, Volume 24, Issue 2, pp 139–152 | Cite as

Lesser Evil Reasoning and its Pitfalls

  • Georg SpielthennerEmail author


We are often faced with dilemmatic situations in which we must choose between alternative courses of action, both of which will have a bad outcome. It is commonly held that in such cases it is both uncontroversial and unproblematic that we have to choose the lesser evil. However, despite its frequent application in ethical decision-making, lesser evil reasoning is not well understood by most of its advocates and it thus occasions much misunderstanding and it presents a number of pitfalls. In this paper I shall thus, first, analyse the concept of a lesser evil; second, offer a critical assessment of lesser evil reasoning; third, discuss the most common pitfalls writers tend to fall into; and finally, explore the application of lesser evil reasoning to choices under uncertainty.


Ethical reasoning Evil Lesser evil reasoning Practical reasoning Principle of the lesser evil 

We are often faced with practical conflict-situations where a greater evil can only be avoided when a lesser evil is caused or permitted. To illustrate this, consider a bank cashier during a robbery. He may be caught in the dilemmatic situation that handing over the money causes a loss to the bank, but not handing it over poses a serious threat to his life. It is commonly held that if we are in a similar situation we have to choose the lesser evil.

Lesser evil reasoning was already common in ancient Greece1 and it was proverbial in Cicero’s time.2 Later, the principle “of two evils, the lesser is always to be chosen” (de duobus malis, minus est semper eligendum) became a moral axiom of Catholic moral thinking, and the practical importance of lesser evil reasoning in the recent literature of ethics can be gathered from the areas where it has been applied in decision-making. It plays an important role in law,3 and has been employed in political science and political philosophy.4 Lesser evil arguments have also been invoked in theology (see McCormick 1978), in medicine,5 in biomedical ethics,6 and even in public debates staged on the media7—and this list is by no means complete.

Yet despite its frequent use, lesser evil reasoning is not only marred by terminological confusions, its requirements are also not quite grasped, I think, by many writers who apply it to ethical problems. Authors often assume, or write as if they assume, that the notion of a lesser evil is clear and they employ it without explaining. But, as we shall see, it needs explaining. And because its logic is not well understood, lesser evil reasoning presents many pitfalls for its advocates also.

My objects in this paper are therefore, first, to offer an analysis of the concept of a lesser evil; second, to critically assess lesser evil reasoning, mainly, but not exclusively, from the logical point of view; third, to explain its most common pitfalls; and finally, to discuss applications of the principle of the lesser evil to practical reasoning under uncertainty, a topic that has received scant attention in the literature.

1 Some Preliminaries

Since authors are not agreed on what exactly lesser evil reasoning is, it will be helpful to clarify this notion and to discuss the requirements of its application.

1.1 The Nature of Lesser Evil Reasoning

Lesser evil reasoning is reasoning in accordance with the so-called principle of the lesser of two evils (hereafter, PLE). Thus, the nature of this type of practical reasoning can best be explained by discussing this time-honoured principle. The correct formulation of the PLE is a matter of controversy. According to some formulations, we should always choose the lesser of two evils (see e.g., Donagan 1977, p. 153). However, this is not valid reasoning. If there are three bad alternatives a, b, and c available, where b is worse than a and c is worse than b then b is the lesser evil than c but it ought not to be chosen. Upon reflection this should be clear. This shows that the PLE requires that a choice needs to be made from only two bad options,8 and a more adequate formulation of the PLE is therefore that if we have to choose between two evils, the lesser evil ought to be chosen.

Some misconceptions about the PLE have their origin in an ambiguity of the term ‘evil’. In the tradition of lesser evil reasoning, ‘evil’ has been used both as a predicate of actions and of state of affairs. In the latter sense, evils can be things like pain, error, sickness or death. Depending on the terminological tradition, they are called physical-, premoral-, or nonmoral evils (see e.g., McCormick 1978, p. 29). Other evils, called moral evils, depend on the reasons from which acts are done. According to traditional Catholic moral theology, they imply a disordered will by which man denies his proper nature and insults God. Some authors use the phrase ‘lesser evil’ to refer to these types of evils, mostly to nonmoral evils.9 It should be noted, however, that in lesser evil reasoning ‘evil’ is a predicate of actions. The evil, which the cashier has to choose in the example given at the beginning of this paper, is an act (handing the money over) and not a possible consequence of his act or a state of affairs. To avoid these terminological confusions, I shall hereafter use ‘evil’ as a predicate of acts only.

In most conflict-situations we have more than two options to choose from. Since the PLE is not a valid principle for such situations, it has been generalized to the principle of the least evil.10 In fact, the principle of the lesser evil is best seen as a special case of the principle of the least evil because if there are only two alternatives available, the lesser evil is the least evil. But it could be argued that even the principle of the least evil is unnecessarily restrictive because it allows only a choice among evils and in most decision situations not all alternatives are bad. It seems therefore reasonable to include good alternatives also and to hold that we should choose the best or least evil course of action. There is, however, no room here to give these matters the attention they deserve and I shall therefore confine myself to lesser evil reasoning only.

