Ethical Theory and Moral Practice

, Volume 15, Issue 4, pp 563–577

Applying Principles to Cases and the Problem of Judgment


    • Department of PhilosophyCalifornia State University, Fullerton

DOI: 10.1007/s10677-011-9311-x

Cite this article as:
Davis, J.K. Ethic Theory Moral Prac (2012) 15: 563. doi:10.1007/s10677-011-9311-x


We sometimes decide what to do by applying moral principles to cases, but this is harder than it looks. Principles are more general than cases, and sometimes it is hard to tell whether and how a principle applies to a given case. Sometimes two conflicting principles seem to apply to the same case. To handle these problems, we use a kind of judgment to ascertain whether and how a principle applies to a given case, or which principle to follow when two principles seem to conflict. But what do we discern when we make such judgments—that is, what makes such judgments correct? The obvious answer is that they are made correct by whatever makes other moral judgments correct. However, that cannot be right, for a principle can be inconsistent with morality yet still apply in a particular way to a given case. If the principle is inconsistent with morality, then morality cannot be what we discern when we judge whether and how that principle applies to a given case. I offer an alternative account of what makes such judgments correct.


Applied ethicsCasesCasuistryJudgementJudgmentMoral practicePractical judgmentPrinciplesSpecification

1 Introduction

Sometimes we decide how to act by applying moral principles to a case. However, applying moral principles is not mechanical or purely deductive. Principles are general and cases are particular, so principles do not always tell us precisely how they apply to a given case, or which of two principles to follow when they appear to conflict. To apply principles, we need a kind of moral judgment I call applicative judgment.1 Unfortunately, competent moral agents can reach very different judgments about where and how a principle applies, thereby reaching very different verdicts about a case. This is sometimes put by saying that principles are indeterminate. Moreover, each agent tends to feel that his or her judgment is correct and other judgments are incorrect. This raises two questions. First, how can we tell which applicative judgment is right? Second, what makes one judgment right and others wrong? This paper is about the second question (progress on the second question may position us for progress on the first). I want to define the truth-conditions for judgments about where and how principles apply to cases—or, in simpler terms, I want to know just what things or states of affairs we are supposed to be discerning when we judge whether and how a given principle applies to a given case.2 This is the problem of judgment.

We can think of this as the beginning of a metaethics for applied ethics; it will turn out that such a metaethics is very different from other metaethics. I will show that neither moral facts (if there are any) nor the descriptive components of principles completely determines the truth of applicative judgments. (Moral facts are particularly tempting as truth-conditions for judgments about how principles apply, for our principles are supposed to reflect moral truth.) I will discuss two theories about the relationship between principles and cases: Determinism, which says the principle and the case alone completely determine the relationship between them, and Indeterminism, which says that a third factor—perhaps moral facts—is needed to completely determine the relationship. I will reject Determinism and the moral facts version of Indeterminism, and argue for a third alternative.

2 Four Uses of Applicative Judgment

Most discussions of applicative judgment focus on conflicts of principles, but the problem is broader than that. In this section I will use a case to illustrate four uses of applicative judgment; later we will see that they each have different truth-conditions.

First, some concepts. Moral practice is deliberation about what is morally required in some particular real world case. More precisely, it is deliberation aimed at reaching a verdict about whether a given act-token is morally permissible, prohibited, or obligatory (Dworkin 2006, pp. 625–626). A case comprises the act-token and all features of the context that are relevant in such deliberation; correct case descriptions include that token and those features, and nothing else. There is more than one kind of moral practice; I am interested in the kind that involves applying principles to cases. A principle of the kind used in moral practice is a practical proposition that includes both a descriptive component—some reference to the kind of case it concerns—and a prescriptive component to tell us whether the act is permissible, prohibited, or obligatory. Such principles are used as heuristics in moral deliberation, and to be useful, they must be somewhat general and brief. This makes them abstract; their descriptive components do not contain enough information, explicitly or implicitly, to fully determine their own application. This is why there is an independent role for judgment.

So principles are indeterminate. So what? Perhaps the range of indeterminacy is usually small enough that we can largely agree about the case; perhaps our principles are imperfect but tight enough to get the job done, like a boat that is seaworthy despite some leaks. Unfortunately, the problem is worse than that, as this case shows:

An elderly patient is permanently comatose. He needs a respirator, feeding tube, and intensive care. His insurance does not cover all healthcare costs, and his treatment is slowly depleting his estate, leaving less for his wife and their self-supporting adult child. Despite never being a consistent churchgoer or student of theology, ten years ago he told his wife that refusing life support is suicide and against God’s will, and asked her to keep him alive as long as possible even if he became incompetent. She merely nodded, saying nothing. Now she is considering taking him off the respirator and feeding tube.

We will apply a principle of respect for autonomy, a principle of nonmaleficence, a principle of promise-keeping, and a principle obligating parents to protect their child’s welfare. Applying them requires using judgment in four situations: 1) when it is not clear whether and how a principle applies to some particular case, 2) when two principles seem to apply to a case but they conflict and we must judge which has more “weight,” 3) when we judge which features are relevant and hence part of a case description, and 4) when we judge whether two cases are relevantly similar.

