You are on page 1of 4

In Defense of Engineer Solutions of the Trolley Problem In my contribution, I discuss the trolley problem as a tool for illuminating and

deciding between the evaluated moral norms. I try to show that examples and similes make the discourse of norms rather more than less complicated. The existence and the status of norms are not thereby doubted, but I will argue for the relative precedence of meaningful situations and examples to (particular) norms (rules), rather than the other way round (precedence of universal rules to understanding of the situations). The trolley problem is a well-known ethical exercise. In one of its default versions: the main character is an observer standing by a switch lever for trolley trails. She/he sees a trolley wagon hurtling down the trail, it is beyond control, and there are five people standing in its way. The crash will kill them. The observer standing by the lever, however, has the possibility to switch the trolley and divert it onto a side-track. There is only one person standing, to be killed by the diverted trolley. Now, this classical thought experiment in ethics ends in the question what the observer should do. Is it (rather) right to switch the lever, or to do nothing? This problem serves as an illustration exposing possible clashes between ethical norms presenting themselves. In the trolley problem these are, roughly, as follows: 1. Five lives are more than one. It is right to choose the option that saves (or is beneficial) for more people, at the necessary expense of a smaller number of people. Hence, it is right to switch the lever, in order to save the five lives, if at the expense of sacrificing the one. 2. Every live is sacred in itself, regardless of number. To do intentionally anything that results in killing a person cannot be justified, despite the number of the putatively saved lives. Hence, it would be wrong to switch the lever; or vice versa, it is right not to switch it. The former line of reasoning comes from the tradition of utilitarian ethics, with its proposal that good and bad deeds can be measured against one other in terms of their (quantified) consequences. The latter line of reasoning is connected rather to the Kantian ethics which prohibits the reduction of any fellow human being to a means necessary for reaching an end considered as higher or greater good. The norms embodied in the two answers seem to be hardly compatible in this particular example. That is one of the reasons why this story is used as an instructive example in various introductions in ethics, illuminating the complicated nature of ethical issues. This practice itself, however, meets (philosophically) interesting problems. If the presentation is not directed to an obliging audience, the unwillingness to take the story in its intended, literal form can be considerable. Thus students of engineering science (for instance), when presented with the trolley problem within an Introduction to Philosophy, may be prone to solutions like why, if I break the lever, the rail switch will be blocked, which stops the trolley and saves everybody. It takes a great effort to explain to them that the particular face value of the example is irrelevant, as it should have served only as an illustration of a more general philosophical problem. In short, the young future engineers were not asked to find out the way of stopping the trolley, but to find out what ethical norm is relevant in situations of such ethical dilemmas. Unfortunately, they somehow refuse to understand that. (I heard Anders Sandberg report this teacher experience.) The reason may be that the example itself is not quite satisfying, and philosophers thus felt the need to introduce further details or variants to make the relevant norm more obvious. The most popular of these additions presents the fat man on the bridge: the observer sees the

hurtling trolley from a bridge above the trail, having a fat man by her/his side. Now the vehicle can be stopped by throwing the fat man down into its way. Even the fat man option can be further qualified: what if the fat man is someone particularly evil (Adolf Hitler)? What if there are little children among the five people in danger? The variety of interpretations of the trolley problem, not to mention the variety of its elaborations is enough to show that the problem is in itself not conclusive enough. It i) cannot show which of the concurrent norms is more relevant, because ii) in terms of the trolley problem, both are applicable (typically on the basis of adding a qualifying detail), but none so that it convincingly rules out the other one, once and for all. This is strange, since in ethics concurrent norms should be liable to evaluation by reflection itself. An example should not be necessary, but if we decide to introduce one, it should facilitate the decision or, at worst, leave the decision equally difficult, not more But the trolley problem fails to facilitate anything. If the trouble is not with the evaluated rivaling norms, can it be with the example? The trolley example apparently lacks sufficient clarity to elucidate or facilitate the decision in the difficult issue. The example seems to require further specifications in order that we can interpret it to shed light on the relevant norm. But with further and further specifications of the example designed to illuminate the issue, it only becomes more and more complicated. Two questions hence arise here: 1. Why is the matter itself difficult to decide in the general form of weighing the norms against one other? 2. Why is the first exemplification (the default trolley problem) and its further additions (the elaborated variants, the additional questions) expected to help to decide the weight of the concerned norms, when they offer only more and more detailed complications? An answer (more or less in line with the engineer-like suggestions) can be the following point: Our ethical reasoning does not begin with an abstract, general reflection upon universal moral norms (laws), but tries to orient within various live situations. The situation presented in the trolley problem has to be further determined, equipped with more details so that it resembles a real life situation; only then one can decide what to do without considerable hesitation. Then one can even justify the decision as confirming to one rule or other. Neither the specified examples, however, can be related to the norm quite directly. Consider the fat man variant of the problem. Throwing the fat man down can be interpreted as conforming to the same norm as switching the lever and thereby diverting the trolley to the side-track. In both cases I prefer the lives of the five to the life of the one. Yet many people who, when asked what they would do, lean to switching the lever are unwilling to throw the fat man down. The likely answer is: it is not the same to kill someone actively and to prevent killing of the five by an action that may result in killing someone later. The implicit arguments standing behind this reserve can be at least twofold: 1. Under further scrutiny, the respondent reveals that the difference in her/his decision is underpinned by difference with respect to the norm operating in the background. It is not the same, to directly kill one for saving the five, and to perform only an operation which is not an act of direct killing. The opening utilitarian intuition is tinged by the Kantian argument: the utilitarian agent feels uncomfortable with the thought that she/he exploits another persons life as a means for an end (though noble). In the former version of the story, the means is the lever, while the one persons life is not, as it were, a part of the act.