1.2 The Range of Lesser Evil Reasoning

Before going on to a detailed consideration of lesser evil reasoning, we need to get clear about its scope. Can it be applied to any conflict-situation where we are caught between Scylla and Charybdis or are there practical problems that do not admit of its application? In this section I discuss two possible restrictions on the range of lesser evil reasoning: the view that there are moral absolutes and value incommensurability. For simplicity of exposition, I shall partition ethical theories into those that hold that there are moral absolutes and theories that deny this. (As will become apparent, this partition cuts across the distinction between consequentialism and deontology.)

(i) According to some moral traditions, notably Catholic moral theology, there exist so-called moral absolutes. That is, absolutely exceptionless moral norms whose violation is intrinsically evil (malum in se), regardless of the circumstances and independently of the consequences (see McCormick 1978, p. 42). Advocates of this view are not agreed on what actions are evil in this sense. Some regard abortion, euthanasia, adultery, and contraception as evil without qualification (e.g., Pope John Paul II in the encyclical Veritatis Splendor, no. 52),11 others defend a less demanding list of moral absolutes (see e.g., Donagan 1977, p. 88). It would be generally agreed, however, that the intentional killing of an innocent human being is absolutely wrong (see e.g., Quirk 1997, p. 512).

Adopting this view restricts the scope of lesser evil reasoning because it implies that some acts can never be a lesser evil. An act that is evil in se is wrong without qualification and must therefore never be performed. However, holding that such an act is a lesser evil implies that it should be performed and is therefore inconsistent with the view that it is evil in se.

I am not, of course, suggesting that absolutists cannot apply lesser evil reasoning. But they need to restrict it to acts that are not malum in se.12 Consequently, Pope Paul VI accepts in his encyclical letter Humanae Vitae that it is sometimes “lawful to tolerate a lesser evil in order to avoid a greater evil,” but he insists that “sexual intercourse which is deliberately contraceptive” cannot be a lesser evil because it is intrinsically evil (no. 14).13

However, many ethicists do not accept the view that certain types of acts are always wrong, regardless of the circumstances and independently of the consequences. For instance, Ross (1930) outlined a deontological theory of so-called prima facie duties. Prima facie duties (e.g., the duty to keep one’s promises or the duty to non-maleficence) are not absolute rules. In the case of a conflict of such duties, we must determine what action would bring about the greatest balance of right over wrong and this requires comparing the competing duties. Depending on the circumstances, the duty to keep one’s promises may be overridden by the duty of non-maleficence. That is to say that breaking one’s promise is, in the circumstances, the lesser evil.

Primarily, however, it has been consequentialists who denied the existence of absolute rules.14 Even though taking this stance it is not an essential component of consequentialism, proponents of this type of ethical theory tend to regard moral rules as rules of thumb. They may regard some norms as virtually exceptionless (e.g., that we must not kill innocents), but they will insist that under extreme circumstances even intentional killing of innocent humans could be right. The scope of lesser evil reasoning is, of course, broader if it can be applied to all types of acts.

(ii) Another possible constraint on lesser evil reasoning is value incommensurability. Obviously, holding that alternative a is the lesser evil than b implies that a and b are comparable. Some critics have held that actions cannot be compared with respect to their values. For example, Ramsey (1978) claimed that options are often incommensurable and therefore there is “no way to determine the greater or lesser good or evil” (p. 71). He admits that there are cases of commensurable values (e.g., in abortion cases when the life of a fetus must be taken to save the life of the pregnant woman), but he insists that in most moral cases a comparison of available alternatives is not possible and lesser evil reasoning can therefore not be applied.15 To see whether this objection against lesser evil reasoning is warranted, we need to have a closer look at value incommensurability. This is, however, not the place for a full-scale discussion of this issue. My present object is, rather, to bring out some points, which are relevant to our discussion.

Generally speaking, holding that the values of two or more things are incommensurable means that these things are not comparable with respect to their values because they cannot be measured on a common scale. There are, however, several forms of measurement. Scales of measurement are commonly broken down into four types, depending on how much information they can provide. Nominal level measurement allows only a classification of objects, ordinal measurement permits a ranking of objects, and on an interval level of measurement, the intervals between adjacent scale values are equal with respect to the attribute being measured. Measuring value on a ratio level would even justify holding that one thing is twice as valuable as another. Since there are different types of measurement, we should neither claim nor deny that actions are commensurable, sans phrase. We rather need to specify on what level of measurement their value is (in)commensurable. Two actions can be incommensurable on a stronger scale but commensurable on a weaker one.

