Intentions and Moral Permissibility: The Case of Acting Permissibly with Bad Intentions

 

 

 

 

Forthcoming in Law and Philosophy

 

 

 

© S. MATTHEW LIAO

New York University, 285 Mercer Street, Room 1005, New York, NY 10003; e-mail: matthew.liao@nyu.edu; www.smatthewliao.com

 


Intentions and Moral Permissibility: The Case of Acting Permissibly with Bad Intentions

 

Abstract

Many people believe in the intention principle, according to which an agent’s intention in performing an act can sometimes make an act that would otherwise have been permissible impermissible, other things being equal.  Judith Jarvis Thomson, Frances Kamm and Thomas Scanlon have offered cases that seem to show that it can be permissible for an agent to act even when the agent has bad intentions.  If valid, these cases would seem to cast doubt on the intention principle.  In this paper, I point out that these cases have confounding factors that have received little attention in the literature.  I argue that these confounding factors undermine the putative force of these cases against the intention principle.  Indeed, when cases without these confounding factors are considered, it becomes clear, so I argue, that intentions can be relevant for the permissibility of an act.   

 


Intentions and Moral Permissibility: The Case of Acting Permissibly with Bad Intentions

 

I. Introduction

What role, if any, do intentions play in determining the (moral) permissibility of an act?  An intention can be understood as what the agent’s aim was when performing a particular act, X, and how she saw X as promoting her objective; and the permissibility of an act, X, can be understood as whether an agent may X.[1]  Many people believe that intentions are sometimes relevant for determining the permissibility of an act.[2]   That is, they accept the

Intention Principle: an agent’s intention in performing an act can sometimes make an act that would otherwise have been permissible impermissible, other things being equal.

This is particularly the case for nonconsequentialists who subscribe to some versions of the Doctrine of Double Effect (DDE), which draws a distinction between intending harm and merely foreseeing harm.  There is no standard formulation of the DDE.[3]  According to one interpretation of it by Frances Kamm, there is a moral constraint on acting with the intention of doing harm, even when the harm will be used as a means to a greater good.[4]  However, it is permissible to act with the intention of employing neutral or good means to promote a greater good, even though one foresees the same harmful side effects, if (a) the good is proportionate to the harm, and (b) there is no better way to achieve this good.  As Kamm explains further, moral constraints such as the DDE are concerned with when it is permissible to harm innocent, nonthreatening persons who would not otherwise be harmed, for whom further life is a good, and who have not consented to being harmed.[5] 

            For example, consider

Terror Bomber: A bombardier is fighting a just war and intends to kill ten civilians by bombing a civilian-only area in order to frighten the enemy into surrendering.

Suppose that the civilians are innocent, nonthreatening persons who would not otherwise be harmed, for whom further life is a good, and who have not consented to being harmed.  Many believe that it would be impermissible for the bombardier to perform the bombing.  However, consider

Tactical Bomber: A bombardier is fighting a just war and intends to bomb a munitions factory in order to get the enemy to surrender, foreseeing though that ten civilians will certainly die as an unintended side effect.

In this case, many believe that it would be permissible for the bombardier to perform the bombing, even though it has been stipulated that the number of civilians who will be harmed is the same in Tactical Bomber as in Terror Bomber.  A difference between Terror Bomber and Tactical Bomber is that the agents’ intentions in the two cases are different.  In Terror Bomber, the bombardier is intending harm, even though the harm will be used as a means to a greater good.   In Tactical Bomber, the bombardier is intending to employ a neutral means, i.e., bombing a munitions factory, in order to promote a greater good, i.e., winning the just war.  Since on the DDE, there is a moral constraint on acting with the intention of harming but there is not a similar constraint on acting when one foresees the same harmful side effects, the DDE seems neatly to explain people’s judgment that Terror Bomber is impermissible while Tactical Bomber is permissible.

            A number of prominent nonconsequentialists have however challenged the intention principle.[6]  One challenge focuses on the difficulty of distinguishing between the intended and the merely foreseen consequences of an act, the so-called “closeness problem.”[7]  Another challenge focuses on presenting cases that appear to show that it can be permissible for an agent to act even when the agent has bad intentions.  A third challenge aims to deflate the role of intentions for permissibility by suggesting that in cases in which intentions appear to be relevant for permissibility, they are only derivatively, and not fundamentally, relevant for permissibility.[8]  A fourth challenge asks advocates of the intention principle to provide a rationale/justification for this principle.[9]

The main focus of this paper is to address the second challenge.  I shall set aside the first challenge altogether, but some of what I have to say will have some bearings on the third challenge.  In the conclusion, I shall briefly address the fourth challenge.  I begin by presenting cases by Judith Jarvis Thomson, Kamm and others that seem to show that it can be permissible for an agent to act even when the agent has bad intentions.  If valid, these cases would appear to cast doubt on the intention principle.  After clarifying the terms of the debate, I argue that Thomson et al.’s cases have confounding factors that have received little attention in the literature.  When cases without these confounding factors are considered, it becomes clear, so I argue, that intentions can be relevant for the permissibility of an act. 

 

II. The Case of Acting Permissibly with Bad Intentions

Thomson motivates the idea that it can be permissible for an agent to act even when the agent has bad intentions by offering the following case involving the use of lethal drugs for pain relief:

Terminal Illness: A patient is terminally ill and suffering great pain. The only course of medication that will relieve this pain will also cause the patient to die.  The patient wants to take this drug.[10]

Thomson asks, does the permissibility of administering the drug depend on the doctor’s intention?  In particular, does it depend on whether the doctor intends to relieve the pain by causing the patient to die, or intends to relieve the pain by giving the drug which will, inevitably, also cause the patient to die?  According to Thomson, it seems counterintuitive to suppose that the intention principle is relevant to such a case.

