as the victim of some form of mind-control, e.g., insanity or a psychological cumpulsion. Also, an agent known to have no motive- if such is possible- would not be held responsible for her conduct. One suspects, however, that the plea 'I don't know why I did it' is either shorthand for 'I don't know why I assented to such a desire' or symptomatic of repression. In neither case would the agent's responsibility be mitigated.

II

One might object that PMR entails that one is responsible for anything that one tries to do, whether one's attempt is successful or not. It is absurd to think that if Jones tries to kill Smith, he is responsible for taking his life even if the attempt goes astray.

PMR, however, does not have this consequence. Strictly speaking, in the case of a successful attempt on someone's life, PMR would have it that the action for which the killer is responsible is, say, his pulling of the trigger, which should be distinguished from the killing itself- its result.5 (There is a sense of 'being killed' according to which Jones' shooting, in a successful attempt on Smith's life, is identical to the killing of Smith, viz., the sense in which a killing is the causing of a death. This is not the sense relied upon in the objection, which is rather 'being dead as the result of someone's or something's action'. It is in this sense that Smith's killing fails to obtain.) The killer's responsibility for the consequence of his action follows from the following principle:

An agent is morally responsible for the results of her action if she intended it to produce those results or if she acted negligently (CP).

E.g., the drunken motorist whose erratic driving results in an accident is morally responsible for it, even though she did not try to cause it, since it is the result of her negligence. Thus, in a case of attempted murder, the perpetrator would be responsible for the same action- the pulling of a trigger- as he would have been had his effort succeeded, though he would not be responsible for a killing. PMR plus CP yields a similar treatment of cases of "nonnegligent ignorance." Were I to discharge a firearm out of curiousity, without understanding its possible effects, my moral responsibility would not extend beyond my pulling of the trigger.

Fischer and Mark Ravizza raise what they call a "divide and conquer" objection to "actual sequence" theories of responsibility such as PMR. (Fischer and Ravizza, [1993b] 36-8).6 Jones in our Frankfurt case is responsible for authoring a vote in favor of the GOP candidate, something to which there was, we shall assume, an alternative, viz., making effective the desire to refrain from voting Republican.

Indeed, it is only on this assumption that Jones is an appropriate candidate for attributions of responsibility. Thus, that to which there is no alternative— voting Republican— and that for which Jones is responsible are different actions. Moreover, at least a revised version of PAP, it seems, must be accepted by a defender of PMR: an agent is responsible for trying to perform an action A only if she could have avoided trying to perform A, i.e., to be responsible one's power of selecting one's motives must be intact, even if its exercise will be futile. PMR appears inconsistent with Frankfurt's case against PAP.

To meet the first half of this objection, it suffices to point out that being responsible for having tried to perform an action does not disqualify one from being responsible for performing it. According to PMR, one is responsible for an action A for having tried to perform it: responsibility is conferred to one's action in virtue of one's forming of its motive. Thus, there is in our Frankfurt case a single action to which there is no alternative and for which Jones is responsible, viz., that of voting Republican.

PMR, then, is consistent with Frankfurt's argument against PAP as applied to non-mental acts. It might still be maintained, however, that the divide and conquer objection suggests the validity of the revised version of PAP sketched above. For, unless one is able to perform both the mental act of making the desire D effective and the mental act of countermanding D, one seems incapable of being responsible for the conduct resulting from D's motivation. There appears to be no Frankfurtian counterexample to this principle. Were Jones incapable of even the latter act- because Black's device would force him to allow his desire to vote Republican to become effective were he to so much as think of resisting it- he would seemingly lack what it takes to be a moral agent, being incapable of bringing his desires into conflict with each other- of controlling himself.

PMR, however, may be defended without making this concession. One can reply that having failed to even contemplate exercising this capacity, its diminishment does not mitigate one's responsibility for the results of one's omission. If one has one's mind 'made up' to allow a certain desire to become effective, the fact that one can not change one's mind does not render one blameless for the ensuing deed.