It should be noted that lesser evil reasoning requires only that the values of options are measurable on an ordinale scale. Given two alternatives a and b, lesser evil reasoning requires holding that a is better than b (a > b), or b is better than a (b > a), or that a and b are of equal value (a ~ b). There is no need for a higher-level measurement of the values and the objections of Ramsey (1978) and others are therefore misguided because they are based on the assumption that alternatives need to be comparable on a more demanding scale.

Nonetheless, the question remains whether all actions are commensurable on an ordinale scale. Authors are not agreed on this. Broome (2004, pp. 81, 168) holds that ordinale incommensurability of values is common, while Griffin (1986) believes that we do not find values that are incommensurable on an ordinale scale (see p. 83). I won’t try to settle this question here. I wish to emphasize, however, that advocates of lesser evil reasoning need to make sure that they do not apply it to unrankable alternatives, and the burden of proof that options can be compared on an ordinale level lies upon those who use this type of reasoning. Unfortunately, many authors are apparently either not aware of the incommensurability problem or they tend to naively assume that alternatives are rankable in terms of their value.

With those preliminaries out of the way, I now approach the main topic of this paper. In the remainder of this essay I shall assess lesser evil reasoning under different conditions, starting with reasoning under certainty.

2 Lesser Evil Reasoning Under Certainty

Practical reasoning is said to be under certainty if the arguer knows, at least for practical purposes, what the outcomes of the available alternatives will be. In this section, I shall discuss lesser evil reasoning of this type.

2.1 Unidimensional Reasoning Under Certainty

The simplest kind of lesser evil reasoning under certainty is unidimensional reasoning, that is reasoning when the outcomes of both evil alternatives can adequately be described in terms of a single attribute. Consider for illustrative purposes the case (admittedly rare) of a pregnant woman whose life is endangered by uterine cancer and who will die unless a hysterectomy (the surgical removal of the woman’s uterus) is performed. According to advocates of lesser evil reasoning, we are faced here with two destructive alternatives. Hysterectomy leads to the death of the fetus, but if surgery is avoided both mother and fetus will die. Since two deaths are worse than one, it has been argued that hysterectomy is the lesser evil and is therefore the act that ought to be chosen.16

The logical structure of this type of lesser evil reasoning is clear. From performing hysterectomy it follows logically a consequence that is better than the consequence of the only alternative (no hysterectomy) and from this it is validly inferred that hysterectomy should be performed. For definiteness, let me put this in logical notation. Let h stand for the statement “hysterectomy is performed,” o for “one human being dies,” and t for “two deaths occur,” and let o > t indicate that one death is better than two deaths. The premises of the lesser evil argument under consideration are then h → o, ¬h → t, o > t and from these premises it follows that hysterectomy ought to be performed.17

The point to emphasize now is that valid reasoning is not ipso facto good reasoning. In my view, it is only a necessary condition for it.18 In fact, the case under consideration reveals two common pitfalls of lesser evil reasoning, which I shall explain now in turn.

(i) A main pitfall to be avoided in lesser evil reasoning is boxing oneself in with two alternatives. This mistake in reasoning is promoted by the very nature of lesser evil reasoning because it suggests that we compare two alternatives and choose then the better one. If authors employ lesser evil reasoning they do not tend to think much about new options. They seem to assume that they know the alternatives open to them and, as a result, many of their decisions are made from an overly narrow set of options. But almost always there are more alternatives available and failing to identify them is a mistake in reasoning.

With regard to our example, we can assume that the available alternatives are not only hysterectomy or no hysterectomy. It might be possible to postpone surgery until the fetus is viable. That is to say, there are not only two alternatives, from which the lesser evil must be chosen, but three; and the right choice is not any longer the lesser of the original two evils. It is, rather, the least evil of the three alternatives that are now available.

(ii) A second common pitfall of lesser evil reasoning is ignoring its defeasibility. It is a characteristic feature of practical reasoning in general that it is defeasible or, in the technical jargon, non-monotonic. That is, if a conclusion follows validly from given premises, it need not follow from a larger set of premises, even if the original premises are included. In other words, saying that a conclusion follows defeasibly means that it follows relative to a given set of options and their outcomes. Considering additional alternatives or new consequences may “defeat” the argument. Let a single example serve to illustrate this: You want to fly to Vienna tonight, and you can get seats on Austrian Airlines and Lufthansa. Since the fare on Austrian Airlines is lower (and there are no other differences) you conclude that you should take this airline. But then you learn that KLM offers a flight to Vienna also, for an even lower fare than Austrian Airlines (again, by hypothesis there are no other differences). This additional premise “defeats” the original argument. It no longer follows that you should take Austrian Airlines; the new conclusion is that you should take KLM.19

It is particularly important to note that lesser evil reasoning is defeasible. Many ethical discussions are confused because debaters tend to argue that one alternative is the lesser evil and seem to think that settles the matter. Apparently, it does not occur to them that their reasoning can be defeated by new consequences not yet considered. To return to our example above, let it be assumed that a further medical examination has revealed that the pregnant woman will not survive hysterectomy. This new information is presumably a “defeater” of the original argument because both options lead now to the death of the woman and her fetus and undergoing hysterectomy causes further excessive burdens on the woman. Arguably, avoiding the removal of the uterus is now the lesser evil; and the new conclusion is that the alternative “no hysterectomy” is the right choice.