First Use: Ambiguous Application

It is not always clear what a principle tells us to do in a given case, or even whether it applies to that case at all. Here it may be unclear whether the parental responsibility principle requires financial sacrifice after the child is a self-supporting adult. If she is an adult, perhaps she is no longer a child for purposes of this principle, but then again, even adult children inherit a parental estate. Moreover, it is not clear how this principle applies. We must judge whether the principle requires the wife to take her husband off life support (to conserve the estate for her daughter), or not (to save the daughter from grief at being the reason for taking him off life support). The daughter’s interests are both served and frustrated by either decision. To take another example, we must also judge whether the wife’s nod constituted a promise; if it did not, then the principle of promise-keeping does not apply here. For most of this paper I will focus on this use of judgment; in Section 5 I extend my account to all four.

Second Use: Conflicting Principles

The first three principles seem to require keeping the patient on life support, while the parental responsibility principle seems to require taking him off it. These principles appear to conflict, giving rise to an apparent (and perhaps actual) moral dilemma. When that happens and there is no lexical ranking for the principles, we typically balance them, and judge which principle has more “weight.” This is the only use most discussions of judgment address.3

Third Use: Which Features are Relevant, and Hence Part of the Case

It is not always clear how to describe a case—that is, what to put into the case description, and what to leave out. We must judge which features of a situation are relevant, and hence part of the case, for whether and how a principle applies turns on how the case is described. Here we must judge whether it is relevant that the patient rarely attended church and never studied his professed religion. In that case (!), the principle of respect for autonomy is arguably less applicable, for his decision seems less indicative of his actual wishes (or perhaps it applies but requires us to honor some earlier wishes he once had, rather than his most recent one). If, however, we leave that fact out, then there is reason to think that respect for autonomy requires honoring his most recent wishes.

Notice that the first and third uses of judgment are intimately connected: If we know which features are relevant, then whether and how the principle applies to that case will be more clear. Conversely, by stating the principle more completely, we may get a descriptive component in the principle that tells us more about which features of the case are relevant. In short, these two uses of judgment seem to work toward each other from opposite directions, and to the extent we have successfully used one of them, we have correspondingly less need for the other.

Fourth Use: Relevant Similarity Between Cases

Sometimes we deliberate about a case by comparing it with other cases where our moral judgments are more certain, in the modern, analogical sense of “casuistry.” Here we might ascertain whether the promissory principle applies to the wife’s nod by judging whether and how closely her nod is relevantly similar to a paradigm case of promise-making.

Now that we have taken a closer look at applicative judgment, let us take a closer look at the problem of judgment. Notice that, as with other kinds of moral judgment, we tend to see our applicative judgments as true or correct. In other words, our applicative judgments seem truth-apt. But if they are, what makes one of those judgments right and others wrong? As James D. Wallace asks, “What are the standards for performing correctly the sort of intuitive judgment or weighing up of reasons that is not simply a matter of mechanically subsuming particular cases under general precepts?” (Wallace 2008, p. 90.)4 The problem of judgment is to identify what can make applicative judgments correct.

This is a deflationary kind of truth; I do not mean that correct applicative judgments represent or correspond to some state of affairs that exists in a realist sense. I mean simply that some judgments are warrentedly assertible and others are not. The question of truth-conditions arises also for those who, like Onora O’Neill, say that principles indicate merely a “context for use” in which we exercise practical judgment in choosing among the many acts which “enact” the principle (2009, p. 225). On this view, judgment in following principles amounts to making good practical choices about how best to live (2007, p. 394). This conception of judgment does not sound truth-apt in the usual sense, but we can still ask what makes a judgment the best judgment. I choose to speak of truth-conditions, but “best” or “correct” work just as well.

3 Two Accounts of Truth for Applicative Judgment

There are two general accounts of what makes applicative judgments true: Either the relationship between principle and case is determined by the principle and case alone, or something extra is required to complete the determination. Both accounts face problems. Here is the first:

Determinism: The relation between principle P and case C is completely determined by P and C and nothing else.5

Determinism is not committed to the claim that all moral truths can be captured or represented in some principle or other, only that the application of whatever principles there are is completely determined by that principle, the case in question, and nothing else.