The situation is really no different than that presented by our Frankfurt case: despite being unable to perform the mental act of trying to avoid voting Republican, Jones is able to execute the intellectual task that would cause him to begin countermanding this desire, the detection of which would activate Black's device. His failure to perform this act- to try to avoid voting Republican, that is, to form the "second-order" desire to resist his tendency to select GOP candidates- is what makes him responsible for assenting to his desire, and, hence, for the ensuing vote. (More on such desires below.) Had he had such thoughts, on the other hand, he would have been accountable for neither act, since he would not have made his desire effective. Responsibility for a non-mental act always stems from a previous mental act: assenting to a desire.

Frankfurt distinguishes between acting freely- "doing what one wants because one wants to-" and acting with freedom of will (Frankfurt [1969 & 1971]). This distinction may be used to explain the just mentioned possibility of being responsibile for making a desire effective, though it should be kept in mind that in being the author of a deed one acts freely and is thus responsible for one's action- whether one acts with freedom of will or not.7

Acting with freedom of will, according to Frankfurt, is a matter of satisfying the conditions for acting freely plus forming one's motive through "reflective self-evaluation." A person, on his view, possesses not only the ability to allow desires to become motives but also the capacity to determine a desire's merit, i.e., reflectively evaluate, in light of one's moral principles and practical interests, possible motives. In so assessing one's desires, one forms "second-order desires." That is, one aspires to have and act upon certain first-order desires rather than others. To act with freedom of will, according to Frankfurt, is to act, freely with one's motive being consonant with one's system of second-order desires. Aligning this notion with PMR, one acts with freedom of will just in case one is the author of a deed whose motive has passed one's reflective self-evaluation.

In terms of second-order desires, then, sense can be made of the notion of responsibility for forming a motive. By developing a second-order desire to act from a specific first-order desire, one disposes oneself to make the latter urge effective. Those of one's assentings that are manifestations of such dispositions are mental acts for which one is responsible, since one has formed their motives. It should be re-emphasized, however, that one may be morally responsible for an action even though its motive is not aligned with one's system of second-order desires. This would be the case were one to assent to a desire about which one was reflectively ambivalent.

At this point, the problem of "artificially induced assentings" presents itself. The intervener in our Frankfurt case could have had his way with Jones by programming him to have the second-order desire to desire to vote Republican, in which case the eventual vote would have stemmed from an assenting rather than the operation of an implanted device. It seems that one may form a motive without being responsible for the resulting action.

Rebutting this criticism involves pointing out that PMR* (PMR augmented by Frankfurt's account of acting with freedom of will) dovetails with the "traditional" view of the will as a "faculty" or system of abilities. For making a desire effective, as noted above, may be regarded as the exercise of one's active powers while reflective self-evaluation can be seen as a function of certain "cognitive capacities."8 The latter provide for the rational assessment of first-order desires, allowing, relative to an agent's interests, the development of a favorable set of potential motives. They are, in other words, the reasoning skills exercised in a critical examination of one's volitions: such a review being conducted in light of one's understanding of one's moral obligations and practical concerns. Thus, the exercise of these capacities promotes higher-order desires or a "rational self-interest," being employed to prioritize a set of first-order desires.

One therefore exercises one's will either in forming a motive or developing one's system of second-order desires. A free will, on this view, may be understood in terms of the proper functioning of the capacities involved in carrying out these tasks, where, teleologically, one's will functions properly just in case it works so as to promote one's highest values and concerns, that is, to borrow a phrase from John Rawls, one’s "plan of life."

An artificially induced assenting, then, turns out not to be an assenting at all. For one's will is not functioning properly in responding to a second-order desire it has not itself formed. Since they are a system, the abilities that make up one's will are supposed to work in concert. In particular, the second-order desires that manifest themselves in the formation of one's motives are to be developed via the exercise of one's cognitive capacities. The "assenting" of a victim of psychological programming is not a respone to an exercise of his cognitive capacity. His will has been circumvented in that he views favorably a potential motive that he himself has not assessed. (The above epistemic requirement also entails that second-order desires that arise from feelings and emotions that have not themselves been appraised, e.g., instinctive shame or disgust over a first-order desire, are not such as to effect an assenting. In that way they are akin to compulsive first-order desires, which mitigate responsibility for the acts that they effect.)