2.2 Multidimensional Reasoning Under Certainty

It is almost a truism that practical problems frequently include multiple and often conflicting values. Reasoning about such problems is multidimensional because more than one attribute is needed to describe the outcome of the alternatives. For illustration, consider Clark’s (2006) study of HIV testing policies in Botswana (and in sub-Saharan Africa in general). Mandatory testing, he says, has the negative effect of infringing upon women’s right to privacy, but would dramatically decrease mother-to-child-transmission of HIV. Voluntary testing, on the other hand, does not violate human rights, but it leads to HIV infection of many infants and their subsequent deaths. He then argues that mandatory testing would be the lesser evil because the death of many infected children (as a result of voluntary testing) is worse than the violation of the right to privacy (if mandatory testing was adopted).20

(i) This example illustrates a third common pitfall of lesser evil reasoning: overlooking crucial consequences of the alternatives. In their attempt to determine the lesser evil, authors tend to focus only on some (usually negative) consequences of the available alternatives and ignore their positive outcomes. This is a mistake in reasoning because practical problems with multiple consequences cannot be resolved by focusing on some consequences only.

In more general terms, suppose that we have two alternatives a and b, the consequences of a are (c a1 , c a2 , …, c an ) and the consequences of b are (c b1 , c b2 , …, c bn ). Suppose further that c a2 and c b2 are bad and c a2 is worse than c b2 . We do now not compare c a2 to c b2 , arguing that b is the lesser evil than a because c a2 is worse than c b2 . To determine the values of the alternatives a and b and to find out which one is the lesser evil, we rather need to compare the compounds (c a1 , c a2 , …, c an ) and (c b1 , c b2 , …, c bn ). The fact that c a2 is worse than c b2 does not warrant the claim that b is the lesser evil because the value of both alternatives depends on their entire set of consequences.

To return to our example, Clark (2006) should have compared the positive and negative consequences of mandatory and voluntary testing. If (and only if) the consequences of both mandatory and voluntary testing are, all things considered, negative (which renders both of them evils) and the outcome of voluntary testing is on balance worse, then he would have been justified to consider mandatory testing a lesser evil than voluntary testing.

(ii) Other writers fall into the related trap of regarding an act as an evil because it has some negative consequences. To illustrate this error in reasoning, consider the example of condom use as a means of preventing HIV infections. Traditionally, the Catholic Church considered condoms as an inappropriate means in the fight against the AIDS pandemic. Recently, however, some bishops held that their use would be the “lesser evil” when it could preserve people from HIV infection.21 If the use of condoms can prevent infection with a potentially fatal disease why is it nevertheless regarded as an evil? In the Catholic tradition, it is an evil because it frustrates the finality of a natural (God-given) faculty and is therefore contrary to the natural purpose of sexual intercourse. That is to say, the use of condoms is an evil because of this putative (non-causal) bad consequence of frustrating the finality of a natural faculty. But this is faulty reasoning because the value of an option must not be determined by one consequence only but by its entire set of consequences. The bad outcome may be outweighed by the positive outcomes and the act is then, all things considered, not an evil (and therefore a fortiori not a lesser evil) but a good.22

3 Lesser Evil Reasoning Under Uncertainty

To this point we have dealt with lesser evil reasoning when, for practical purposes, we can know the consequences of each alternative before acting. But certainty is hard to come by, and most of the time, we make do with less. We must, therefore, turn now to the more realistic situations in which we do not know what the consequences will be until after acting—no matter how much time and thought we expend. Such decisions are said to be made under uncertainty.

3.1 Probabilistic Reasoning Under Uncertainty

Ethical decision problems exist normally in an uncertain environment, and not knowing exactly what would happen if a particular course of action were to be adopted adds a new layer of complexity to lesser evil reasoning because it makes it more difficult to determine the lesser evil. But in many situations uncertainty is unavoidable and effective decision-making requires then confronting it by judging the likelihood of different outcomes and assessing their possible results. In this section, I explore probabilistic lesser evil reasoning, that is reasoning when we do not know the consequences of our options before acting, but we can say of the outcomes of each of our options how likely it is.