Determinism is easy to motivate. After all, applicative judgments are about how cases apply to principles; what else could determine that relation besides the two relata? And how else could a principle apply than by picking out the act-token to which it applies? According to Determinism, principles contain act descriptions which, if fully stated, could be matched, feature-for-feature, with case descriptions. Determinism works with the fact that a situation can be described at greater or lesser levels of generality. The idea here is that principles as we express them tend, for practical reasons, to be stated at a very general level of description, but when they are fully stated you get a completely detailed description of the case, plus a prescription. For example, “Keep your promises” could be unpacked to state a comprehensive definition of “promise,” what is involved in keeping one, and so on. Principles of this kind are rarely stated in their complete form, but if we unpack them, the act description will be the same as a complete description of the case, for both descriptions include all and only those features of the context of the act-token that are relevant to whether the act-token is morally permissible, prohibited, or obligatory. In all four uses of judgment, we use applicative judgment to discern the descriptive components of principles that are not fully stated.6 If Determinism is correct, the relationship between the first use of judgment (ambiguous application) and third (relevant features) is particularly close: what makes it true that P applies to C in some determinate way is the description they share, and a judgment is true when it discerns that common description. The first and third uses of judgment both discern that common description.

Unfortunately, Determinism fails. There are two problems. First, Determinism does not seem capable of handling conflicts between principles unless we imagine that such conflicts are always merely apparent—that if the principles are sufficiently unpacked, we will see that only one of them applies, or that there is some lexical ordering between them. This is a questionable assumption; it is possible that all conflicts of principles are like this, but it is also possible that some are not.

The second and deeper problem with Determinism lies in the fact that the descriptive component of a principle must fit more than one case. As noted, if the act description maps feature-for-feature onto the case description, then it implicitly contains a complete description of the case in every detail. However, this means the principle can apply to only one case, unless we suppose that our principles actually contain infinitely long disjunctions of descriptions of every possible case they could apply to. The only way a principle can apply to more than one case is for the act description to be general enough to take in more than one case. In that event, the case description must be equally general. But if the case description is general, how do we know the case has been described in the correct general terms? We must engage in the third use of judgment, and decide which features of the context are relevant, and under which general descriptive terms. What makes such judgments correct? It cannot be the fact that the act description and the case description match feature-for-feature, for they no longer can, unless we know what makes a general case description correct—and that amounts to knowing the truth-conditions for applicative judgments about how to describe a case. Thus, Determinism suffers from circularity; it defines the truth-conditions of applicative judgments in terms of other applicative judgments.

The second account of what makes judgments true is:

Indeterminism: The relation between principle P and case C is not completely determined by P and C, but is completely determined by P and C plus something else.

What is the “something else”? The obvious answer is moral facts, for our principles are supposed to reflect them; surely moral facts determine the truth of our judgments about their application. Thus, a judgment about how P applies to C is correct if the act it tells us to perform in C is the one we really should perform. Call this Moral Facts Indeterminism. On this account, principles are approximations or abbreviations of moral truth, and the part of morality that is left out is what bridges the gap between principle and case. According to Moral Facts Indeterminism, we use applicative judgment to discern that unstated portion of morality.

However, Moral Facts Indeterminism faces two problems. The first arises if claims about the validity of moral principles are neither true nor false; the second arises if such claims are true or false. First, if expressivism is correct, there are no moral facts. However, applicative judgments seem truth-apt even if other moral beliefs are not. For example, an expressivist would deny that a principle forbidding killing except in self-defense is either true or false, but should agree that this claim is true: “This principle prohibits snipers from killing unarmed people in the street.”7 Therefore, moral truth cannot be what makes applicative judgments true, not because expressivism is true, but because applicative judgments have truth-value even if expressivism is true.

Second, if expressivism is false, and claims about the validity of specific moral principles do have truth value, then surely some of our putative moral principles are false. Of course morality cannot guide the application of a false moral principle. In this respect principles are like laws: we hope that applying them yields justice, but even if they are bad or unjust laws, they can still be applied correctly or incorrectly. For example, a principle that says it is wrong for women to engage in politics gives false moral guidance, but it is true that this principle prohibits women from running for Congress. Whatever makes that true cannot be moral facts. (If we suggest that the descriptive term in the principle determines its application, then we are back to Determinism.) This is why the metaethics of applicative judgment is different from the metaethics of other kinds of moral judgment; whatever makes applicative judgments truth-apt cannot be whatever makes other kinds of moral judgment truth-apt.

We might dismiss this problem by saying that we are interested in true principles. If we mistakenly subscribe to a false principle, what difference does it make whether we apply it correctly? This response might work if we had reason to think that the truth-conditions for true principles are different from those for false principles. However, this is implausible. If something makes it true that a false principle applies to a case in a particular way, then it should work with true principles as well, for it works with principles simply by virtue of their being principles (otherwise it could not work with false ones). The truth-conditions for judgments will match moral truth if the principle is a true one, but this does not mean that moral truth supplies the truth-conditions.