            Thomson goes even further and claims that the intention principle is not relevant even in

Vengeful Doctor: Everything is the same as Terminal Illness.  In addition, the doctor hates the patient and wants the patient to die.[11] 

According to Thomson, while a doctor who acts with such vengeful intentions is a bad person, as long as the act relieves the patient’s pain, the act remains morally permissible.  As Thomson explains, “That she will inject for that reason matters morally, not by way of fixing that it is morally impermissible for her to proceed, rather by showing something morally bad about her.”[12]  In other words, Thomson believes that intentions are relevant not to the permissibility of the doctor’s act but only to the assessment of the doctor’s moral character.  Moreover, in Thomson’s view, if the intention principle is relevant at all, then it should be relevant in a case like Vengeful Doctor.  However, given that the intention principle does not seem to be relevant in Vengeful Doctor, this suggests, according to Thomson, that intentions are never relevant for the permissibility of an act.[13]  

            Thomson’s objection appears to retain its intuitive force in other types of cases such as variations of Terror/Tactical Bomber.  For example, Kamm offers a case suggested by Thomson, what might be called

Child-hating Bomber:  A bombardier is fighting a just war and intends to bomb a munitions factory in order to get the enemy to surrender, foreseeing that innocent children will certainly die as an unintended side effect. The bombardier who has been selected to bomb the munitions factory is someone who would not have taken the job unless he knew that innocent children would be killed as a side effect. That is, it has always been a goal of his to kill some innocent children, and he sees his bombing of the factory as a means of achieving this goal.[14]

Kamm questions whether the fact that this bombardier intended to kill innocent children and would not go on the mission if it were not possible for him to kill innocent children means that it is impermissible for him to bomb the munitions factory.  Kamm’s response is “I think not. I agree with Thomson that an act is permissible because of its properties not because of the intentions.”[15]  In his recent book, Moral Dimensions, Thomas Scanlon also concurs with Thomson and Kamm about these cases.[16]  If these cases were valid, those who thought that the intention principle was correct would appear to be mistaken.

 

III. Clarifying the Terms of the Debate

To avoid talking past one another, it is useful to clarify the terms of the debate.

First, Thomson et al. are not claiming that intentions are not relevant for moral assessments.  Recall that Thomson believes that intentions are relevant for assessing a person’s moral character.  Scanlon is sympathetic to Thomson’s distinction between assessing an action and assessing an agent, but he believes, e.g., that the doctor in Vengeful Doctor could have acted permissibly with a bad intention and not be a bad person, if the doctor were acting out of character.  For this reason, Scanlon proposes a narrower distinction between the deliberative use of a principle, which considers whether a certain action is permissible, and the critical use of a principle, which assesses an agent’s decision making process.  According to Scanlon, on this distinction, intentions are relevant for assessing, not the agent’s overall character, but rather “the quality of the particular piece of decision making that led to the action in question.”[17]  In other words, Scanlon believes that the vengeful doctor could be someone who generally had a good character, but who on a particular occasion failed to give proper weight to certain considerations, which then led to an action that was not ideal but nevertheless permissible.  In either case, for Thomson et al., intentions can be relevant for some kind of moral assessment, just not for the permissibility of an act.[18]

            Second, it may be useful to distinguish between an intention’s being relevant for permissibility in a fundamental sense and an intention’s being relevant for permissibility in a derivative sense.  Scanlon, for example, accepts that intentions can be relevant for permissibility but only in a derivative sense.  He identifies two ways in which intentions can be relevant in this sense.  For one thing, an agent’s intention can have predictive significance, that is, an agent’s intention can tell us how her action is likely to be guided such as to which aspects of her situation she will attend and how she is likely to respond if things do not go as expected.[19]  Scanlon argues that this way in which an intention is relevant for permissibility is derivative, because

What are of fundamental relevance in these cases are the effects of the agent’s action on the world around her (or what it is reasonable to expect those effects to be).  Her intention is relevant in the ways just described only because it tells us something about those effects.[20]

The second way in which intentions can be relevant for permissibility, according to Scanlon, is that the intention with which an agent acts can affect what he calls the meaning of an action, and in some cases of what he calls expression and expectation cases, the meaning of an action can affect permissibility. As Scanlon defines it, the meaning of an action is the significance, for the agent and others, of the agent’s willingness to perform that action for the reasons that he or she does.  An example of an expression case is your calling a sick relative to inquire about her health and presenting yourself as being concerned about her when in fact you are not moved by this reason (e.g. you may have called her so that your parents would stop nagging you).  An example of an expectation case is your entering into a conversation with another only because she assumes (perhaps without your having done anything to encourage this assumption) that you have certain intentions (e.g. that you fancy her).  In both cases, Scanlon agrees that facts about your intentions are relevant to the permissibility of your action.  But he argues that the intentions are “relevant only in a derivative way, as a consequence of a more basic moral requirement not to mislead others or take advantage of their mistaken beliefs about one’s intentions.”[21]

            It is worth noting that some people do not think that this distinction between permissibility in a fundamental sense and permissibility in a derivative sense is relevant.  For example, Doug Husak believes that if intentions can be shown to be relevant for permissibility fundamentally or otherwise, then the claim that intentions are not relevant for permissibility should be rejected.[22]  In this paper, I shall not challenge this distinction because I shall identify cases that appear to support the idea that intentions are relevant for permissibility in a fundamental sense.  I shall henceforth assume that we are concerned with whether intentions can be relevant for permissibility in a fundamental sense, other things being equal.  To save words, I shall drop the phrase ‘in a fundamental sense, other things being equal’ unless the context requires it.

