Moral Intuitions in the Moral Luck Discourse

1. Introduction to the problem

The paradox of moral luck is usually presented in the literature as a conflict between two opposite moral intuitions. The former, termed the control principle (or the control condition), claims “we are morally assessable only to the extent that what we are assessed for depends on factors under our control”.² The latter can be expressed by denial of the claim above, i.e. if we accept that moral luck exists, we agree that “an agent can be correctly treated as an object of moral judgment, despite the fact that a significant aspect of what he is assessed for depends on factors beyond his control”.³

The problem with moral luck arises if we presuppose morality to be a non-contradictory set of beliefs. From there we should offer an explanation as to why one might hold two contradictory beliefs – that the control principle remains valid and that there is moral luck. Neither Williams⁴ nor Nagel⁵ offer such an explanation, and the most common

strategy among other philosophers is to deny one intuition and support the other.⁶ Adherents of the control principle argue that the intuition of moral luck is either mistaken or irrational, and that when we make moral judgments based on it, we actually have something else in mind. On the other hand, their opponents accept that moral luck affects our moral judgments, and they maintain that the control principle seems to be an implausible reflection of human nature. What is more, a third position can be adopted. We can face the fact that there are two equally strong intuitions, rendering ourselves unable to mitigate the conflict that exists between the two.⁷

Although I will not provide any satisfactory solution to it, an attempt will be made to establish a basic framework for discussion. The main aim of this study is to explain, with some support from modern moral psychology, 1) how we should understand the term “moral intuitions”; 2) what the nature of our moral intuitions is as concerns moral luck; 3) whether our moral intuitions are reliable; and, finally, 4) whether metaethical considerations might help us understand these problems. Although one may find these answers not fully conclusive, I believe at the end of the day they provide substantial insight into understanding the above problem.

2. Moral intuitions in the psychological sense

2.1. Social Intuitionist Model

When we make moral judgments on particular cases or abstract moral dilemmas, we often follow some sort of intuition. The Social Intuitionist Model (SIM), a psychological theory proposed by J. Haidt, holds that these intuitions should be understood as

judgments, solutions, and ideas that pop into consciousness without our being aware of the mental processes that led to them. When you suddenly know the answer to a problem you’ve been mulling over, or when you know that you like someone but can’t tell why, your knowledge is intuitive.⁸

Moral intuitions (intuitions in the psychological sense will be labeled, after Antti Kauppinen,⁹ as intuitions*, in contrast to intuitions in the philosophical sense – intuitions^) are just a subclass of intuitive knowledge “in which feelings of approval or disapproval pop into awareness as we see or hear about something someone did, or as we consider choices for ourselves”.¹⁰. The background of such an account of intuitions* is the so-called dual-process theory, which divides the processes occurring in the human brain into two groups: conscious (most often characterized by slowness, controllability, sequentiality, requiring effort, manifesting in recent – and only human – stages of evolution) and unconscious (characterized by speed, uncontrollability, simultaneity, associativity, occurring quite commonly among animals in the early stages of evolution).¹¹ Therefore, moral intuitions*, which seem to be

grounds for our moral decisions, are results of the latter processes. By contrast, conclusions of moral reasoning, defined as “conscious mental activity that consists of transforming given information about people in order to reach a moral judgment”¹² are results of the former processes. Moral reasoning serves only to provide post factum justification for already-made moral judgments; it acts like a lawyer or a press secretary whose job it is to defend his or her boss’ decisions. I will not determine here whether the SIM account is correct; I maintain only that it is highly probable our actual moral judgments are largely based on intuitions*. It is presumed that more research will be devoted to the subject in the future.