I suspect that most writers on judgment implicitly hold a version of Moral Facts Indeterminism. O’Neill is an example. She says we almost never apply just one principle to a case; typically we must comply with many principles at once; like most other writers on judgment, she focuses on tension or conflicts between principles (the third use of judgment).8 A good judgment correctly resolves such tensions or conflicts, much the way we compromise among competing practical considerations in nonmoral contexts (like bringing a building project in on time, under budget, and to specifications). Such judgments involve practical reasoning (e.g., making good practical choices (2007, p. 394)) and making “justificatory arguments” in favor of one principle over another (2007, p. 403). Moreover, she does not discuss false principles; her examples are all plausible moral principles, such as respect for marriage or respect for freedom of expression (2009, pp. 226, 228). On this account, practical judgment and practical reasoning sound like reasoning about what is morally correct. (Note that this account does not obviously fit the other three uses of judgment; it does not tell us what makes a judgment correct or good when we determine how a seemingly ambiguous principle applies, or which features are relevant, or whether two cases are relevantly similar. None of these involve choosing between competing practical considerations.)

O’Neill does not try to say what makes such judgments or arguments correct: “I have said nothing about the justificatory arguments that might be given in favour of some norms or against others. I have made no assumptions about the quality of arguments that we may be able to offer in favour of specific norms that are important in specific aspects of life…”9 (Perhaps she says no more because the answer is obvious: moral facts.) She does discuss reforming our institutions and practices to minimize tension between principles, and prescribes regret and compensation when principles unavoidably conflict (2009, pp. 228–230). However, conflicts cannot always be avoided, and we want to regret making the right judgment, not regret making the wrong one. In short, O’Neill’s account of judgment leaves off where my project begins. (They may, however, be largely compatible.)

4 A Solution to the Problem of Judgment

In this section I offer a version of Indeterminism that avoids these problems. Here the “something else” which completes the determination of a principle is not moral facts but a hypothetical piece of moral theory construction. The basic idea is that judgments are true by reference to a counterfactual situation: A principle applies to a case if we would have derived that principle from that case had we engaged in moral theory construction. To distinguish this version of Indeterminism from Moral Facts Indeterminism we must assume that moral theory construction sometimes produces false principles which have a determinate application; more on this later.

I have a particular method of moral theory construction in mind; I call it the Derivation Account. My account of truth for applicative judgments works for principles constructed or discerned in this particular way, but not necessarily for other kinds of principles, if any exist. I will defend this choice later, but first I want to show how it sets the truth-conditions for applicative judgments. The method draws heavily on Rawls’s account of principles as heuristic generalizations (Rawls 1999), though this idea is not unique to Rawls.10 On this account, the usual way to discern or construct principles involves deriving them (my term for it) from what Rawls calls “considered judgments” about cases. The phrase “considered judgments” is broad enough to confuse; note that I use it in a very restricted sense. As I use the term, considered judgments are not just any moral judgments we have carefully considered; they are pretheoretical moral judgments about cases alone. They are not judgments about how a principle applies to a given case. (Even when they concern applying a principle to a case, considered judgments focus on whether the resulting verdict is intuitively acceptable, not on whether and how the principle is applicable to the case in the first place.) Considered judgments might concern easy cases, where the right thing to do is immediately obvious, or hard cases, where the answer is clear only after we have considered and discussed the case at length.

Either way, we start with a case where we feel sure of what should be done—what I call a confident considered judgment—and then derive a principle from that considered judgment by describing the case and adding a prescriptive component. Thus derivation starts by describing the case with all and only its relevant features (leaving out proper names) at a level sufficiently abstract to be heuristically useful, and adding a prescriptive component to tell us whether the act-token is permitted, prohibited, or obligatory. The next step is to make the description more abstract so that it can be part of a principle that is general enough to be heuristically useful. To do this, the descriptive component will not be merely an abbreviated version of the complete description; it will not typically include a complete list of the relevant features of the cases to which it applies (if it does, the names of the features may have indeterminate application). The descriptive component in a principle will not be a complete description, but it must be abstracted from a complete description. For example, the descriptive component might not look like a description; it might be a single term that refers to but does not restate a set of relevant features that tend to occur together—like “promise.” The descriptive component in a principle applies to a set of cases if the principle could be derived from all of them. Suppose, for example, that we contemplate a case where A did X after assuring B that A would do Y, and B foreseeably relied to his detriment on A’s assurance. If we feel sure that A’s conduct was wrong, we might restate that considered judgment by describing the case in terms of all and only its relevant features, then restate that description more abstractly and incorporate it into a principle of detrimental reliance.11

Hare discusses something like this when he talks about universalizing the descriptive component of a moral judgment (Hare 1977, pp. 10–14). Derivation might also involve looking for patterns in our considered judgments, then developing a principle that describes all the cases in that pattern; here we start with several cases instead of just one (induction in the sense of inferring the principle from particular judgments). We might refine the principle by testing it against considered judgments about still other cases, or seeking reflective equilibrium between that principle and considered judgments in other cases.12 Once the description is general enough to be heuristically useful, the principle will not pick out a case by listing its relevant features; there will still be relevant features, but the principle will not contain them.

Because principles are derived from confident considered judgments, we can use a principle heuristically to decide a case that is relevantly similar to the case the principle was derived from, reaching the same verdict we would reach if the case were easy for us, and we had a confident considered judgment about it.13 (As I discuss later, what makes cases relevantly similar is that we would derive the same principle from both of them.) Principles can do this by highlighting morally relevant features of a case, or arranging issues so that those with implications for other issues get decided first (as when we avoid doing harm before we benefit others).