Let us consider a simple example to illustrate lesser evil reasoning under such conditions.23 Sixty-eight-year-old Mr. Goldman had lost the vision in his left eye and now his right eye has deteriorated due to a cataract. Even with glasses, his corrected visual acuity is now only 20/50. His doctor raised the possibility of cataract surgery and estimated that the chances of successful surgery would be 80%, which would restore his visual acuity to 20/30 (corrected). However, unsuccessful surgery would erode his vision to 20/100 (corrected). The figure illustrates the structure of Goldman’s decision problem. Open image in new window

Goldman considers his two options “cataract surgery” and “no cataract surgery”. Surgery has an 80% chance of a fairly good consequence (20/30 acuity) but involves a 20% risk of a very bad outcome (20/100 acuity). Given this bad possible consequence, he regards the alternative “cataract surgery” overall seen as an evil. On the other hand, if he chooses “no surgery”, his vision remains bad (20/50 acuity), which means that this option is also bad. He is therefore in a dilemma and has to choose between two evils.

Although the logic of practical reasoning under uncertainty is more controversial than the logic of reasoning under certainty, it is clear that determining the lesser evil in a situation like this requires weighting the values of the outcomes with their respective likelihood. That is to say, to estimate the (dis)value of the alternative “cataract surgery”, Goldman must weight the value of the consequence (20/30 acuity) with its likelihood (80%) and combine this with the weighted (20%) badness of the consequence (20/100 acuity). As a result of this combination, he can determine how bad the alternative “cataract surgery” is and he can then compare it to the alternative “no cataract surgery” to find the lesser evil.

A vital point to notice here is that the value of the alternative “cataract surgery” is only an expected value. Suppose that Mr. Goldman opts for surgery because he thinks it is the lesser evil. The outcome is then either positive (20/30 acuity) and in this case surgery was no evil at all, or it is very bad (20/100 acuity) in which case surgery was the greater evil. “Cataract surgery” should therefore be called the lesser expected evil. A lesser expected evil may turn out to be no evil at all or it may be the greater evil. Nonetheless, when we reason probabilistically we never know what the outcome of our options will be before acting, and therefore it is reasonable to employ what we may call the “principle of the lesser expected evil.” That is, the principle that when we have to choose between two options under conditions of probabilistic uncertainty, the lesser expected evil ought to be chosen.

(i) A further difficulty of lesser evil reasoning that needs to be addressed in this connection is the epistemic pitfall of asserting unjustified premises. So far in this paper, I was mainly concerned with the logical assessment of lesser evil reasoning. A purely logical evaluation is concerned with the logical structure of arguments, but it is indifferent to the truth or justification of the premises. However, logically valid arguments do not necessarily provide a reason for believing the conclusion. A simple example will suffice to make this clear. “Everybody is happy; therefore Jones is happy” is a valid argument, but the premise provides no reason for believing that Jones is happy because it is unjustified and false.

All too often in lesser evil reasoning, authors make claims about the consequences of available options without providing evidence that the asserted (deterministic or probabilistic) relation between an alternative and its consequences really holds. But when we employ lesser evil reasoning we want to give reasons why one alternative ought to chosen rather than the other; and this requires that we justify our premises. If Mr. Goldman (in our example) holds that he has a reason for choosing surgery he needs to have a justification for believing that the success rate of cataract surgery is really 80% and that his visual acuity will be restored to 20/30 if it is successful.

An epistemic assessment concerns the agent’s justification in believing the premises of his reasoning; and it should be noted that lesser evil reasoning is epistemically defective if the arguer draws a conclusion but has no sufficient justification for believing the premises.24 There are, of course, many problems in determining just when the premises that occur in a piece of practical reasoning are justifiably believed, but these problems are not peculiar to lesser evil reasoning and I shall therefore not pursue this issue here.

(ii) Holding that divorce is a lesser evil than continuing marital strife implies not only the claim that both divorce and marital strife are evils, but also that the latter is the greater evil. In short, lesser evil reasoning implies value judgements, which need to be justified.25 That is to say that lesser evil reasoning is valuationally faulty if the arguer does not provide reasons for his valuations also.26 Ignoring this means falling into the pitfall of making unjustified value judgements.

Some examples will make this point clear. Many authors hold that abortion is the lesser evil when the life of a pregnant woman can only be saved if the fetus is aborted. But some claim that abortion is the greater evil even in such cases.27 The disagreement is here not about the consequences of the different options but about the values of these consequences. Both sides disagree on whether the direct killing of the fetus is worse than letting the woman and the fetus die; and resolving this disagreement requires reasoning about the values involved.

This can also be illustrated by an example discussed by I. Kant in his notorious essay “On a supposed right to tell a lie from altruistic motives.”28 He describes there a scenario where a murderer is pursuing an innocent man who takes refuge in your house to escape the attempted murder. The murderer, however, expects that the man may be in your house and asks you if he has taken refuge there. According to Kant, you have to tell the murderer the truth. Apparently, Kant regarded giving away the hide out of the innocent man a lesser evil than lying. As the discussion of this case in the philosophical literature shows, most authors disagree. Once more, the disagreement is about values, and justifying that truth telling (or lying) is, in the circumstances, the lesser evil requires giving reasons for the valuations involved.