2.2. Moral heuristics

One of the approaches to understanding the nature of moral intuitions* is the heuristic-based methodology. According to a broad definition adopted by C. Sunstein, “Heuristics include any mental short-cuts or rules of thumb that generally work well in common circumstances but also lead to systematic errors in unusual situations”.¹³ A narrow definition states that heuristics operate by unconscious attribute substitution. W. Sinnot-Armstrong and his colleagues¹⁴ explain it in the following way:

A person wants to determine whether an object, X, has a target attribute, T. This target attribute is difficult to detect directly, often due to the believer’s lack of information or time. Hence, instead of directly investigating whether the object has the target attribute, the believer uses information about a different attribute, the heuristic attribute, H, which is easier to detect. The believer usually does not consciously notice that he is answering a different question: ‘Does object, X, have heuristic attribute, H?’ instead of ‘Does object, X, have target attribute, T?’ The believer simply forms the belief that the object has the target attribute, T, if he detects the heuristic attribute, H.

So if we ask the question “Which city is bigger, Washington or Memphis?”,¹⁵ the person to whom the question was directed will try to answer the question “Which city is more familiar to me?” The target attribute in answers concerning moral blameworthiness is moral wrongness,¹⁶ the nature of which is embroiled in a long-standing controversy about the nature of wrongness. Similarly, we often do not try to answer the question “Is the drunk driver morally blameworthy for killing a person in an accident?” but a rather different one – one that concerns our feelings of outrage, because the answer to the former question is typically epistemically inaccessible (i.e. we do not know whether the pedestrian ran into the street, or whether the brakes of the car broke, and perhaps a sober driver would not have prevented the accident). Even if we assume that some moral dilemmas have objective answers (this does not necessarily imply any metaphysical assumptions), we must admit that they are highly inaccessible.¹⁷ This is probably the reason why we use moral heuristics to make moral judgments, even if using them exposes us to the risk of failure.

Therefore, may using heuristics be rational? There are two positions that are prima facie opposites. On one hand, as W. Sinnot-Armstrong et al.¹⁸ write, “Kahneman suggests that people who use heuristics exhibit a kind of irrationality insofar as their responses violate rules of logic, mathematics, and probability theory”.. On the other hand, Gerd Gigerenzer and his followers place emphasis on the fact that employing heuristics leads to the best solution available in typical environmental conditions and that they are “fast and frugal”.¹⁹ However, while proponents of both views agree that decision-making processes based on heuristics in unusual circumstances might be responsible for mistakes, they can lead to very effectual results under normal conditions. This disagreement is apparent because both parties have a point. Heuristics are irrational, if we consider that the rational method of decision-making must observe principles of logic and thus lead to undeniable and fully-justified conclusions. Nevertheless, if we assume that it is reasonable for the being to adopt even slightly irrational strategies which will increase the efficiency of our operations and reduce their costs, then, in this sense, heuristics seem to make us work smarter. If we use heuristics in almost every area of life, why not also use them in moral judgments? Indeed, it is highly probable that we do.²⁰

3. Moral intuitions in the philosophical sense

3.1. Moral intuitionisms

Although there is a temptation for psychologists to equate intuitions^ with intuitions*,²¹ it seems clear that ethical intuitionists have something else in mind when they talk about moral intuitions. But there is a long-standing controversy and disagreement on their nature. Kauppinen distinguishes three types of moral intuitionism: (a) self-evidence intuitionism; (b) seeming-state intuitionism; and (c) coherentism.²² The first account is represented today by R. Audi, who maintains that moral intuitions^ have to meet four conditions: 1) they must be non-inferential, so “the intuited proposition is not – at the time it is intuitively held – held on a basis of a premise”²³; 2) they must be the result of “moderately firm cognition”,²⁴ i.e. beliefs should be hard to overcome by doubts or counter-evidence; 3) their holder should have at least a minimal understanding of their content; and 4) they should be independent of any former theories and cannot be theoretical hypotheses themselves.²⁵ Therefore, Audi’s moral intuitions^ are beliefs that are non-inferentially justified simply by their self-evidence. For W.D. Ross, some examples of such intuitions^ are prima facie moral principles.²⁶ As Kauppinen distinguishes, new versions of self-evident intuitionism (e.g. that of Shafer-Landau) neither need to appeal to special intuitive cognition nor claim that intuitions^ must be self-evident for