Now back to applicative judgment. Where and how principles apply turns on their heuristic function. Applying principle P to case C is meant to guide us to the same verdict about C as the verdict implied by the confident considered judgment we would have about C—if we had a confident considered judgment about C. P can do this if P is derivable from C, for if P derives from our (hypothetical) confident considered judgment about C, then it will lead us to the verdict implied in that very considered judgment. After all, it came from that case, or could have. Thus, P applies to C if P is derivable from case C.14

Moreover, P applies to C only if P is derivable from C. This stems from the fact that we do not want our principles to guide us to the correct decisions only by accident. Suppose A sincerely assured B that she would do X, but she had no reason to believe that B might rely on this assurance to his detriment. Is A obligated to compensate B if she does not do X? If we apply the principle of detrimental reliance, the answer is “no,” for that principle requires that A had reason to anticipate B’s reliance. If we apply a principle of veracity, again the answer is “no,” for A was sincere: she really did intend to do X. Two principles lead to the same verdict, but the second leads there only by accident. Thus the first principle applies but the second does not, for only the first is derivable from the case.

I call this account Derivation Indeterminism. According to Derivation Indeterminism, we use applicative judgment to discern whether P is derivable from C; in effect, we are trying to discern a hypothetical piece of moral theory construction.15 Such judgments are correct when they discern such facts; they are false when they fail to do so.16

This is not a version of Determinism. According to Determinism, the relationship between a principle and a case is determined by that principle and that case alone—and nothing else. However, according to Derivation Indeterminism, that relationship is determined by whether the principle is derivable from a case, and that information is not contained in the principle, the case, or the two of them put together. When I say that a principle is derivable from a given case, I do not mean that the principle somehow contains a complete description of the relevant features of the case the way Determinism claims. We do not make applicative judgments by unpacking a description that is latent in a principle to see whether it matches the case feature by feature, as Determinism claims. Moreover, Derivation Indeterminism avoids circularity. Determinism is circular because it defines the truth-conditions of applicative judgments in terms of other applicative judgments. Derivation Indeterminsim, by contrast, avoids this by using two kinds of judgment: applicative judgments and considered moral judgments. Thus, Derivation Indeterminism does not define applicative judgments in terms of other applicative judgments.

Nor is this Moral Facts Indeterminism. We do not make applicative judgments by discerning moral facts. Although Derivation Indeterminism refers to constructing moral theory by deriving putative principles from considered judgments about cases, considered judgments can be false, and the process can result in false moral principles. Thus, constructing moral theory through principle-derivation does not guarantee true moral principles, and the Derivation Account does not require them. When our moral principles are false, they result from a good-faith but misdirected attempt at moral theorizing; either we began with the wrong verdict about that case, or we expressed that verdict badly in a poorly-conceived principle. Applicative judgments about false principles are true or false by reference to how a piece of mistaken moral theorizing would be extended to the case in question. (Where O’Neill would say that “enactments” of a principle resulting from good practical judgments are “best,” I say that the best enactment or judgment is the one that is appropriately related to the relevant hypothetical moral reasoning.)

Now let me introduce two complications. First, I have been speaking as if one applies a principle that one personally derived from the case in question. However, we derive very few moral principles for ourselves; we acquire most of them from other people. Thus, I might apply a principle that I would not derive from that case if I had a confident considered judgment about that case, and which I would reject if I were better informed. How can I correctly apply what I would, if better informed, regard as a false principle? The answer is that the principle applies to a case if those who would endorse the relevant considered judgment (the one the principle derives from) would derive that principle from that case. For example, the principle forbidding women to engage in politics clearly forbids them from running for Congress. If that result troubles me, then I should question whether I selected a valid principle, but not whether that principle applies to that a case, for those who disapprove of such activity would derive it from that case. Whether a principle is morally correct and whether it has been correctly applied are different questions.

Second, it is possible that I would derive the principle from my confident considered judgments about some cases, but not from my (hypothetical) confident considered judgment about this case. But what about those who would derive it from this case? Does the principle apply here because some people would derive it from here, or not apply here because I would not derive it from here? We cannot simply say that it applies because those who have the relevant considered judgment about the case would derive it, for I would derive it too—but not from here.

The answer turns on the heuristic purpose of principles. I want my principles to guide me to the correct verdicts, and if I have reason to disagree with the considered judgments others reach, then I should also disagree with their derivation and application of the principle. If a principle guides some deliberators to their confident considered judgment, but does not do so for other deliberators, then that principle is a good heuristic for the first group of deliberators, but not for the second. This implies a kind of relativism by way of pragmatism; a principle applies if it guides you to the target verdict, and that target will vary with the deliberator. This is not metaethical relativism about considered judgments in question, or about the validity of principles. This is relativism to deliberators merely when it comes to the heuristic purpose of principles. However, provided that either I would derive the principle from the case, or I would not derive it from any other cases, this complication will not arise.

