In a seminal essay anticipating the contemporary preoccupation with the empirical ground of our normative judgments, Daniel Callahan wrote, “What is, is all we have in the universe. We have to admit that if nature is not the source of morality, it does not have any source.”Footnote ¹ The attitude that Callahan has characterized in these stark terms has become a commonplace today, and efforts have shifted toward determining an adequate empirical basis for ethics, and what its acceptable uses are, rather than justifying in general terms the fact that “what is” is relevant for “what should be,” and refuting the so-called naturalistic fallacy.Footnote ² At a minimum, a “naturalized bioethics” has to “resist the pull to purity”: it must acknowledge the complexity of ethical judgments and actions, promote curiosity for real-life dilemmas, and affirm the necessity that our ethical proposals be accountable to facts.Footnote ³ Empirical ethics avails itself of a variety of methods. However, whereas qualitative and quantitative studies are common practice and their relevance is widely discussed,Footnote ⁴ the use of individual cases has not been scrutinized with a view to assessing the reasons why, and the extent to which cases can constitute a valid empirical basis for ethical judgments. This is all the more crucial because a case anchors moral reasoning to practice in an intuitive and immediate way and often suggests insights or solutions to particular and even general issues: we are all familiar with the impact that cases have on our understanding of situations of ethical unease, with the satisfaction that we get when we feel that the solution we are suggesting accounts for several prominent cases we have encountered or read about, and with the heuristic function that cases play with respect to our common beliefs and practices. Indeed, one might argue, as Jonathan Dancy does, that cases are the only suitable tool of empiricism in ethics, because only cases can serve as tests of a sort for moral principles.Footnote ⁵

Medical ethics, geared as it is toward the improvement of medical practice, is at the forefront of the naturalistic movement. It is a fact that cases do play an important role in medical ethics, and it is not surprising that they should: clinical medicine itself is rightly characterized as an art and to that extent deals primarily with particular cases rather than with general diseases. Cases are very often described in medical ethics publications: whether short or long, detailed or sketchy, current or exceptional, they capture our imagination and often provide a foundation for the authors’ conclusions. Moreover, cases, whether fictional or real, are central to narrative ethics, a prominent approach to bioethical reflection and education today: thanks to their temporal organization, the selective nature of the events recounted, and the thickness of their texture, narratives can better account for the lived experience of those involved in a medical decision,Footnote ⁶ and to that extent they constitute a necessary tool for a context-sensitive bioethics.Footnote ⁷ Finally, cases are the nuts and bolts of clinical ethics consultations that can be considered as the laboratories of medical ethics: here our ethical theories are daily put to the test of experience, and new concepts and approaches are constantly elaborated and fine-tuned to adjust to ever-changing practices. Indeed, the term “case” not only indicates the individual situation and moral decision with which we are involved but also refers to the way in which such individual experiences are recounted and used to nourish our ethical approaches and our moral experience.

However, although cases appear to be an irreplaceable tool in medical ethics, they are perceived as a disturbing element in our ethical reflection, for several reasons. First, by definition a case is unique, and to that extent it may appear of no use for addressing a general issue, especially if it is recounted in detail and in a way that emphasizes the idiosyncratic elements of the situation. Intractable cases of conflict between advance directives and proxies, for example, do not by themselves challenge the wisdom of the principle of advance directives. Even in clinical ethics consultations, certain details of the case at hand have to be disregarded because they prevent consultants from getting the situation into focus. Cases thus seem to be in need of being stripped of their uniqueness, which is precisely what makes them cases. Second, cases often have a strong emotional impact on us and therefore are liable to have a distracting effect on our ethical reflection. A doctor, for example, maintains that she prefers not to meet a 70-year-old man who would like to have access to assisted reproductive technologies: his presence and his reasons rooted in his particular story might upset entrenched ethical convictions and guidelines that set the limits at 65. Third, cases may even appear to be utterly disruptive of sound ethical reflection rather than conducive to ethically well-informed decisions, especially when they represent exceptional circumstances. Cases that receive wide publicity in the media might challenge a working consensus and suggest maxims of action that might be unacceptable if generalized. This is true, for example, for individual requests for active euthanasia in countries where the practice is not authorized; it is clear that a law should not be changed only to solve particular cases but should be changed to address what is perceived as a common problem. As such, cases do not seem sufficient for providing ethical arguments against the status quo; in some instances they can even have an adverse effect on serious ethical debate and can be used for political and ideological manipulation.