            Third, a note on methodology.  It may be tempting to want to settle the issue of whether intentions are relevant for permissibility by appealing to some larger, overarching philosophical theory.  But doing so would not be satisfactory for our purpose.  To see why, note that most consequentialists do not recognize the relevance of intentions for permissibility.  Suppose a consequentialist were to say to a nonconsequentialist who believes in the relevance of intentions for permissibility that on her larger, overarching philosophical theory about how permissibility works, namely on consequentialism, intentions are not relevant for permissibility.   This would hardly persuade this nonconsequentialist.  Hence, to advance this debate, it seems that one should begin by finding arguments/cases that even one’s opponent would have to admit have some intuitive plausibility.  As I see it, this is what Thomson et al. believe they have done, namely, they seem to have found cases that undermine the intention principle, and that even advocates of the intention principle would have to admit have some intuitive plausibility, even if they do not share Thomson et al.’s intuitions.  My methodology in this paper will be the same as Thomson et al.’s. 

            Finally, let me mention one possible way of responding to Thomson et al.[23]  My purpose here is to explain briefly why I shall not be taking up this approach in this paper, and not to argue that this approach could not succeed.  Some might think that acting for the wrong reason and not acting at all are not the only alternatives; there is a third alternative which is to act for the right reason.  For instance, it might be thought that what the bombardier ought to do is bomb the munitions factory for the right reason.  I am sympathetic to this response, but I think that more needs to be said for this response to be a problem for Thomson et al. [24]  To see why, recall that in Child-hating Bomber it is stipulated that the bombardier has bad intentions and that this case nevertheless appears to be a case of permissibility.  Two ways of responding to this case are relevant for us: either one has the intuition that this is a case of permissibility or one does not.  Suppose that one has the intuition that it is a case of permissibility.   It would seem to be beside the point to argue that what the bombardier ought to do is bomb the munitions factory for the right reason, because one has already accepted that it is permissible for the bombardier to act even for the wrong reason.  Suppose instead that one does not have the intuition that this is a case of permissibility.  Merely to state that the bombardier ought to bomb the munitions factory for the right reason seems just to be reporting that one has a different intuition than Thomson et al. regarding this case.  But lest this debate become just a case of reporting clashing intuitions, it would be better if one were able to offer additional arguments that could advance this debate further. 

           

IV. The ‘Must Act’ Factor

Thomson et al.’s cases contain some known confounding factors that make them less than decisive.  For instance, the terminal illness cases may be drawing support from the idea that it is not a bad thing, morally speaking, for a person in such circumstances to die sooner rather than later.[25]  Indeed, if it is beneficial for the patient to die, then it could be argued that one does not harm the patient when one causes the patient to die in these cases.  Also, it has been stipulated that the patient wants to be given the lethal drug.  This raises the issue of whether the patient has consented to being given the drug.  If the patient has consented to being given the lethal drug even by the vengeful doctor, then it is not surprising that one would judge that it is permissible for the vengeful doctor to administer the drug to the patient. 

The bombing cases do not have these confounding factors, since it can hardly be beneficial for the innocent children to be killed from the bombing, and since typically the children would not have consented to being killed.  However, the bombing cases have another known confounding factor.  As Thomson and Kamm point out, the context of war may make acts that would otherwise be impermissible permissible.  Kamm writes, “we may have to consider the possibility that the context of war is introducing special factors that override in some ways a general principle like X,” where X is a principle of constraint that tells us when it is permissible to harm innocent persons.[26]  Similarly, in Thomson’s discussion of the permissibility of self-defense, she says that “I will have to bypass as too hard the question how the fact of war affects questions about self-defense . . . That means we here bypass the question what makes it permissible for the pilot to drop his bombs in [Tactical] Bomber, and impermissible (if it is) for him to do so in Terror Bomber.”[27]  Given that the bombing cases have this particular confounding factor, this is a reason to be cautious about drawing conclusions from these cases.

  In fact, however, the terminal illness cases and the bombing cases have other confounding factors that have received little attention in the literature.  As I shall now argue, these confounding factors appear greatly to undermine the putative force of these cases against the intention principle. 

One confounding factor is what can be called the ‘must act’ factor.   Thomson et al. appear to want us to think that the agents in their cases ‘must act’[28] and that it would be impermissible for the agents not to act.  For example, in the terminal illness cases, one might think that the agent must give the patient the drug because the patient is suffering great pain and wants the drug, and because it is not even obvious that one would be harming the patient if one were to give the patient the drug.  In the bombing cases, one might think that the agent must bomb the munitions factory because, for example, the good of winning the war is overwhelmingly great.[29]  Indeed, Thomson et al. have made retorts such as ‘What is the agent supposed to do, not act?’ and ‘Do we have to wait until we find an agent with the right intentions in order for a particular act to be done?’  These retorts invite us to think that ‘not acting’ is not an option and that a particular act must be done now.[30] 

 Likewise, in the course of arguing that the option of acting for the right reason is not available to an agent who does not recognize the right reason, Scanlon also asks us to consider a case in which the agent must act:

Rescue Case: Suppose that a person is in danger of dying.  An agent hates the person who needs help and would be happy to see him die, but she saves him anyway because she does not want him to die right then, since that would mean that his heir, with whom she is locked in a bitter political contest, would have much more money to spend on his campaign.[31]

Scanlon considers the suggestion that “it might be impermissible either to bring about a result with certain bad reasons in mind or to fail to bring it about at all, and that the only thing that would be permissible would be to bring it about for the right reasons.”[32]  In response, Scanlon also asks rhetorically, “What is an agent supposed to do in such a situation, not save the person?”, thereby implying that one must save the person and that it would be impermissible not to save the person.[33] 

            It is understandable that Thomson et al. would want to appeal to ‘must act’ cases to argue that intentions are not relevant for permissibility of an act.  After all, it seems to be a tall order to show that an agent’s intentions can change a ‘must act’ to a ‘may/must not act.’  However, this strategy misses the target because the intention principle is not committed to showing that an agent’s intentions can change a ‘must act’ to a ‘may/must not act.’  Recall that the intention principle says that an agent’s intention in performing an act can sometimes make an act that would otherwise have been permissible impermissible, other things being equal.  This can be restated as an agent’s intention in performing an act can sometimes make an act that an agent otherwise ‘may do’ into an act that the agent ‘may/must not do,’ other things being equal.  In other words, the intention principle seems only committed to the idea that an agent’s intentions can sometimes change a ‘may act’ in to a ‘may/must not act.’