To put the point succinctly, lesser evil reasoning that is based on unwarranted valuations does not provide a reason for the conclusion—even if it is logically and epistemically convincing.

3.2 Nonprobabilistic Reasoning Under Uncertainty

We are often facing problems that are so vague that we cannot assign probabilities. That is the extreme end of uncertainty. In Goldman’s case (3.1) it might be extremely difficult to settle on a set of numerical probabilities and therefore we may need to rely on nonprobabilistic decision criteria, which are based solely on the value of the outcomes and ignore their probabilities. In a nonprobabilistic analysis of Goldman’s case, we need to compare the two options “cataract surgery” and “no cataract surgery”. The outcome of “no surgery” is, for practical purposes, certain (20/50 acuity). The alternative “cataract surgery” has two possible outcomes, the good (20/30 acuity) if surgery is successful and the evil (20/100) if surgery is not successful. Ho, the probabilities of these outcomes are unknown.

This raises the question of the logic of deciding in such circumstances. There is much disagreement on this. Various criteria have been proposed, but none has firmly established itself. The rationale underlying the so-called Maximin Principle, which is often recommended for such cases, is to find for each alternative the worst outcome and then choose the option for which the worst outcome is best. If we apply this principle to Goldman’s case, then he should not undergo surgery because the result of “no surgery” (20/50 acuity) is better than the worst possible result of the alternative “surgery” (20/100). The option “no surgery” is therefore the lesser evil.

However, the exclusive focus on the worst outcomes has caused much criticism in the literature and it is all too easy to construct examples for which the Maximin Principle gives unreasonable results. Some authors have therefore recommended the so-called Maximax Principle. According to this principle, the decision maker should find the best outcome for each action and then choose the action for which this best outcome is largest. On this view, Mr. Goldman should undergo cataract surgery because the best consequence of this option (20/30 acuity) is better than the outcome of “no surgery”.

It is worth observing that advocates of the Maximax Principle will not even regard the option “cataract surgery” as an evil because it has a possible good outcome and their focus is on the positive results only. This shows that in nonprobabilistic reasoning, it is even unclear whether or not an action is an evil. As we have seen, lesser evil reasoning is clear under conditions of certainty and it needs to be broadened to lesser expected evil reasoning in probabilistic reasoning. But in nonprobabilistic reasoning the notion of a lesser evil may not be applicable. This result is a further restriction on the scope of lesser evil reasoning.

The discussion in this section highlights a seventh pitfall to watch out for: over-focusing on the negative. Some advocates of lesser evil reasoning give exclusive attention to the downsides of their alternatives. They then compare the two alternatives with respect to their negative outcomes and argue that the lesser evil should be chosen. This, however, is faulty reasoning. In the case of the pregnant woman who needed hysterectomy (2.1), it was a mistake to compare the deaths only because both alternatives will have positive effects also, which need to be taken into account no less than the negative outcomes. The reasoning for mandatory HIV testing (2.2) was defective because it compared only the negative outcomes of the available options and ignored their possible positive effects; and arguing that Goldman should not have surgery because the consequence of this alternative is better than the worst possible outcome of its alternative is erroneous because of its exclusive focus on the worst consequences. Good practical reasoning requires seeking a balance among the competing alternatives, which requires that we lay out their good and bad consequences. Finding an appropriate trade-off is often difficult, but not trying to balance our options is a mistake in reasoning.


  1. 1.

    Aristotle holds that we must sometimes "as the people say, take the least of evils." (see the Nicomachean Ethics II, 9).

  2. 2.

    In De officiis (iii, 28) he quotes the saying that we should choose the least of two evils (minima de malis eligenda).

  3. 3.

    For example, Arnolds and Garland (1974) apply it to the problem of punishment.

  4. 4.

    It has, for instance, been held that democracy is the lesser evil than totalitarianism (see Dubil 2004); compare to this also Ignatieff (2004), Winkler (1990), and Bellamy (2007).

  5. 5.

    See, for instance, Allison and Saag (2006) who apply it to the problem of prescribing analgesics that have negative side-effects; compare to this also Brinker (2000) or Schover (2008).

  6. 6.

    Compare, for example, Kanniyakonil (2007) who regards the principle of the lesser evil as one of the biomedical principles and Clark (2006).

  7. 7.

    For example, BBC (28 July 1998) reports that in the UK "tolerance zones" for prostitutes are regarded as a lesser evil than the situation that exists now.

  8. 8.

    In the classic formulation of the PLE, this is expressed by the phrase "De duobus malis …" (see the introduction).

  9. 9.

    For instance, Kamm (1996, p. 196) seems to regard the false beliefs that students may acquire in her lectures as a lesser evil.

  10. 10.

    As I have indicated above, already the ancient Greek and Roman formulations hold that we have to choose the least evil (minima de malis).