Thus, despite their conspicuous presence in medical ethics, cases seem to involve a paradox that can undermine their use as a valid basis for an empirically sensitive bioethics: they are interesting for precisely those reasons that make them problematic, namely their uniqueness. The purpose of this article is to address this basic difficulty. I shall analyze the nature and use of cases in the bioethical literature and practice and put it in a more coherent and systematic perspective. By drawing on the rhetorical tradition of example, which is the topic of more systematic analyses, I try to better understand and justify the relevance of cases for an empirically informed bioethics. I define in broad terms two kinds of cases—”tamed” and “wild” cases—and describe their respective characteristics, uses, and modus operandi. Tamed cases look promising for solving the paradox: if judiciously used, they can fulfill several important functions, such as illustrating moral standpoints, guaranteeing the applicability of ethical principles, and refuting or supporting general claims. However, they raise some major difficulties of their own and therefore appear as an insufficient empirical basis for medical ethics. I finally sketch the heuristic and subversive function of what I shall call wild cases and argue that, if they are carefully used, they can be of the highest value for empirical ethics, despite the fact that, unlike tamed cases, they retain their paradoxical nature.

Rhetorical Examples and Ethics: Tamed Cases and Argumentation

The usefulness of cases for reaching valid conclusions has been theorized in the field of rhetoric and argumentation theory under the label of arguments from example. Rhetorical examples can be defined in the following binary way: “The evoking of history which has or has not actually occurred, which is similar or related to the matter under discussion, which is implicitly or explicitly brought into connection with this matter as argument (evidence or model) OR as ornament, and which takes the form of a narration, a name-mentioning or an allusion.”Footnote ⁸ Indeed, the word “example” can be understood in two ways: as either a particular kind of inductive argument or as a positive model for action. When examples are placed before the conclusion, Aristotle writes in Rhetoric (II.20),Footnote ⁹ they create conviction and constitute a kind of proof. But, in a less technical sense, examples can also follow the conclusion of an argument and make it more forceful by evoking a striking instance. In this second sense, examples can also function as exemplars, personal stories that have the power of inspiring behavior and motivating a particular course of action, rather than forcing an audience to accept a hypothesis. An exemplar serves as either a positive or a negative model for action and consists of an authoritative figure whose traits are idealized and whose explicit message remains somewhat ambiguous.Footnote ¹⁰ Like examples, cases can also be used as both argumentative tools and striking and inspiring personal histories. In what follows I shall describe the first function of examples in more detail.

In the first sense, namely as a kind of argument, examples belong to the category of technical proofs, as opposed to more direct proofs based on witnesses, laws, and contracts. Whereas enthymemes, that is, rhetorical deductive arguments with a missing premise, aim at deriving particular conclusions from general premises, arguments from examples use particular instances as evidence for establishing general conclusions. Examples, therefore, consist of a truncated form of induction in which a well-chosen single case, or a short series of similar cases, constitutes evidence for a general conclusion. In a normative rather than descriptive context, as in practical reasoning, an argument from example has the following form: A did x, which is commonly judged as a good action; therefore a general maxim prescribing x is a good maxim. Arguments from example, as opposed to full-blown inductive generalizations, have two main features: they appeal to our emotions in addition to our rational faculties, and they are elliptical. Chaim Perelman stresses the strong confirmation effect that illustrations can have on hypotheses, insofar as they are based on a vivid example.Footnote ¹¹ To that extent, arguments from example have a higher rhetorical value than plain inductive generalizations and serve as tools of rational persuasion concerning general statements. Also, because their conclusions may be left unstated, arguments from example are more powerful and flexible than ordinary inductive arguments and therefore more difficult to refute. In France, the example of a well-publicized case of withdrawal of nutrition and hydration that led to a patient’s painful death could serve as evidence for general conclusions as diverse as “nutrition and hydration are not medical treatments and therefore should not be withdrawn” and “active euthanasia should be legalized in order to prevent such events from happening.”Footnote ¹²