To see that intentions can be relevant for the permissibility of an act, let us consider cases in which the ‘must act’ factor is not present.  For example, consider

Standard Trolley:  A runaway trolley is headed toward five innocent people who are on the track and who will be killed unless something is done.  Abigail, who is passing by, can push a button, which will redirect the trolley onto a second track, saving the five people.  However, on this second track is an innocent bystander, who will be killed if the trolley is turned onto this track.

In Standard Trolley, it seems that it would be permissible for Abigail to push the button.[34]  In this case, most nonconsequentialists, including Thomson and Kamm, also believe that there is no requirement to push the button.[35]  That is, they believe that while Abigail may choose to kill the one, Abigail is not required to kill the one.  In other words, Abigail may allow the five to die if she chooses.  An explanation is that if Abigail does not push the button, then she merely lets the five die.  But if she pushes the button, she is actively deciding to the kill the one.  Within certain limits, so the thought goes, agents should not be required to cause harm to others, especially if they have deep personal commitments against doing so.[36]  Hence, Standard Trolley is a ‘may act’ case rather than a ‘must act’ case.

Now consider the following:

Enemy Trolley:  A runaway trolley is headed toward five innocent people who are on the track and who will be killed unless something is done.  An enemy, who is passing by, can push a button, which will redirect the trolley onto a second track, saving the five people.  However, on this second track is an innocent bystander, who will be killed if the trolley is turned onto this track.  The enemy does not care about saving the five.  He hates the innocent bystander on the side track, he does not try to hide the fact that he hates the bystander, and he sees this as his opportunity to kill the bystander.  In addition, the enemy is not committed to trying to kill the bystander by other means if this means of redirecting the trolley becomes unavailable.

In Enemy Trolley, it seems that it would be impermissible for the enemy to push the button.  Indeed, if the enemy were to push the button, it seems that the enemy would have murdered the bystander.  Suppose that there is no requirement to push the button in Standard Trolley.  There should likewise be no requirement to push the button in Enemy Trolley.  But if the enemy is not required to push the button, then it seems perfectly cogent to require the enemy not to push the button if the enemy is going to push the button with the wrong intention/for the wrong reason.  The intention principle straightforwardly explains how the enemy’s bad intention can change a ‘may act’ into a ‘may/must not act’ in this case.  

Indeed, suppose Thomson et al. were to ask, ‘What is the enemy supposed to do, not push the button?’  It seems that ‘yes, the enemy should not push the button’ would be a perfectly sensible response, given that there is no requirement to push the button in the first place.  Or, suppose Thomson et al. were to ask, ‘Do we have to wait until we find an agent with the right intentions in order for the redirection to be done?’  Again, since there is no requirement to push the button, it seems cogent to reply that ‘yes we should wait until we find someone who has the right intentions.’

Furthermore, even if Scanlon is right that the option of acting for the right reason is not available to an agent who does not recognize the right reason, in a case like Enemy Trolley in which an agent may act but is not required to act, the option of not acting would remain open.  If all of this is right, this shows that once the ‘must act’ factor is removed, intentions appear to be relevant for the permissibility of an act. 

 

V. The Institutional Role Factor

There is another confounding factor in Thomson et al.’s cases that has received little attention, which we can call the institutional role factor.  In particular, in both the terminal illness cases and the bombing cases, the individuals performing the actions are agents with role responsibilities acting within certain kinds of institutional structures.  When an agent is part of an institutional structure that is just, and when this institutional structure has authorized the agent to do certain things in certain circumstances, it might be thought that an agent’s intentions do not matter.[37]  For example, in Vengeful Doctor and Child-hating Bomber, the vengeful doctor and the child-hating bombardier are agents of institutional structures that are just.  The doctor belongs to a hospital, the aim of which is to help patients.  The bombardier belongs to a military, the aim of which is to win a just war.  Suppose that the institutions that license the doctor’s and the enemy’s actions generally act with good intentions, that is, suppose that they are just.  It might be thought that the doctor and the bombardier act permissibly, even if they have bad intentions, as long as they are only following protocols and orders issued and authorized by just the institutional structures. 

Let us grant for the sake of argument that an agent’s intentions do not matter when the agent is part of a just institutional structure that has authorized the agent to do certain things in certain circumstances.  This way of explaining why Vengeful Doctor and Child-hating Bomber are cases of permissibility should not cause problems for the intention principle.  The reason is that advocates of the intention principle can maintain that intentions can be relevant when agents are outside of such institutional structures.  Since Vengeful Doctor and Child-hating Bomber involve agents acting within certain institutional structures, these cases would be silent on the question of whether intentions are relevant when agents stand outside of institutional structures.

To see that intentions can be relevant when agents stand outside of institutional structures, we can consider cases without the institutional role factor.  For example, it is possible to construct variations of the terminal illness cases and the bombing cases in which the agents are not part of any institutional structure.  These cases would of course inherit the other confounding factors mentioned above.   But it may nevertheless be instructive to consider them.  Consider, first, variations of the terminal illness cases.