  11. 11.
  12. 12.

    In the history of Catholic moral theology, lesser evil reasoning was originally limited to cases of counselling. For example, St. Alphonsus Ligouri (Eighteenth century) teaches that it is licit to counsel the lesser evil to someone who is determined to realize the greater one. Alphonsus regarded it therefore as permissible to induce someone to drunkenness in order to avoid a greater sin, such as sacrilege or homicide. Later, however, lesser evil reasoning was applied to a much broader category of problems. Compare to this, Bretzke (2007) and Donagan (1977, p. 153) who seems to accept this very narrow scope of lesser evil reasoning.

  13. 13.

    This passage has been approvingly quoted by Pope John Paul II in his encyclical Veritatis Splendor (no. 80); see to this also Keenan (2000). The encyclical letter Humanae Vitae is available from:

    In this connection, it should be noted that the group of Catholic theologians who have been labelled ‘Proportionalists’ (e.g., R. Mc Cormick and B. Schüller) are in agreement with the magisterium of the Catholic Church that there are moral absolutes. They only disagree on which rules are absolutes and how to determine moral absolutes. It was this disagreement which caused Pope John Paul II's criticism that proportionalism is "not faithful to the Church's teaching" in his encyclical Veritatis Splendor (no. 75–78); compare to this also Quirk (1997, pp. 504–506).

  14. 14.

    By ‘consequentialism' I mean normative ethical theories that hold that the moral rightness of actions is exclusively determined by their (causal or non-causal) consequences. On the other hand, I define deontological theories as theories that are not consequentialist. That is, they hold that at least some actions are not right or wrong because of their consequences. Deontologists may (but need not) accept that the rightness of some actions is determined by their consequences.

  15. 15.

    A similar view has been defended by May (2003) who is here in agreement with Germain Grisez.

  16. 16.

    See McCormick (1978, p. 27) who discusses a similar case.

  17. 17.

    Observe, that I could have used any of a number of phrases to express the conclusion because ordinary language does not uphold a sharp distinction between different deontic modalities. For example, I could have said just as effectively that hysterectomy has to be performed or that is must be performed.

    It should also be noted that in this study, I will take it for granted that lesser evil reasoning can be logically conclusive; that is, that we can validly derive a conclusion from its premises. This view is by no means beyond dispute. Many logicians and philosophers endorse it, but some writers disagree. Since I have argued for the validity of practical arguments elsewhere (see Spielthenner 2008), I shall not pursue this issue further here.

  18. 18.

    I am grateful to an anonymous referee at this journal for drawing my attention to the problem that there is no agreement among authors whether validity is a necessary condition for good reasoning. But nothing substantial in this essay hinges on this issue and therefore we need not settle it here.

  19. 19.

    Philosophers have long been aware of this feature of practical reasoning and have therefore held that practical reasoning is only prima facie valid or that it provides only pro tanto reasons or ceteris paribus reasons—that is reasons "other things being equal."

  20. 20.

    See Clark (2006, p. 8). I have presented Clark's argument in a slightly simplified but, I think, not distorted form.

  21. 21.

    See, for instance, the Belgian Cardinal Danneels, according to "Catholic World News" (10 March 2006). Accessed 12 Sept. 2008. Compare to this also Cardinal Carlo Maria Martini who held that the use of condoms can certainly "constitute in certain situations a lesser evil"—Los Angeles Times (22 April 2006). Accessed 10 Aug. 2008.

  22. 22.

    This very common mistake has also been made by Anthierens et al. (2007) who consider prescribing benzodi-azepines (drugs used as anxiolytics) as an evil because of their negative side-effects, even though their prescriptions has, of course, positive effects also (e.g., it reduces the burden of suffering). Once more, whether prescription of benzodiazepines is an evil depends on its positive and negative effects and if the positive effects outweigh the bad effects it is, overall seen, not an evil but a good.

  23. 23.

    The example has been adapted from Hammond et al. (1999). Smart choices. New York: Broadway Books, pp. 131–133.

  24. 24.

    In passing, be it mentioned that this is exceptionally difficult for proponents of deontological theories. Ross (1930), for instance, may hold that breaking a promise is a lesser evil than treating someone unjustly because the prima facie duty of justice overrides, in the circumstances, the duty to keep one's promise. But to justify this he can only rely on intuition, which is a quite unreliable method of justification. And a formidable task lies ahead of theologians who try to justify that the use of condoms is an evil because it frustrates the finality of a God-given natural faculty.

  25. 25.

    I ignore here the crude emotivistic view that value judgements cannot be justified because they only serve to express and arouse affective states.

  26. 26.

    So-called instrumentalists will disagree. Instrumentalism can roughly be defined as the view that acting rationally means acting in a way that contributes to the agents intrinsic desires whatever they are. This implies that these valuations need not be justified. But many philosophers have critiqued instrumentalism at length, and I therefore will not deal with this view here.