Thus, assenting, when properly understood as the exercise of a part of a faculty, is sufficient for moral responsibility. A morally responsible agent, according to PMR*, acts of her own free will. This is just what one would hope for from an analysis of the common sense notion of moral responsibility.

PMR* is a version of ahistorical compatibilism. An agent's personnel background, on this view, is irrelevant to the question of whether or not she is responsible for her action. What counts is simply the way in which her motives were formed. Her action's approximate causes are not to be taken into account. Moreover, PMR* does not require a "break" in the chain of causes leading up to a free action: it is enough that the active power of an agent’s will is exercised.

It might be objected that this approach is inadequate, that an agent's history must be looked at in order to determine whether or not she should be held accountable for her action. Questions of responsibility often turn on the issue of "reasonable alternatives": an agent is blameless for acting improperly if she is justifiably unaware of options besides the action she performs. If whether or not an agent is justified in holding a particular belief depends upon her personal history, theories such as PMR* are false.9

But to posit such a connection is merely to commit the Genetic Fallacy: that is, conflate a belief’s warrant with its etiology. It is not her personal history per se that determines which beliefs an agent is warranted in holding. Rather, it is the cognitive effect of her experiences that is here relevant. Specifically, it is her total doxastic state at the time at which she acts that settles the question of what she may then justifiably believe. To settle whether or not she holds a given belief, it is enough to determine her current dispositions towards the propositions that warrant its acceptance and the logical principles licensing its inference therefrom.

Moreover, even in the case of second-order desires, the development of which takes time and thus may be thought of as establishing a link between personal history and moral responsibility, an agent’s prevailing inclinations are decisive, since one would disavow an artificially induced disposition were one unable to recall having formed it via the exercise of one’s cognitive capacities. And if such a "memory" had been effected along with the disposition, it would entail a belief to which one was not entitled, given other propositions one holds. Thus, there appears to be no situation in which a finding of moral responsibility requires an investigation of an agent's personal history in addition to one of her current beliefs/dispositions.

III

The question now becomes, does PMR* authorize our reactive attitudes? It can be shown that it not only legitimizes our responses to particular actions but also our assessments of character. Consider blaming. Knowing that an individual has authored an evil deed, we blame her for it. Realizing further that she has not merely assented to a reprehensible desire but reflectively endorsed it, our reprobation increases as we then become aware of her vicious character. Also, one's blame for a harmful action is mitigated, if one made an earnest attempt to keep its motive, which proved overwhelming, from becoming effective: we realize the limitations of our will, though we judge more harshly those wrongdoers who fail to exercise it or do so only 'half-heartedly'.

PMR* also legitimizes praising. The agent who conforms to our standards of interpersonal conduct is praised for her efforts toward 'doing the right thing'. That it is not merely such conformity that is praiseworthy is indicated by the fact, stressed by Kant, that one who fails to act properly is nevertheless rightly praised, if she attempted to suppress the recalcitrant desire that led to her transgression. Moreover, the praiseworthy agent whose motive conforms to her system of second order desires is respected for her integrity. The virtuous person, children are taught, practices self-control, which may be regarded as an exercise of the will’s cognitive powers.

Since they are closely linked to blaming and praising, one might expect similar connections to exist between the other reactive attitudes and the constitutive notions of PMR*: assenting and self-evaluation. Resentment is attendent upon blame. It stems not only from the belief that one has been wronged, but also from the conviction that the wrongdoer has willfully satisfied a harmful desire at one's expense. Such a desire, the injured party would angrily contend, should have been suppressed rather than assented to. Moreover, such an attitude is consistent with the belief that an agent lacks control over most of her action’s causes, since it is not based upon UMR. It is instead an agent's control over his motives that justifies others in resenting his wrongdoings.10

The same may be said of forgiveness. The wrongdoer's contrition entails an exercise in self-evaluation the outcome of which is an improvement in character: the development of a second-order desire whose intended effect is the prevention of further harms. The recognition of such an effort elicits forbearance.