Now back to my claim that applicative judgments have truth-value. I argue for this by reference to the derivation model of moral theory construction. But is this model is the right one? Those who reject it seem to have reason to reject my entire account. The answer is that I do not assume that this method is the only right way to construct or discern the kind of moral principles we use in moral practice. I assume that it is one of the plausible methods of moral theory construction, and that many of us use it (often without realizing it). It generates one kind of principle, for heuristic purposes in moral practice. When a principle is constructed in this way, then applicative judgments about it will have truth-conditions, for such principles are heuristic devices to help us reach the same case verdict we would reach if we had a confident considered judgment about that case, and it will be either true or false that a particular application of a principle has led us to that verdict. This is not to say that there are not other kinds of moral principles, perhaps generated in a different way, with other purposes. Kantians, for example, have a very different account of principles; they are not generalizations derived from considered judgments about cases. However, the kind of principle I discuss is common, and principles generated this way have the truth-conditions I discussed.

5 Four Uses of Judgment

In this section I will use the Derivation Account to explain the truth-conditions for all four uses of judgment.

5.1 First Use of Judgment: Ambiguous Application

A principle applies ambiguously when it is not clear what action is required, prohibited, or permitted by that principle in a given case, or even whether the principle applies to that case at all. For example, in the life support case, it was unclear whether the parental responsibility principle requires financial sacrifice after the child is a self-supporting adult. Here are the truth-conditions for judgments about ambiguous application:
A principle applies to a given case if and only if it has clear implications for a proposed act-token by either requiring or forbidding that action, and

a. that principle was derived from a considered judgment about that case, or it is the principle that those who endorse that considered judgment would have derived from that case had they engaged in moral theory construction from a confident considered judgment about that case, and

b. precisely how the applicative judgment says the principle applies yields the same verdict about the case as the considered judgment that the principle was or would have been derived from.

(If the action is neither required nor forbidden by that principle, but merely permissible, then the principle is neutral with regard to that action, and therefore inapplicable.)

5.2 Second Use of Judgment: Conflicting Principles

The truth-conditions for judgments about principles that seem to conflict are more complex. In the case above, the first three principles seem to require keeping the patient on life support, while the parental responsibility principle seems to require taking him off life support. When principles appear to conflict, we face a moral dilemma, where the agent seems morally obligated to do two different things but cannot do both. (I take no position on whether all moral dilemmas are merely apparent, so that principles never really conflict.) When we apply multiple principles that are not lexically ordered, we use judgment to weigh the conflicting principles to see which one to follow. Many writers complain that the weight metaphor explains very little, and that such judgments are notoriously difficult to explain and argue for.

My account of moral dilemmas is essentially Rossian (or at least on an interpretation of Ross according to which prima facie duties remain valid principles despite being outweighed by other principles). However, Ross’ account is hard to understand, and my account helps to explicate his. Ross tells us that when prima facie duties conflict, one’s actual duty is the prima facie duty with “the greatest balance of prima facie rightness”—another way of saying it has more weight (Ross 1930, p. 41). Our actual duty is our duty proper, while a merely prima facie duty has merely a “tendency” to be our duty:

Tendency to be one’s duty [e.g., a prima facie duty] may be called a parti-resultant attribute, i.e. one which belongs to an act in virtue of some one component in its nature. Being one’s duty [e.g., the duty that weighs the most] is a toti-resultant attribute, one which belongs to an act in virtue of its whole nature and of nothing less than this (Ross 1930, p. 28).

The first step toward understanding this passage is to bear in mind that principles are constructed or discerned through the process of derivation, and this makes it puzzling that they could ever conflict in the first place. I have been speaking as if each considered judgment yields only one principle. However, if each considered judgment generates only one principle, and a principle applies to a case if and only if it is derivable from that case, then how does it happen that there are cases where more than one principle seems applicable, and how can they conflict?

I do not know whether cases where two principles both apply with equal weight are possible, such that we cannot correctly say that one principle outweighs the other. For two principles to apply with equal weight, they would both have to be derivable, yet if our considered judgment leaves us feeling torn about how to respond to the case, then we are not likely to derive either principle from it. Perhaps the actions consistent with both principles are permissible, and our considered judgment would be that there is a range of actions we can take. Anyway, such cases are likely to be rare, so I will not discuss them further.