Desert Island Enemy: A patient is on a desert island where there are many people, but no authorized doctors.  The patient is terminally ill and suffering great pain. The only course of medication that will relieve this pain will also cause the patient to die.  The patient wants to take this drug.  A mortal enemy of the patient has learned about the patient’s condition.  The mortal enemy hates the patient, does not try to hide the fact that he hates the patient, and he sees this as his opportunity to kill the patient.  In addition, the mortal enemy is not committed to trying to kill the patient by other means if this means of giving the patient the drug becomes unavailable.

And,

Desert Island Bystander: A patient is on a desert island where there are many people, but no authorized doctors.  The patient is terminally ill and suffering great pain. The only course of medication that will relieve this pain will also cause the patient to die.  The patient wants to take this drug.  A bystander, who is not a doctor, has learned about the patient’s condition, and the bystander wants to help alleviate the patient’s suffering.

Since no one on the desert island is part of any institutional structure, no one on the desert island has the proper authorization to administer the drug.  If having proper authorization were necessary, one would have to say that it is impermissible for both the bystander and the mortal enemy to administer the drug to the patient.  But it seems that it is at least permissible for the bystander to administer the drug to the patient.  At the same time, it seems impermissible for the moral enemy to administer the drug.  Indeed, it seems that this is so precisely because of the mortal enemy’s intentions.  Notice that since the mortal enemy does not try to hide the fact that he hates the patient, that his act would be impermissible would not be due to the meaning of his action.  Moreover, since the mortal enemy is not committed to killing the patient by other means if the means of giving the patient the drug becomes unavailable, that his act would be impermissible would also not be due to the predictive significance of his intention.  If all of this is right, it seems that intentions can fundamentally determine the permissibility of an act in these cases. 

            Of course, one may think that Desert Island Enemy is a case of permissibility, because one thinks that it is not worse for the patient to die sooner rather than later and/or because one thinks that the patient has consented to be given the lethal drug even by his enemy.  However, if these are the reasons one thinks that Desert Island Enemy is a case of permissibility, then, for reasons given above, one should already think that the terminal illness cases are not the right kind of cases for testing the intention principle. 

Consider next modified versions of the bombing cases without the institutional role factor. 

Thug:  A thug living in a region without governments has learned that a munitions factory could be bombed in order to win a just war and innocent children would be killed as a side effect.  The thug hates these children, he does not try to hide the fact that he hates them, and he sees this as an opportunity to kill them.  He does not care about whether the just war is won or not.   In addition, the thug is not committed to trying to kill the children by other means if this means of bombing the munitions factory becomes unavailable.

And,

Neutral:  An individual living in a region without governments has learned that a munitions factory could be bombed in order to win a just war and innocent children would be killed as a side effect.  The individual cares that the just war is won.

Since neither the thug nor the neutral individual is part of any institutional structure, neither has the proper authorization to bomb the munitions factory.  If having proper authorization were necessary, one would have to say that it is impermissible for both the neutral individual and the thug to bomb the munitions factory.  But even though the neutral individual does not belong to any institutional structure, given that what the neutral individual cares about is for the just war to be won, there are some reasons to think that it would be permissible for the neutral individual to bomb the munitions factory.  In contrast, given the thug’s intention, intuitively, it seems impermissible for the thug to bomb the munitions factory.  And, were the thug to bomb the factory and kill the innocent children, it seems that the thug would have murdered these children.  Moreover, since the thug does not try to hide the fact that he hates the children, that his act would be impermissible would not be due to the meaning of his action.  Similarly, since the thug is not committed to killing the children by other means if the means of bombing the munitions factory becomes unavailable, that his act would be impermissible would not be due to the predictive significance of his intention.  Since a key difference between Thug and Neutral is that the agents have different intentions, these two cases also suggest that intentions can fundamentally affect the permissibility of an act. 

            Again, one may think that the context of war is special such that the thug should be permitted to bomb the munitions factory despite his bad intentions.  But if this is the reason why Thug is a case of permissibility, then, as explained before, we should already question whether the bombing cases are appropriate for testing the intention principle.

At this point, it is worth considering whether there are cases without the institutional role factor that we can use to test the intention principle.  Possibly.  Many writers, including Thomson and Kamm, believe that trolley cases can be instructive because one can remove many confounding factors from them.  Of course, some people regard the ‘virtue’ of the trolley cases as their ‘vice.’  That is, some people think that the trolley cases become too artificial when all the confounding factors are removed.  It is beyond the scope of this paper to settle this matter here.  Since Thomson and Kamm have both made (extensive) use of the trolley cases, and since we have also already used them, I shall assume that the trolley cases can be used in this context.  Before proceeding, it may be interesting to note that Philippa Foot’s original version of the trolley case also has the institutional role factor.  As Foot describes it, you are a trolley driver of a runaway trolley and you are about to drive the trolley into five track workers unless you redirect the trolley onto the sidetrack where another track worker sits.[38]  Interestingly, it is Thomson who notices this institutional role factor, and who proposes a version of the trolley case without this institutional role factor, what she calls Bystander at the Switch.[39]  As Thomson explains, “Some people may feel a difference between [Foot’s case and Bystander and the Switch]. In the first place, the trolley driver is, after all, captain of the trolley. He is charged by the trolley company with responsibility for the safety of his passengers and anyone else who might be harmed by the trolley he drives. The bystander at the switch, on the other hand, is a private person who just happens to be there.”[40]   Hence, of all people, Thomson should agree that the institutional role factor in the terminal illness cases and the bombing cases is a confounding factor that should be removed.

Consider again Standard Trolley and Enemy Trolley, which we have discussed previously.  In these cases, both Abigail and the enemy are without role responsibilities.  If they were to act, they would act outside of institutional structures.  As we have said, in Standard Trolley, it seems that it would be permissible for Abigail to push the button.  In contrast, as we have said, in Enemy Trolley, it seems that it would be impermissible for the enemy to push the button.  These two cases are structurally similar apart from the intentions of the agents.  Moreover, since the enemy does not try to hide the fact that he hates the patient, that his act would be impermissible would not stem from the meaning of his action.  And since the enemy is not committed to killing the bystander by other means if the means of redirecting the trolley becomes unavailable, that his act would be impermissible would also not be due to the predictive significance of his intention.  Assuming that the trolley cases are appropriate in this debate, these two cases provide strong evidence that once the institutional role factor is removed, intentions can fundamentally determine the permissibility of an act.   