Gratitude is partly a response to the benefits one receives from another’s responsible action. But at a deeper level it involves one’s understanding that another saw fit to act out of a concern for one’s interests. That is why one may be grateful for another’s failed attempt at assistance, recognizing the ‘good intentions’ of one’s would be benefactor.

IV

Thus, PMR* satisfies a desideratum of a theory of moral responsibility, justifying the attitudes we adopt toward a person in finding her responsible for her conduct. It is appropriate to foster one of the reactive attitudes towards a morally responsible agent given that she is capable of forming her motives and evaluating her desires. This rationale is valid despite the fact that she lacks control over each one her action’s causes. Though failing to meet the demands of the hard determinist, PMR* recapitulates our everyday use of the term 'morally responsible'.

Notes

* I thank the following persons for commenting on earlier drafts of this paper: Paul Benson, Sarah Buss, Randolphe Clarke, John Martin Fischer, Harry Frankfurt, Walter Glannon, Ishtiyaque Haji, Robert Kane, Kenton Machina, Lawrence Powers, Eleonore, Stump, Michael Zimmerman, and an anonymous referee from this journal.

1. Cf. Pereboom for a defense of hard determinism. Pereboom's argument is adapted from Van Inwagen (1982, 1983). Cf. McKay and Johnson for a recent discussion of the latter's reasoning. Cf. also Frankfurt (1969), Van Inwagen (1978), Zimmerman, Fischer (1982), Fischer and Ravizza (1993b), and Clarke.

2. Cf. Frankfurt for the original presentation of such an example.

3. Cf. Allen for a defense of Frankfurt's claim that Jones lacks alternative possibilities.

4. PMR is consonant with Frankfurt's own reading of the case, viz., that Jones is responsible since he did not vote Republican only because he could not do otherwise (Frankfurt, 151-52). Smith's voting occurred, as it turned out, not because of the operating of Black's device. Rather, it was the result of Jones allowing his desire to vote Republican to become effective, which is what, according to PMR, makes him responsible.

5. Cf. Lombard (144-55) and Thomson for a discussion of the relationship between a shooting and the attendent killing.

6. This objection originated with William Rowe (Rowe, 276-78).

7. For discussions pro and con of Frankfurt's "hierarchical"

view of motivation see Thalberg, Zimmerman, Shatz, Slote,

Friedman, Christman, Taylor, Watson, Velleman, Abelson, Wolf Stump, and Anderson. Frankfurt defends his thesis in Frankfurt (1971a&b, 1975, 1976, 1987). Eleanore Stump and Fischer disagree as to what in Frankfurt's account is required for responsibility: acting freely, according to Stump; acting with freedom of will, according to Fischer (Stump, 213; Fischer [1986], 46). PMR reflects Stump's view given that Fischer's, as Stump points out, has counterintuitive results: one cannot be responsible for a deed unless one has subjected its motive to critical self-evaluation.

8. The term ‘cognitive capacities’ is taken from Joseph Campbell. Cf. Locke (Bk.2 Ch.XXI), Kane, and Campbell.

9. Cf. Machina.

10. Pereboom (36-41) argues that an acceptance of causal determinism- the view that every event is the inevitable outcome of the events preceding it- rationally requires an abdication of certain aspects of the reactive attitudes. This presupposes that it is UMR upon which the adoption an instance of one of these stances is based. I have been at pains to show that the reactive attitudes are linked instead to PMR.

References

Allen, Robert. "Re-examining Frankfurt Cases," unpublished.

Abelson, Raziel. (1988) Lawless Mind (Philadelphia: Temple University Press).

Anderson, Joel. (1995) "Knowing What a Person Really Wants," Presented at the 1995 American Philosophical Association Central Division Meeting.

Campbell, Joseph. (forthcoming) "A Compatibilist Theory of Alternative Possibilities," Philosophical Studies.

Christman, John. (1986) "Autonomy: A Defense of the Split-Level Self," Southern Journal of Philosophy 25: 281-93.

. (1991) "Autonomy and Personal History," Canadian Journal of Philosophy 21: 1-24.