The second conception of moral intuitionism considers moral intuitions^ as intellectual appearances. Michael Huemer defines them as follows:²⁹ “An intuition that p is a state of its seeming to one that p that is not dependent on inference from other beliefs and that results from thinking about p, as opposed to perceiving, remembering, or introspecting.” An example of such p may be the statement, “If A is better than B and B is better than C, then A is better than C.”³⁰ Before we think about arguments for and against this statement, we can say that it seems to be true, unlike a proposition such as “Assisted suicide is immoral”, which is controversial from the beginning. What is important is that the state of an utterance seeming true is a result of just thinking about the statement. Kauppinen calls it “non-doxastic, presentational, and compelling experience”.³¹ The third moral intuitionist theory is the so termed coherentism, which has been primarily represented by John Rawls with his famed idea of reflective equilibrium (RE). In its wide variant, there are three components: a set of moral principles; a set of considered moral judgments, which are primarily moral judgments on particular cases; and a set of background theories, both scientific and philosophical. Equilibrium should be the end of a deliberative process in which we revise our initially credible, considered judgments so that they “cohere not only with substantial moral principles but also with social science and with ideals of the person and of society”.³² After presenting these two approaches, psychological and philosophical, the question remains: What is the relationship between moral intuitions* and moral intuitions^? Although I must admit I do not have a precise answer and I believe much more psychological research and philosophical self-reflection is needed here, we can conduct some analyses. First, we cannot reduce intuitions to intuitions* – at least not all intuitions. Usually intuitions^ are not formed by emotional gut feelings, but rather by some kind of gradable understandings. Some people are better at this task, and some are not as good. Even if these intuitive beliefs are the result of unconscious processes, they can be confirmed by conscious, careful, firm reflection. We can argue about them and even be convinced that we were astray. Some, like Joshua Greene,³³ assert that “nonconsequentialist moral theory is epistemically unsupported”³⁴ because its premises are emotional nonconsequentialist intuitions* that are undermined as they are “sensitive to morally irrelevant features”.³⁵ This argument, which Kauppinen calls the aetiological debunking argument, is nevertheless far too simplistic if we are aware of the difference between intuitions^ and intuitions*.³⁶

Even more problematic is the epistemic status of moral intuitions^. They must be noninferentially justified. But what does this really mean? We can be adherents of the so-called epistemic liberalism, according to which “we are justified in taking things to be as they appear to be, unless we have sufficient reason to doubt their appearances”.³⁷ But there are plenty of doubts concerning the epistemology of moral intuitions^ that undermine their non-inferential epistemic justification, i.e. that which leads to truth or objective rightness.³⁸ They can be justified instrumentally – i.e. “if denying those intuitions would disrupt everyday life in harmful ways”³⁹ (cf. moral heuristics section) – if they form a feeling of belonging in their holders within the group. Such instrumental justification is, however, not what we hope for. On the other hand, there is a way out. Our moral intuitions* will not be susceptible to the moral skeptic’s argument if we adopt a nonlinear conception of justification.⁴⁰ In contrast to the linear theory, there is no need for belief A to be justified by belief B, and belief B by belief C, and belief C by belief D, and so on. The nonlinear conception of justification states that belief A can be justified by the conjunction of beliefs B, C, D, and probably many others. Hence, belief B is justified by the conjunction of belief A, C, and D, and belief C by beliefs A, B, and D. Such a position allows us not only to bypass the problem of regress ad infinitum – it is also resistant to the circularity objection. One of the variants of the nonlinear conception of justification is John Rawls’ aforementioned theory of reflective equilibrium.