 

VI. Conclusion

Thomson et al. have presented cases such as Vengeful Doctor and Child-hating Bomber to cast doubt on the intention principle.  As we have seen, however, these cases have confounding factors that have not been noticed in the literature and that undermine their putative force against the intention principle.  One confounding factor is that these cases appear to be cases in which the agent ‘must act.’  The ‘must act’ factor is a confounding factor because the intention principle is not committed to the idea that an agent’s intentions can change a ‘must act’ into a ‘may/must not act.’  The intention principle is only committed to the idea that an agent’s intentions can sometimes change a ‘may act’ into a ‘may/must not act.’  An example of this is Enemy Trolley, where the enemy’s bad intention makes an act, i.e., push the button, that would otherwise be permissible impermissible.

Another confounding factor is that the individuals performing the actions in Vengeful Doctor and Child-hating Bomber are individuals with role responsibilities acting within certain institutional structures.  Hence, these cases may be cases of permissibility because the doctor and the bombardier are following protocols set by institutional structures that are just.  However, this does not undermine the intention principle since these cases are silent on circumstances in which the institutional role factor is not present.  Indeed, when we examined cases such as Desert Island Enemy, Thug and Enemy Trolley, which do not have the institutional role factor, it becomes clear that intentions can be relevant for the permissibility of an act.   

As I mentioned at the outset, advocates of the intention principle have often been criticized for not having provided convincing theoretical justification for the intention principle.  I shall conclude by offering two brief remarks regarding how the intention principle can be given additional justification. 

First, we should not rule out the possibility that the intention principle may just be a basic principle.  By a ‘basic’ principle, I mean that either further theoretical justifications given for it are not plausible; or even if they have some plausibility, they are not as plausible as the principle itself.  Consequentialism seems to be such a basic principle, and the intention principle may be also.  Given that it does not seem to be a problem for consequentialism to be a basic principle, it should not necessarily be a problem for the intention principle either.  If the intention principle were indeed a basic principle, then the only way to defend it would seem to be to provide intuitive justification for it and to explain away rival intuitions, which is what we have done above.

            Second, drawing on Kant’s and Warren Quinn’s work, let me propose what I take to be a familiar and intuitively plausible way of providing the intention principle with some theoretical support: acting with certain intentions towards certain beings is sometimes tantamount to disrespecting those beings, and disrespecting certain beings is relevant for the permissibility of an act.[41]  Disrespecting certain beings involves regarding and treating those beings as mere means to one’s ends, and not as ends-in-themselves.  In particular, those beings are regarded and treated as available to be used for one’s own purposes.  To see that acting with certain intentions towards certain beings is sometimes tantamount to regarding and treating those beings as available to be used for one’s own purposes, consider the following:  if an enemy acts with certain intentions, e.g. he tries to kill you for his own satisfaction, it seems that the enemy would be regarding and treating you as available to be used for his own satisfaction, which is tantamount to disrespecting you.

Disrespecting certain beings is relevant for the permissibility of an act because if a being deserves our respect and ought to be respected, then it seems prima facie impermissible to disrespect the being.   An example of this is the case of disrespecting persons.  As beings with great intrinsic value – perhaps owing to their capacity for moral agency – many would accept that persons deserve our respect and ought to be respected.  If so, it seems prima facie impermissible to disrespect persons.  If acting with certain intentions towards certain beings is sometimes tantamount to disrespecting those beings, and if disrespecting certain beings is relevant for the permissibility of an act, then it seems that acting with certain intentions towards certain beings is sometimes relevant for the permissibility of an act. 

            This respect account can accommodate the cases we have discussed above.  For example, Terror Bomber seems impermissible because in intending to kill the civilians in order to frighten the enemy into surrendering, the bombardier seems to regard and treat the civilians as available to be used for the bombardier’s own purposes, which is tantamount to disrespecting the civilians.  Similarly, Desert Island Enemy also seems impermissible because in intending to kill the patient whom he hates, the enemy seems to regard and treat the patient as available to be used for the enemy’s purposes, which again is tantamount to disrespecting the patient.  In contrast, in Tactical Bomber, the civilians are not regarded nor treated as available to be used for the bombardier’s purposes.  Indeed, if the civilians had shelters, the bombardier would see no point in their refusing to use them. 

Obviously, much more can and needs to be said regarding the respect account.  In any case, a full defense of the intention principle would need to address the closeness problem, something I have not done in this paper.  Nevertheless, I hope to have shown that once confounding factors such as the ‘must act’ factor and the institutional role factor are removed from Thomson et al.’s cases, it becomes clear that intentions can be relevant for the permissibility of an act.[42] 

 



[1] See, e.g., Scanlon, T. Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Belknap Press, 2008, pp. 9-11.  Understood in this way, an intention can be distinguished from acting intentionally, according to which an agent is aware of what she is doing or realizes that something would be a consequence of her action.  Scanlon calls acting intentionally the wider sense of ‘intention’ (Moral Dimensions, pp. 10-11). 

[2] Many consequentialists would think that intentions are not relevant for the permissibility of an act, since consequentialists typically judge the permissibility of an act by the consequences of the act rather than by an agent’s intentions.  However, see Adams, R.M. “Motive Utilitarianism.” Journal of Philosophy 73 (1976): 467-81, for a version of consequentialism that seems to allow intentions to play a significant role in determining the permissibility of an act.