Given that we often do think one principle outweighs another, most cases where principles conflict should have the following structure. Imagine a case with two relevant features: F1 and F2. Imagine further that when that case has both F1 and F2 and we find it easy, we would have the considered judgment that A1 is the proper action in that case. If we derived a principle from that considered judgment, we would derive principle P1. However, if that case were easy but contained only feature only F2, we would have the considered judgment that A2 is the right action, and we would derive principle P2. Thus, the case having both F1 and F2 may suggest both principles, for it contains a “smaller” case containing only F2. To say that P1 outweighs P2 in this case is simply to say that we would have derived P2 from the case if F2 were present but F1 were absent, but we would derive P1 when both F1 and F2 are present (and thus P2 is not derivable). Principle P1’s status as our actual duty (the one having more weight) is “toti-resultant” because it belongs to the act (or case) “in virtue of its whole nature”—that is, in virtue of both features F1 and F2. Principle P2 is merely a prima facie duty, for it belongs to the act only “in virtue of some one component of its nature”—in virtue of feature F2 alone. To the extent this picture is Rossian, I am reading Ross to say that, even though P2 is a prima facie duty, it’s still a valid principle. It is a principle because it would be derived from an easy case containing F2 but not F1, and it is outweighed by P1 in a case containing both F1 and F2. It is a valid principle because it would be derived from a case with F2 alone, even though this is not one of those cases, and it is outweighed because this case contains F1 as well. (If this is not what Ross meant by “prima facie duty,” then my account is quasi-Rossian.)

Other moral dilemmas may have a more complex structure, but they will be elaborations of this structure. We are not likely to derive principles from cases that present moral dilemmas, for we tend to derive principles from cases where we feel pretty sure about what should be done, not cases where we feel conflicted. However, the truth-conditions for judgments about conflicting principles will be a function of which principles we would derive from the case presenting a moral dilemma, were we to engage in derivation from that case. (To the extent that we argue for a judgment by engaging in practical reasoning about which principle should give way to another, as O’Neill suggests, this is what such arguments must establish.) Thus,
A judgment that principle P1 has more “weight” than principle P2 when they conflict in a given case is correct when

a. those who would endorse the relevant considered judgment did or would have derived P1 from that considered judgment about that case, and

b. the case contains a subset of features from which, if those features were considered as a complete case on their own, they would have derived P2.

This model of how conflicting principles can be derived from a case also suggests an explanation of degrees of weigh: why a principle seems to weight a lot more than its rival, or only slightly more. Suppose removing F1 produces a case that seems very similar to the case that contains F1 and F2. Then P1 will seem to have a little more weight than P2. If, however, removing F1 produces a case that is strikingly different from the case that contains both features, then P1 will seem much more weighty than P2. Or so I surmise—this is only a hypothesis for now.

5.3 Third Use of Judgment: Which Features are Relevant, and Hence Part of the Case

Whether and how a principle applies can vary with how the case is described. When we are deciding how to act, we do so in a context with many features. Some of them are relevant to the act-token that is contemplated and therefore part of the case, and some are not. A complete case description includes all and only features of the context that are relevant to the moral status of the act-token in question. As Lawrence A. Blum says, the “idea of moral judgment as bridging general rule and particular situation depends on a prior individuating of ‘the situation.’ It is moral perception which does that…” (Blum 1994, p. 42). We judge whether some feature is relevant and hence part of the case description, and our judgment is true under these conditions:

A feature of the context of an act-token is relevant and hence part of a case if and only if its presence or absence in the context affects our moral evaluation of the act-token.

If the feature is present, we are more likely to judge the act wrong; if it is absent, we are more likely to judge the act right (or vice versa). This is why we speak of “making a case” when we marshal facts to show why some act should or should not be done; we consider things which make no difference to be irrelevant and not part of the case.17 In the case earlier, the fact that the wife nodded may be relevant; that she was sitting down was not, for our verdict will not vary with her posture. On this account, the first and third uses of judgment are related through derivation; a feature is relevant if it affects our judgment of the act-token, and what principle we would derive from that judgment depends on the features that affect that judgment.

5.4 Fourth Use of Judgment: Relevant Similarity Between Cases

Finally, we might deliberate about the life support case using the method of analogical casuistry, where we ask whether our case is relevantly similar to another, paradigm case where we feel more certain about what should be done. Suppose we judge whether the promissory principle applies by asking whether and how closely the wife’s nod is relevantly similar to a paradigm case of promise-making. What would make such a judgment true?

A case comprises all relevant features of the context of the act-token under consideration. If we would derive the same principle from both cases, then both cases will have the same relevant features, for we can derive different principles from them only if one of them contains a feature which affects our judgment about what to do, and which the other case lacks. (I am assuming that the relevant features are stated without using proper names and the like.) Therefore,

A case is relevantly similar to another case if and only if those who would derive a given principle from the first case would also derive it from the second case, and similar to the degree that the principle is specific rather than general.

If two cases are very dissimilar, a principle will have to be more abstract in order to be applicable to and derivable from both of them, but they can still be relevantly similar. The more specific the principle shared by two cases, the more similar those cases will be. For example, we should follow the paradigm case of promising and classify the wife’s nod as a promise when we would have derived a principle of promise-keeping from both cases. However, because the wife’s nod lacks many features found in the paradigm case, the principle they share might be something more general—not a principle of promise-keeping as such, but perhaps something like Ross’s principle of fidelity, which includes express promises, implied promises, detrimental reliance, and the like.