3.2. Commentary

It is also important to draw a distinction between (a) first-level moral considerations involving making moral judgments, i.e. assessing whether a person or action is good or bad, praiseworthy, or blameworthy, and (b) second-level moral considerations concerning criteria for grasping and using moral concepts that underlie first-level considerations. Therefore, while at the first level, we think whether Smith is morally responsible and blameworthy for the death of Jones; at the second level, we analyze the very concept of moral responsibility and blameworthiness and the criteria of assigning them. In other words, the latter considerations are the metalevel for the former. Research on the above considerations can be divided into (a) descriptive, which focuses on identifying which moral judgments we actually make and the criteria for the use of moral concepts that we in fact apply, and (b) normative, which works to determine what moral judgments we should reach and which conditions of moral concepts we should adopt. Nonetheless, I do not believe that this distinction lies between the psychology of morality and normative ethics. Descriptive studies can surely include, for example, metaethical attempts to reconstruct the folk theory of moral concepts (the so-called modest conceptual analysis).

4. Moral luck and moral intuitions

4.1. Which level?

Let us come back now to the problem of moral luck, which was asserted to be a conflict of two moral intuitions. First, when we consider our judgments that concern moral luck

from the descriptive perspective, we can follow both intuitions^ and intuitions*. The latter are the object of many recent studies.⁴¹ The former, as claimed by moral intuitionists, can be the grounds for our moral judgments, but, as stated earlier, it is probably a much rarer case (descriptively). However, in philosophical literature on moral luck, we will find intuitions^ rather than intuitions*. To see how intuitions work in the case of moral luck, let us return to Thomas Nagel’s article on moral luck, which began this discussion. Nagel explicitly writes that “the problem develops out of the ordinary conditions of moral judgment”,⁴² which in other places are called “intuitively acceptable conditions of moral judgment”.⁴³ What are these conditions? Nagel claims that “prior to reflection it is intuitively plausible that people cannot be morally assessed for what is not their fault, or for what is due to factors beyond their control”,⁴⁴ which means that “without being able to explain exactly why, we feel that the appropriateness of moral assessment is easily undermined by the discovery that the act or attribute, no matter how good or bad, is not under the person’s control”.⁴⁵ On the other hand, Nagel observes that there is another: moral luck intuition, i.e. “a perception of one of the ways in which the intuitively acceptable conditions of moral judgment [control condition] threaten to undermine it all”.⁴⁶ In these passages, it is quite clear that Nagel treats control intuition and moral luck intuition as parts of second-level moral consideration. Before we make moral judgments about some case, we have to think if it is appropriate to make any moral judgments at all. For example, when we see a man stricken by lightning, we do not blame his neighbor for its happening because we believe that the lightning strike is completely independent of the neighbor’s realm of control; thus, he cannot be subject to moral evaluation. On the other hand, when we see one person shooting another, we have a second-level belief that we can form a moral judgment concerning the shooter because he had control over the situation. Only then can we move to the lower level – when we descend to this lower level, we wonder whether or not the shooter is blameworthy (e.g. his shooting may be justified if he were acting in self defense). This view that the control condition is formed at the second level of moral considerations seems to be supported in the majority of later literature on moral luck.. Just few examples follow: David Enoch and Andrei Marmor assume that “the condition of control has enough intuitive support and philosophical credentials that one would need very strong reasons to discard it”⁴⁷; Dana Nelkin, in her invaluable article in the Stanford Encyclopedia on Philosophy, admits that the “[...] Control Principle […] is intuitively compelling”⁴⁸; and, in her recent work, Diana Hsieh tries to “determine whether our conflicting intuitions about control and responsibility might stem from a faulty view of the requirements of moral praise and blame”.⁴⁹

Now the (descriptive) question concerns the kind of intuitions we follow at the second level of moral consideration: Do we follow intuitions^ or intuitions*? Although, on the basis of this one text, it is hard to decide what Nagel’s theory of moral intuitionism is in detail; he probably has in mind intuitions^ which, according to his theory, are pre-theoretical (“prior to reflection”) appearances (“the seemingly natural requirement of fault or responsibility”) that are not non-inferentially justified (“without being able to explain exactly why”). I think most philosophers share the view that intuitions^ are involved in the discussion on moral luck. However, there are some researchers who believe that our control principle and moral luck intuition are intuitions*.⁵⁰ We are not able to say which option is right, because it depends on, first, the epistemological reliability of different concepts of moral intuitions, and, second, further empirical research on intuitions. Therefore, I leave the matter open. An unequivocal decision is not necessary for the purposes of this article.