[3] Many have cited a passage from Aquinas as the origin of the DDE (Aquinas, T. Summa Theologiae, II, ii, q. 64, a. 7). 

[4] See Kamm, F. M. Intricate Ethics: Rights, Responsibilities, and Permissible Harm. New York: Oxford University Press, 2007, p. 93, for this formulation.  For a different account of the DDE, see Quinn, W. Morality and Action. Cambridge: Cambridge University Press, 1993.

[5] Kamm, Intricate Ethics, p. 138.

[6] Thomson, J. J. “Self-Defense.” Philosophy and Public Affairs 20 (1991): 283-310; Thomson, J. J. “Physician-Assisted Suicide: Two Moral Arguments.” Ethics 109 (1999): 497-518; Kamm, Intricate Ethics, op. cit.; Kamm, F. M. “Failures of Just War Theory: Terror, Harm and Justice.” Ethics 114 (2004): 650-92; Scanlon, T. “Intention and Permissibility I.” Proceedings of the Aristotelian Society Suppl., no. 74 (2000): 301-17; Scanlon, Moral Dimensions, op. cit.

[7] The literature on the closeness problem is vast.  For some discussions, see Hart, H.L.A. "Intention and Punishment." In Punishment and Responsibility. Oxford: Clarendon Press, 1968; Foot, P. "The Problem of Abortion and the Doctrine of Double Effect." Oxford Review 5 (1967): 5-15; Bennett, J. Morality and Consequences. Vol. II, The Tanner Lectures on Human Values. Salt Lake City: University of Utah Press, 1981; FitzPatrick, W. "The Intend/Foresee Distinction and the Problem of “Closeness”." Philosophical Studies 128 (2006): 585-617; Hills, A. "Intentions, Foreseen Consequences and the Doctrine of Double Effect." Philosophical Studies 133 (2007): 257-83.

[8] Scanlon, Moral Dimensions, op. cit.

[9] See, e.g., Scanlon, Moral Dimensions, p. 2.

[10] Thomson, “Physician-Assisted Suicide,” p. 513.

[11] Thomson, “Physician-Assisted Suicide,” p. 516.

[12] Thomson, “Physician-Assisted Suicide,” p. 516.

[13] See also Hanser, "Permissibility and Practical Inference." Ethics 115 (2005): 443-70, p. 455, for this interpretation of Thomson.  As Hanser says, “I think that it is clear from the rest of the passage . . .  that Thomson means to argue against the relevance of intentions to any sort of permissibility judgment” (p. 455).

[14] Kamm, “Failures of Just War Theory,” p. 666.

[15] Kamm, “Failures of Just War Theory,” p. 666.

[16] See, e.g., Scanlon, Moral Dimensions, op. cit.; Scanlon, “Intention and Permissibility I.,” op. cit.

[17] Scanlon, Moral Dimensions, p. 28.

[18] Given that Thomson et al. believe that intentions can be relevant for some kind of moral assessment, it might be asked, what would be lost if intentions were not relevant for the permissibility of an act?  One may of course care about this issue just because one cares about the truth of the matter.  In addition, one may care about this issue because one believes that full moral theories should not just assess characters and/or one’s decision processes; they should also be action-guiding, that is, they should help us determine what we must and must not do and what we may do.  In so far as some people believe that the intention principle is a core component of any plausible, full moral theory, some might think that the intention principle should therefore be able to help determine what we must and must not do and what we may do.

[19] Scanlon, Moral Dimensions, p. 12.

[20] Scanlon, Moral Dimensions, p. 13.

[21] Scanlon, Moral Dimensions, p. 40.

[22] Husak, D. “The Costs to Criminal Theory of Supposing That Intentions Are Irrelevant to Permissibility.” Criminal Law and Philosophy 3 (2009): 51-70, p. 57.

[23] There are other possible ways of responding to Thomson et al.  Again, owing to lack of space, I can only briefly explain why I shall not be taking up these approaches here, and I shall not argue that these approaches could not succeed.  For example, it might be thought that it is permissible to do an act with a wrong intention as long as it is the type of act that someone with a good intention could perform.  According to this line of thought, Vengeful Doctor and Child-hating Bomber are cases of permissibility because the types of act perpetrated by the vengeful doctor and the bombardier are the types of act that someone with a good intention could perform.  Kamm has though responded that in deciding whether an agent with a good intention could do a particular act, one is in effect considering the properties of the act, e.g., whether it leads to a greater good, and not considering the intention of the agent (“Failures of Just War Theory,” p. 666).  William FitzPatrick offers a more sophisticated version of this kind of response to Thomson et al., according to which ‘possible’ intentions are relevant for the permissibility of an act-type (“Acts, Intentions, and Moral Permissibility: In Defence of the Doctrine of Double Effect.” Analysis 63, no. 4 (2003): 317-21).  In particular, FitzPatrick holds that an act of the relevant type (e.g. bombing the munitions plant) is permissible just in case there exists a justification for this act-type in terms of a sufficiently worthy end (e.g. winning the war) such that it is possible to perform this act-type without intending anything illicit as a means.  However, among other things, there remains an issue about whether ‘possible’ intention of the sort that FitzPatrick has in mind is doing any real work in determining whether an act is permissible.

As another possible response to Thomson et al., Matthew Hanser has an insightful, inferential account of permissibility, according to which, if an agent φ’s with the intention of achieving E and he thereby acts permissibly, he might have acted impermissibly had he either (i) done something else with the intention of achieving E, or (ii) φ’d with the intention of achieving something else ("Permissibility and Practical Inference," op. cit.).  All too briefly, for our purpose, Hanser’s account is a purely formal account of permissibility, as Hanser himself has said.  This means that, e.g., with respect to Child-hating Bomber, the inferential account would only say that whether the bombardier is acting permissibly or not turns on whether his reason for acting is sufficient to justify his doing what he does.  The inferential account would however be silent on whether the bombardier’s reason for acting is in fact sufficient to justify his doing what he does.