2. Under further scrutiny, the respondent reveals that the difference in her/his decision is underpinned by difference with respect to the result of the act. It is not the same because unlike the fat man doomed to die, the person standing on the side-track could perhaps jump aside, or could survive the clash with the trolley because she/he stands in a greater distance than the five and the vehicle could perhaps slow down by that time, etc. In the latter case, we witness again an attempt to cheat oneself out, as it were, of the exact wording of the example. I enter further details into my reading of the situation so that I could say: there is a way I could act without killing anyone. The engineer solutions of the trolley problem blocking the rail switch, or many more options the young engineers are able to offer are of similar kind. These suggestions, however, should not be understood as attempts at cheating out of the problem. They are attempts to make the example comprehensible to an agent who is on the one hand not a trained philosopher, but on the other hand someone with a particular history and a unique sum of personal, familiar, cultural, historical, professional, etc. experiences. The philosopher objection is: but this really is cheating. The point of the trolley problem is to construe a situation of a dilemma and on this background to illuminate the problem of ethical norms that can be incompatible. The cunning engineers only deny that there is a dilemma. But dont dilemmatic ethical situations really occur in our lives, quite commonly? Yet posing the trolley problem as a paradigm for ethical dilemmas misses the point. Dilemmatic situations are such where the agent has already investigated all the possibilities she/he has been able to see and has not found any one acceptable. The abstract, non-specific outline of the trolley problem is far from having all the possibilities covered. When we are, having investigated all the options and additional details of a situation, not capable to seeing any desirable, we make a difficult decision. The difficult decisions we make in reality are difficult because we feel despite all the moral reasoning we perform that there remains at least one rule that we have not ruled out as irrelevant or secondary, and yet we violate it. All the weighed details of the situations comprise a situated logic the decision does have, despite the fact that it violates an acknowledged rule. In other words, despite the violation of a rule, the committed action can be reasonably explained and perhaps justified. The explanation contradicting the authority of a rule does not take shape of another, perhaps more specified rule, but of listing the particulars of the situation. (In explanations like I could not refuse Jacks request to help him rob the bank /I ought to have helped him/, because I have known him for more than 20 years, since our childhood, and we have experienced a lot together I appeal to no norm, yet my actions are thereby explained and some may even say justified.) I take such examples as reminders (rather than direct arguments) that situations falling under a general rule in an unequivocally assessed way are rather rare and mostly are abstractions. Once we remove finally the underdetermination of the concerned situation so that we can say the best and the most right thing to do here is this-and-this, because this-andthis, let us pose the question in what other situations is the identical justification because this-and-this equally and equally meaningfully applicable? Let us imagine here the above mentioned because I have known Jack... This explanation establishes certain rule, but this does not mean that there are any other examples apart from the above one where because I have known Jack... is a meaningful explanation as well as justification. We will soon see that rules instituted this way do not reach far. Here the rule is not it is right to do anything Jack asks, because of the shared childhood, but not even it is right for me to do anything Jack asks, because of the shared childhood (not to mention it is right to do anything a friend from childhood asks, because of the shared childhood). It may be right

for me to rob the bank on the basis of this reason, but perhaps not to lend him money. So not to do anything. (Compare also It is not right that I divert the trolley to the side track because it is Jack standing there, and I have known him... with It is not right to stop the trolley by throwing a fat man down into its way, because the fat man standing by my side is Jack, and I have known him... Or It is right that I divert the trolley to the side track, because Jack is standing in its way, and I have known him... with It is right to stop the trolley by throwing a fat man down into its way, because Jack is standing in its way, and I have known him...) The question how anyone else but me could follow such rules only adds to that. I do not want to deny the existence of any universal norms (i.e. to argue that only individual rules can exist and exercise a power). Rather I think that the trolley problem and other such similes allow us to focus on the way rules pervade the life and the decisions of an individual. A rule can govern an individuals action only insofar as it appears meaningful to her/him, in terms of her/his present situation as she/he understands and experiences it. (For another person who is an observer, there may be a different rule making sense of the situation in a different way, resulting in a different moral evaluation.) We conceive of rules in very different sense when we expect them to be building blocks for a universalistic ethical system and when we follow them forming and making sense of the (moral) life of individuals within which they play a role. In this sense, the engineers refusing to answer the trolley problem in the prescribed philosophical way to for either answer and finding a uniquely personalized way to orient within the example may be those who understand the problem more properly. (It would be perhaps interesting to consider the status of the regimentations determining that and when one acts according to a utilitarian-like calculus. They are likely to act as, so to speak, context-specific regulations. E.g., commanding officers of army units are sometimes expected to make such decisions as sacrificing a number of their troops for the sake of successful action of other troops or ensuring their lives or the lives of the civilians. It is no wonder that decisions like that usually prove to be unsuitable for a subsequent moral evaluation: if someone says I had my mission that I had to accomplish, so I decided to give these three soldiers this task, it does not really answer the question was it right to give them this task that cost their lives? neither as yes, nor as no , but rather suggests that the question may be misplaced.) In short, trolley problem can show us that it may be quite legitimate to answer that one ought to do sometimes one thing and sometimes another (and that it neednt be stupid to try to find another, third option). The example in its default, non-specific form as it is usually presented does not include the reason deciding for one option. Various reasons are inherent to its various personalized readings. References D. Z. Phillips, Interventions in Ethics, Albany: SUNY Press, 1992. Rush Rhees, Moral Questions, Basingstoke: Macmillan Press, 1999. Anders Samberg, What do Cars Think of Trolley Problems: Ethics for Autonomous Cars. Presented at the conference Beyond AI 2013: Artificial Golem Intelligence (1214 November, Plze, Czech Republic). Judith Jarvis Thomson, The Trolley Problem, The Yale Law Journal 94 (1985): 1395-1415.

You might also like