  27. 27.

    For instance, Pope Pius XI (1930) writes in his encyclical letter Casti Connubii (no. 64): "[H]owever much we may pity the mother whose health and even life is gravely imperilled in the performance of the duty allotted to her by nature, nevertheless what could ever be a sufficient reason for excusing in any way the direct murder of the innocent [fetus]?" Accessed 17 Oct. 2008.

  28. 28.

    The essay "Über ein vermeintes Recht aus Menschenliebe zu lügen" was first published in the Journal Berlinische Blätter 10, Blatt (6 September 1797, pp. 301–314) and it appears now in vol. 8 of the Royal Prussian Academy edition of Kant's works, pp. 423–430.


  1. Allison, J.J., and K.G. Saag. 2006. Unbalanced regulation of over-the-counter analgesics: The lesser of two evils? Pharmacoepidemiology and Drug Safety 15: 410–411.CrossRefGoogle Scholar
  2. Anthierens, S., H. Habraken, M. Petrovic, and T. Christiaens. 2007. The lesser evil? Initiating a benzodiazepine prescription in general practice. Scandinavian Journal of Primary Health Care 25: 214–219.CrossRefGoogle Scholar
  3. Arnolds, E.B., and N.F. Garland. 1974. The defense of necessity in criminal law: The right to choose the lesser evil. The Journal of Criminal Law & Criminology 65 (3): 289–301.CrossRefGoogle Scholar
  4. Bellamy, A.J. 2007. Dirty hands and lesser evils in the war on terror. The British Journal of Politics and International Relations 9: 509–526.CrossRefGoogle Scholar
  5. Bretzke, J.T. 2007. The Lesser Evil. America: The National Catholic Weekly. Accessed 12 Nov. 2008.
  6. Brinker, J.A. 2000. De duobus malis minus est semper eligendum (of two evils the lesser should always be chosen). Pace 23: 384–385.Google Scholar
  7. Broome, J. 2004. Weighing lives. Oxford: Oxford University Press.CrossRefGoogle Scholar
  8. Clark, P. 2006. Mother-to-child transmission of HIV in botswana: An ethical perspective on mandatory testing. Developing World Bioethics 6 (1): 1–12.CrossRefGoogle Scholar
  9. Donagan, A. 1977. The theory of morality. Chicago: University of Chicago Press.Google Scholar
  10. Dubil, H. 2004. The lesser evil: Moral approaches to genocide practices. New York: Routledge.Google Scholar
  11. Griffin, J. 1986. Well-being: Its meaning, measurement, and moral importance. Oxford: Clarendon.Google Scholar
  12. Ignatieff, M. 2004. Lesser evil: Political ethics in an age of terror. Princeton: Princeton University Press.Google Scholar
  13. Kamm, F.M. 1996. Morality, mortality, vol. 2: Rights, duties, and status. Oxford: Oxford University Press.Google Scholar
  14. Kanniyakonil, S. 2007. The fundamentals of bioethics: Legal perspectives and ethical approaches. Oriental Institute of Religious Studies, India: Vadavathoor.Google Scholar
  15. Keenan, J.F. ed. 2000. Catholic Ethicists on HIV/AIDS Prevention. Continuum International Publishing Group.Google Scholar
  16. May, W. 2003. An introduction to moral theology. Huntington: Our Sunday Visitor.Google Scholar
  17. McCormick, R.A. 1978. Ambiguity in moral choice. In Doing evil to achieve good: Moral choice in conflict situations, ed. R.A. McCormick and P. Ramsey, 7–53. Chicago: Loyola University Press.Google Scholar
  18. Quirk, M.J. 1997. Why the debate on proportionalism is misconceived. Modern Theology 13 (4): 501–524.CrossRefGoogle Scholar
  19. Ramsey, P. 1978. Incommensurability and indeterminacy in moral choice. In Doing evil to achieve good: Moral choice in conflict situations, ed. R.A. McCormick and P. Ramsey, 69–144. Chicago: Loyola University Press.Google Scholar
  20. Ross, W.D. 1930. The right and the good. Oxford: Clarendon.Google Scholar
  21. Schover, L.R. 2008. A lesser evil: Prophylactic mastectomy for women at high risk for breast cancer. Journal of Clinical Oncology 26 (24): 3918–3919.CrossRefGoogle Scholar
  22. Spielthenner, G. 2008. The logical assessment of practical reasoning. SATS—Nordic Journal of Philosophy 9 (1): 91–108.Google Scholar
  23. Winkler, H.A. 1990. Choosing the lesser evil: The german social democrats and the fall of the weimar republic. Journal of Contemporary History 25: 205–227.CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media B.V. 2009

Authors and Affiliations

  1. 1.Department of Philosophy and Applied EthicsThe University of ZambiaLusakaZambia

Personalised recommendations