Clarke, Randolph. (1996) "A Revised 'Direct Argument' for Responsibility-Incompatibilism," presented at the 1996 APA Pacific Division Meeting.

Fischer, John, ed. (1986) Moral Responsibility (Ithaca, NY: Cornell Univ. Press).

Fischer, John. (1995) The Metaphysics of Free Will. (Boston: Basil Blackwell)

Fischer, John and Mark Ravizza, eds.. (1993a) Perspectives on Moral Responsibility , (Ithaca, NY: Cornell Univ. Press).

. (1993b) "Introduction," in Fischer and Ravizza (1993a) : 1-41.

Frankfurt, Harry. (1969) "Alternate Possibilities and Moral Responsibility," in Fischer (1986): 143-52.

. (1971a) "Freedom of Will and the Concept of a Person." in Fischer (1991): 650-668.

. (1971b) "Three Concepts of Free Action: II" in Fischer (1986): 113-23.

. (1976) "Identification and Externality," in The Identities of Persons, ed. Amelie Oksenberg Rorty (Berkeley: Univ. of California Press, 1976): 239-51.

. (1982) "What We Are Morally Responsible For," in Fischer and Ravizza (1993): 286-95.

. (1987) "Identification and Wholeheartedness, "in Fischer and Ravizza (1993): 170-87.

Friedman, Marilyn. (1986) "Autonomy and the Split-level Self," Southern Journal of Philosophy 24: 19-35.

Kane, R. (1985) Free Will and Values Buffalo: State University of New York Press.

Locke, John. (1979) An Essay Concerning Human Understanding, Oxford: Clarendon Univ. Press.

Lombard, Lawrence. (1986) Events: A Metaphysical Study (Routledge & Kegan Paul).

Machina, Kenton. (1996) "Compatibilism, Injustice, and Oppression," presented at the 1996 APA Central Division Meeting.

McKay, Thomas and Johnson, David. (1995) "A Reconsideration of an Argument Against Compatibilism," presented at 1995 American Philosophical Association Central Division Meeting.

Pereboom, Derk. (1995) "Determinism Al Dente," Noûs 29: 21-45.

Reid, Thomas. (1983) Inquiry and Essays, Indianapolis: Hackett Publishing Company.

Rowe, William. (1991) "Responsibility, Agent-Causation, and Freedom: An Eighteenth-Century View," in Fischer and Ravizza (1993a): 263-285.

Slote, Michael. (1980) "Understanding Free Will," Journal of Philosophy 77: 136-51.

Strawson, Peter. (1963) "Freedom and Resentment," in Fischer and Ravizza (1993a): 45-66.

Stump, Eleanore. (1988) "Sanctification, Hardening of the Heart, and Frankfurt's Concept of Free Will," Journal of Philosophy 85: 395-420.

Taylor, Charles. "Responsibility for Self," in Rorty (1976): 281-99.

Thalberg, Irving. (1978) "Hierarchical Analyses of Unfree Action," Canadian Journal of Philosophy 8: 211-26.

Thomson, Judith Jarvis. (1971) "The Time of a Killing," Journal of Philosophy 68: 115-32.

Van Inwagen, Peter. (1978) "Ability and Responsibility," in Fischer (1986): 153-76.

. (1980) "The Incompatibility of Responsibility and Determinism," in Fischer (1986): 241-49.

. (1982) "The Incompatibility of Free Will and Determinism," in Free Will, ed. Gary Watson (Oxford: Oxford Univ. Press, 1983): 55-78.

. (1983) An Essay on Free Will, (Oxford: Oxford Univ. Press).

Velleman, J. David. (1992) "What Happens When Someone Acts?," in Fischer and Ravizza (1993a): 188-210.

Watson, Gary. (1975) "Free Agency," Journal of Philosophy 72: 205-20.

Wolf, Susan. (1990) Freedom Within Reason, (Oxford: Oxford Univ. Press).

Zimmerman, David. (1981) "Hierarchical Motivation and Freedom of the Will," Pacific Philosophical Quarterly 62: 354-68.