[24] Scanlon offers a different response to this line of inquiry (Moral Dimensions, p. 61).  According to Scanlon, it is odd to say that in the case of a person who has a bad reason to do something, the only permissible course of action for her is to act for the right reason.  The reason, as Scanlon argues, is that this presupposes that it is open to the person to choose to act for the right reason, but Scanlon believes that while it is open to the agent to choose to act in one way or another, it is not open to her to choose to see a certain consideration as a reason for doing something.  I see it as a virtue of my argument that I do not need to settle this matter in order to show that the case Scanlon has presented proves too much.  This said, here are some reasons why I am skeptical of Scanlon’s claim.  For one thing, it seems that one can sometimes choose to see a certain consideration as a reason for doing something.  To give an example, it seems that I can choose whether to see Duchamp’s Fountain as just a toilet or not, and depending on how I see it, this may or may not then give me a reason to go to the museum to see it.  Moreover, it seems important to distinguish cases in which someone is ignorant of certain consideration from cases in which someone is knowledgeable about certain consideration.  Scanlon’s argument has more force against the former, since if a person is ignorant of certain consideration, it is indeed difficult to see how she can choose to see that consideration as a reason for doing something.  But, Scanlon’s denial notwithstanding, it seems that Scanlon’s argument has less force against the latter, since the fact that a person is knowledgeable about certain consideration suggests that it is open to her to choose to see that consideration as a reason for doing something.   For our purpose, the latter kinds of cases appear to be the ones that are at issue.  Consider Vengeful Doctor and Child-hating Bomber.  It is not the case that the vengeful doctor and the bombardier are ignorant of the good considerations of administering the drug and bombing the munitions factory.  They know that doing so would relieve the suffering of the patient and cause the enemy to surrender.  However, they knowingly choose to act from some other consideration.  Given that they are knowledgeable about the good considerations, it seems that it is open to them to choose to see these considerations as reasons for action.

[25] Scanlon, Moral Dimensions, p. 19.

[26] Kamm, “Failures of Just War Theory,” p. 672.

[27] Thomson, “Self-Defense,” p. 298.

[28] I am taking ‘must act’ to be synonymous with terms such as ‘obligated to act’ and ‘ought to act.’

[29] However, if the good of winning the war were that great, it would raise the question of why Terror Bomber would also not be a case of permissibility. 

[30] See, e.g., Thomson, “Physician Assisted Suicide,” p. 514;  Kamm, “Failure of Just War Theory,” p. 666; Scanlon, Moral Dimensions, p. 57, p. 69.

[31] Scanlon, Moral Dimensions, p. 57.

[32] Scanlon, Moral Dimensions, p. 59.

[33] Scanlon, Moral Dimensions, p. 57.

[34] While many people share this intuition (see, e.g., Cushman, F., L. Young, and M. Hauser. "The Role of Conscious Reasoning and Intuition in Moral Judgment." Psychological Science 17 (2006): 1082-89), it should be acknowledged that not everyone does.  Recently, Thomson has offered theoretical arguments for why it would be impermissible for Abigail to redirect the trolley (“Turning the Trolley.” Philosophy & Public Affairs 36 (2008): 359-374.  FitzPatrick has argued though that Thomson’s arguments are not decisive ("Thomson's Turnabout on the Trolley." Analysis 69 (2009): 636-43).

[35] Thomson, J. J. “The Trolley Problem.” The Yale Law Journal 94 (1985): 1395-1415, p. 1398; Kamm, Intricate Ethics, p. 14.

[36] See, e.g., Williams, B. A Critique of Utilitarianism. In J. J. C. Smart & B. Williams (Eds.), Utilitarianism: For and Against. Cambridge: Cambridge University Press, 1973; Scheffler, S. Reject of Consequentialism. Oxford: Clarendon Press, 1982.

[37] Personally, I am inclined to think that an agent’s intentions can matter even when an agent is part of a just institutional structure.  But I can see how some people might think otherwise.  For one thing, they might think that an agent who is part of a just institutional structure would be able to comply with her duties more easily if she were able to act in a simple rule guided way without having to monitor her intentions.  Also, they might think that it would be too demanding to require an agent who is part of a just institutional structure repeatedly to reason in the correct way.  I thank Victor Tadros for these suggestions.

[38] Foot, “The Problem of Abortion and the Doctrine of Double Effect,” op. cit.

[39] Thomson, “The Trolley Problem,” p. 1397.  It should be mentioned that the institutional role factor is not, for Thomson, the primary reason to change from Foot’s case to Bystander at the Switch.

[40] Thomson, “The Trolley Problem,” p. 1397.

[41] See, e.g., Quinn’s defense of DDE (“Actions, Intentions, and Consequences: The Doctrine of Double Effect.” Philosophy and Public Affairs 18, no. 4 (1989): 334-51), which also draws on Kant’s work.

[42] I would like to thank David Wasserman, William FitzPatrick, Alec Walen, Jeff McMahan, Doug Husak, Guy Kahane, Jussi Suikkanen, Matthew Hanser, Steve Clarke, James Edwards, Caspar Hare, Colin Marshall, Steve Sverdlik, Victor Tadros, Wibke Gruetjen, Collin O’Neil, Ben Sachs, Greg Bognar, Joseph Raz, Gerald Dworkin, Thomas Scanlon, Frances Kamm, Alison Hills, Arthur Kuflik, David Velleman, and audiences at the International Society for Utilitarian Studies at University of Berkeley, the James Martin 21st Century Advanced Research Seminar at Oxford University, and the Bioethics Colloquium at New York University, for their very helpful comments on earlier versions of this paper.