6 Conclusion

This paper concerns the metaethics of applied ethics. I have argued that judgments about the application of principles to cases are true if and only if we would have derived those principles from those cases had we engaged in moral theory construction. I then used that account to define the truth-conditions for judgments about principles that apply ambiguously, judgments about which of two conflicting principles has more “weight,” judgments about which features of a situation are relevant and hence part of a proper case descriptions, and judgments in casuistry about whether two cases are relevantly similar, and to what extent.


Some writers speak of intuition, practical wisdom, or perception here. Those terms are not interchangeable, but in this context they all refer to cognizing the relationship between principles and cases in a noninferential and nonsensory way.


My discussion is not affected by the debate between particularism and generalism. If generalists are right, then principles work the way we always thought they did. If particularists are right, then principles are rough generalizations which may fit a great many cases, but not all, for the moral valence of a feature mentioned in the principle may switch in certain cases, depending on the presence of other features. However, we can still ask whether and how principles apply to cases where the moral valence of features has not switched.


See Richardson (1990) and O’Neill (2007, 2009).


See also Richardson (1990, pp. 287–288) and O’Neill (2007, p. 399).


My distinction between Indeterminism and Determinism is a modification of Wallace’s distinction between “rule formalism” and “intuitionism,” though his distinction concerns what guides us in applying principles, while mine concerns what determines the relation between principles and cases (Wallace 2008, pp. 84–90).


Wallace can be read as something like a Determinist, for he argues that we are guided not only by the explicit, “manifest” content of a principle, but also by the implicit, “latent” content, which includes the point or purpose of the principle and “how this principle functions in concert with other principles.” (Wallace 2008, pp. 96, 100.)


Alan Gibbard’s “norm expressivism” exploits this contrast, earning cognitivism through the application of norms that we endorse expressively (Gibbard 1990).


O’Neill’s distinguishes between cases where there is “tension” between principles and one principle is defeated by another, and cases where it is “impossible” to satisfy both principles—conflict in a stronger sense, where neither principle is defeated. I take her to mean that, when a principle is defeated by another principle, the right act need not satisfy the defeated principle (2009, pp. 226–228). In my terminology these are both conflicts, but I acknowledge the distinction.


Nor is that her project: “My question has only been whether whether if we had such arguments we would still find that norms were impotent to guide action…” (2007, pp. 403–404). In other words, her project here is to ascertain whether practical reason not only validates principles, but can also validate particular actions. Her answer: Yes.


See Gert et al. (2006, pp. 3–49), and Mill’s discussion of common-sense moral rules as heuristics for utility-maximizing (Mill 1861, p. 24).


Some see moral principles as truth-makers for particular moral verdicts, the way the natural law that copper is electrically conductive makes it true that particular bits of copper conduct electricity. Others see moral principles as generalizations made true by particular verdicts, the way “all coins in my pocket are copper” is made true by particular coins. (McKeever and Ridge (2006, p. 12).) The heuristic principles I discuss are summaries of confident considered judgments; they are generalizations. This does not preclude the possibility of moral principles that are truth-makers, or the possibility that our generalizations might closely approximate whatever truth-making moral principles may exist.


Even philosophers who seem to suggest that we can intuit principles directly, such as Ross, hold that principles must be consistent with our intuitions about particular cases, and are discerned through making many such intuitions even if we are not aware of this. This is close to the Derivation Account (Ross 1930, pp. 32–33; see also Rawls 1971, pp. 46–53).


We do not have a confident considered judgment about every case. However, I assume that moral truth is always knowable, at least under suitably idealized conditions.


I am not suggesting that only one principle can be derived from each case. I do assume that, given the considered judgment you reach about a case, you will try to derive the most useful principle you can—the one most likely to get you to the verdict in relevantly similar cases that you would reach if you had confident considered judgments about them.


I have no idea how we do this; that is a question for empirical psychology.


There is room for the possibility that applicative judgments can alter our pretheoretical considered judgments about cases and vice versa, for we sometimes revise our considered judgment about a case when it conflicts with what a principle tells us about that case, or the principle when it conflicts with the considered judgment, depending on which we have more confidence in—the considered judgment or the principle. This is a familiar aspect of reflective equilibrium. To the extent that our considered judgments about cases and our principles are both revised in the course of seeking reflective equilibrium, it may seem difficult to separate the truth-conditions for applicative judgments from those for other moral judgments. However, the fact that we seek coherence between them does not mean they have the same truth-conditions. For example, if we revise a considered judgment in light of what our applicative judgments tell us about how a principle applies to a case, we might form a false considered judgment, if the principle we apply is false.


See definitions of relevance in Gert (1998, p. 227) and Hare (1989, pp. 193–4).



I thank Carl F. Cranor, Michael Cholbi, Mark LeBar, Graham McFee, Sean McKeever, Stephen R. Munzer, Michael Zimmerman, and two anonymous reviewers for comments.

Copyright information

© Springer Science+Business Media B.V. 2011