5. Normativity of moral intuitions

5.1. What is the “real” moral judgment we expect?

From the descriptive point of view, regardless of whether intuitions are intuitions^ or intuitions*, the normative perspective still remains – what criteria for moral judgment and responsibility should we adopt? Do intuitions, in this regard, have normative power? I believe not, or at least I doubt it. But let us start by answering the questions of which concept of moral judgment we accept, and what should be the “real” moral judgment. Kenneth and Fine note:

There are a number of ways in which philosophers might approach the question of which is the ‘real’ moral judgment. First we might ask which process, automatic or controlled, is most influential in the determination of moral attitudes or moral judgments. Second we can ask which process best answers to our concept of moral judgment. Third, we can ask which judgment has normative authority.⁵¹

The answer to the first question is sought by psychologists; the answer to the second relates to second-level moral considerations, which are interesting to both philosophers concerned with conceptual analysis and psychologists who study the actual use of concepts. The third question, however, seems to remain outside the scope of tools owned by psychology. I believe that neuroscience, psychology, or sociology can help us in determining what the motivation is or what the limits of human cognition are, but in the case of normative inquiry I stand on the side of Armchair Traditionalism rather than Ethical Empiricism.⁵². But what determines normativity? In this paper, I accept the approach presented by Kenneth and Fine, who refer to the theory of Michael Smith and John Deigh. Kenneth and Fine⁵³ cite Deigh, who claims that “One cannot coherently think, I ought to

do such and such, though there is no reason for us doing it.”⁵⁴ Smith discerned explanatory and motivating reasons.⁵⁵ When we ask for the explanation of why my neighbor calls me an immoral person, the explaining reasons will be the conscious and unconscious processes which led him to make such judgment, and the motivating reason could be that he is cautious of me, but also that he simply wishes to decry me. But justification of his claim and belief is something entirely separate from explanation. So my neighbor will not be justified in his judgment because he lacks factual reasons (e.g. he mistakenly thinks I have committed some act but actually I have not) or moral reasons (e.g. it is true that I have committed the act but my neighbor incorrectly believes it was morally wrong when, actually, it was permissible). When a father punishes his son for doing something bad, the son can always ask the question: why is it bad and why do I deserve punishment? If the father answers “It is bad because it is bad”, the son probably will not regard this justification as valid because it lacks moral reasons. In response to Haidt’s SIM, Kenneth and Fine write:

When an individual makes a moral judgment, it is plausible to suppose that the reasons implied or adduced in support of the judgment must be reasons which the agent herself (albeit perhaps mistakenly) takes to justify and not merely that explain the judgment. Otherwise the judgment can have no normative authority for her.⁵⁶

That is why intuitions* underlying our moral judgments can only explain why we made such a moral judgment, though they may not justify it. For example, if John, being a witness to a fatal car accident, forms a judgment that the driver killed the pedestrian, but did not check whether the cause of the accident was indeed a violation of traffic rules by the driver and not the failure of the brakes, we can explain why he issued such a moral judgment (he had an intuition*), but we cannot say that it was justified. The adherents of SIM seek to demonstrate that such an approach is not realistic – that people, in fact, do not act in this way; instead of making judgments based on reasons, they remain with feelings, or intuitions*, that unwittingly appear in their minds. It is probable that such a statement is empirically true, but let us focus on Haidt’s method of conducting research, which resulted in the discovery of the moral dumbfounding⁵⁷ phenomenon. Researchers expected from the respondents a justification of their moral judgments based precisely on justificatory reasons; thus, the presupposed concept of moral judgment, the “real” one (in the normative sense), was the same as that of Deigh, Kennet, and Fine. Even if intuitions* in SIM terms are responsible for how people make moral decisions, they are still not the reasons that determine the normativity of moral judgments. Moral intuitions* might have normative power, which needs however to be demonstrated first. Moral intuitions* are also objective reasons, and that in the SIM is missing.

5.2. The two-levels theory of moral thinking

But what about the normativity of intuitions^? Most ethical intuitionists assume, after all, that intuitions are normative so long as they are non-inferentially justified. On the

one hand, we return to the problem of moral skepticism and doubts about the certainty and reliability of our intuitions^. Even if we philosophically trust them, there is still the question raised by R.M. Hare: How do we choose between two equally justified intuitions^ for which there are two moral obligations that cannot be fulfilled at the same time? There is often no intuitive metarule that can help us make a non-arbitrary choice. To solve this problem, Hare delivered his metaethical theory of two levels of moral thinking.⁵⁸ His metaethical beliefs can be classified as a) non-descriptivism, b) weak cognitivism, and c) anti-realism.⁵⁹ This means that a) the meaning of moral judgments is determined not only by syntax and truth conditions but also by some normative component; b) although moral judgments can be true or false, making them correctly does not require “establishing epistemological contact with the independently constituted state of things”⁶⁰; and c) there are no moral values, facts, and properties which cannot be reduced to any non-moral natural properties and exist sui generis. Hare

divided normative moral reflection into two levels: critical and intuitive. In his view, normative moral theory should disintegrate into separate but complementary areas of considerations: the aim of the former is to look for the perfect standard method of moral thinking, and the aim of the latter to answer the question of to what extent and how a person, as a cognitively and psychologically limited being, should use this pattern.⁶¹

The levels correspond to two types of principles: prima facie principles, which are results of thinking at an intuitive level, and critical rules. The first of them expresses prima facie duties, such as “Do not kill”, “Do not lie”, etc. They are uncomplicated, which is a psychological limit, because they are usually used in common cases from everyday life and because we acquire them in the process of upbringing and education.⁶² Prima facie principles are simple because, like in the case of heuristics, we need “a practical guide […] unspecific enough to cover a variety of situations, all of which have certain salient features in common.”⁶³ They are also

an indispensable help in coping with the world (whether we are speaking of moral decisions or of prudential or technical ones, which in this are similar), namely the formation in ourselves of relatively simple reaction-patterns (whose expression in words, if they had one, would be relatively simple prescriptive principles) which prepare us to meet new contingencies resembling in their important features contingencies in which we have found ourselves in the past.⁶⁴

Although upbringing and education as well as life experience may be the possible sources of many moral intuitions. Hare could not know about such innate mechanisms developed during evolution which are compatible with his view on moral intuitions. The simplicity of prima facie rules becomes a problem when we encounter situations in which we face moral dilemmas. As Hare says:

Although the relatively simple principles that are used at the intuitive level are necessary for human moral thinking, they are not sufficient. Since any new situation will be unlike any

At this point, it is also necessary to introduce the “ultimate” level of moral thinking – the critical one. On this level, we attempt to establish critical rules, universal prescriptions, which can be complex and sophisticated because they are not limited by human psychological capabilities. Using such a rule, we can justify one correct solution for every moral conflict from the intuitive level. What is the relationship between prima facie principles and critical rules? Saja writes “critical thinking is epistemically more elementary than prima facie principles, i.e. it can be used as a tool for criticism and their changing”.⁶⁶ Hare claims that prima facie principles are designed by critical thinking to be the best reflection of critical rules for psychologically limited human beings. If humans were archangels (Hare’s characters that have ideally fast and precise mental processes, which is why they could only use critical thinking – also why this archangel character is able to make a moral judgment in even the most complicated case), prima facie principles would be unnecessary. Of course, there are no absolute proles (these Orwellian characters use only intuitive principles) and no absolute archangels; in each of us there is a little prole and a little archangel (the proportion of the two causes our moral judgments either to be based more on intuitions or more on critical thinking). Although Hare claims that both levels of thinking are rational, there is no single concept of rationality here. The set of prima facie principles is rational when it allows us to make the best moral judgments without wasting too much time and energy on rational reflection, where “the best” denotes the quality of being as similar as possible to the set of critical rules. Therefore, the resemblance to moral heuristics seems legitimate. The critical level is rational in a more classic way, as well as normative because we do not simply follow prima facie intuitive principles, but look for non-intuitive reasons. Its aim is to search for the one ultimate solution that would actually be right. But why should we rely on rational thinking? In order to reconcile feuding subjective affections, it is necessary to find a guarantee of objectivity. If we cannot fall back on the intuitive level, the only way, claims Hare, is to trust conclusions from the analysis of the logic of moral language. Perhaps it is not so fast and frugal; to the contrary, it is rather time- and energy-consuming, but at least it is definite.⁶⁷ Of course, Hare’s theory is not only a descriptive one, though he is aware of some features of human mechanisms of making moral judgments. It is also a normative model that tries to solve a problem which Hare notes in ethical intuitionism. Effectually, it grounds the making of final moral decisions not in moral intuitions, but in non-intuitive reasons. Of course, this theory can expose oneself to criticism in that the level of critical thinking is also based on certain intuitions, but this is another problem that will not be resolved here. The question remains, however, whether this model is useful for second-level moral considerations. Hare’s prima facie principles are rather norms from the first level of moral considerations – we decide on their basis whether someone is blameworthy or not and what moral decision we, ourselves, will make. Nevertheless, it seems that when it comes to the general

notion of the two levels of moral thinking, it can refer both to the moral considerations of the first level as well as the second level. On the second level, we also have to deal with certain principles – perfect examples include the control principle and moral luck intuition, which both seem to be valid and justified. The main conclusion from applying Hare’s theory to the problem of intuitions in the discussion on moral luck is that, unmodified and alone, it cannot provide us with definitive answers and decent justification for them. We need to develop some sort of conflict rule between the two intuitions (e.g. with the wide reflective equilibrium procedure) or look for conclusive reasons somewhere outside the intuitive level.

6. Conclusions

What is the image of the debate on moral luck in the light of the above considerations? I am convinced that it is rather clarified. Now, we are prepared for reading about intuitions concerning control and moral luck and we know that we have to deal with the ambiguous concept. Psychologists suggest that moral intuitions* have an affective nature that makes them sensitive to morally irrelevant features. On the other hand, philosophers claim that moral intuitions^ are something else, that they are reliable beliefs which can serve as a justification for moral assessments. However, there is not enough discussion between these two academic worlds. There is also no adequate methodological reflection on the relationship between description and normativity of moral intuitions. So far, these problems have not been the subject of deeper consideration within the discussion on moral luck. However, the Empirically Informed Ethics program makes us think about the meaning of the term “moral intuitions” and the appropriate methodological approach to research on them.⁶⁸ In the case of moral luck, psychologists try to explain why people differentiate moral assessments of similar cases and what they understand as the control condition. But they miss each other with philosophers who are more interested in the question of whether people should apply the control principle and whether it is morally correct (or even permissible) to make judgments based on the moral luck intuition. Maybe these two aims do not go in hand but surely philosophers and psychologists have to be aware of both domains. I hope this paper makes it easier. The other problem is, however, that there are good reasons to question the normativity of both moral intuitions* and moral intuitions^. The former are not good candidates for objective moral reasons for action and the latter cannot give us definite answers for moral dilemmas. That, of course, makes discussion on moral luck – which is, after all, full of arguments appealing to moral intuitions – much harder because we have to find reasons to justify our position somewhere else than the intuitive level of moral judgment. The question “Where?” is, however, out of the scope of this paper. It seems, though, that there is a promising position which gives hope for an adequate description and perhaps solves the problem of moral luck. It is the wide version of reflective equilibrium, the coherentist theory, which is not susceptible to the moral sceptic’s argument concerning the noninferential nature of moral intuitions and to Hare’s idea objection. Moreover, it allows us to take into account the process of moral reflection, both a priori moral principles such as the control principle and the considered judgments on particular cases, such as the practice of moral judgment based on moral luck intuition. However, much work remains to be done to apply this theory to the problem of moral luck and verify it.