But what exactly allows examples to play such a crucial argumentative role? In order to answer this question, we need to look more closely at what cases are. Here too we can draw on the rhetorical tradition. The etymology of the term “example” points to two basic characteristics. The Greek term paradeigma is related to the verb “to show”: examples have an ostensive function insofar as they indicate and exhibit something more than they describe it. The rhetorical force of examples resides precisely in their ability to point to an objective entity open to public and common scrutiny: “Examples are usually labeled neatly. . . . The example seems to proclaim the speaker’s common cause with the audience in looking at the world the way it is.”Footnote ¹³ Cases, like examples, point to a strong form of reality, that of an individual personal story rather than an abstract entity or concept. Indeed, even a fictional case appears to be real, especially when it is described with a wealth of details and respects literary verisimilitude. However, the Latin translation of the Greek term exemplum introduces another dimension in the analysis of cases. Exemplum is related to the verb “to cut” (eximere): insofar as they are recounted, examples are isolated from their original setting and placed in a new argumentative context. This is the reason why the same event or personal story can give rise to several cases according to context and to the function it is designed to play in argumentation; different cases will highlight different features of the event and will emphasize selected aspects of the situation.

The fact that cases are “cut out” of their original context has two important consequences. First, to the extent that they are always reconstructed and considered from a particular point of view, cases, like examples, are never unique but potentially refer to a series of similar cases, even though these are not evoked explicitly. Aristotle himself writes that, insofar as they have to serve for reaching inductive generalizations, several similar examples should be preferred to a single one. Even when cases are described in a detailed, as opposed to a sketchy, way, they strive toward generality: “When used in an example, an entity loses its autonomy and unicity. . . . Occurrences and events not only can but also must be both themselves and representative of something else—of the subordinating concept—and must be replaceable in that function with other events or entities.”Footnote ¹⁴ As we shall see, this is the basis of casuistic reasoning. The second consequence concerns the fictional—as opposed to the real—nature of cases. Following Aristotle, who distinguished historical from fictional examples,Footnote ¹⁵ we can classify cases as representing either real or fictional events; the latter can be either literary cases, imaginary cases, or counterfactual cases. The former are often used for pedagogical purposes:Footnote ¹⁶ it is assumed today that fictional cases represented in the literature are a tool of moral education. As for imaginary cases, everybody is familiar with “scare cases” used in controversial arguments about cloning and euthanasia: a dictator’s cloned children or an elderly patient killed for lack of money can be used as arguments of sorts against the legalization of cloning and euthanasia. Counterfactual cases are cases in which, by thought experiment, so to speak, one important variable of the real event is changed: what would have happened if Mr. X had survived his accident as a quadriplegic, after being resuscitated? Such hybrid cases can be an important aid in moral reflection. They can confirm the well-foundedness of a decision by verifying that it can be considered as justified whatever the upshot: if, even in the hypothetical case of severe handicap, the resuscitation still appears as a good decision, that might confirm, a posteriori, the moral soundness of the decision. On the contrary, if the decision appears as unjustified given a different outcome, a counterfactual case can highlight an important morally relevant property that has been crucial in the decision—in our case, the role played by the chances of recovery after resuscitation.

However, despite the intuitive distinction between real and fictional cases, we can argue that cases, like examples, are always fictional to a certain extent, whether or not they are meant to represent a real underlying event. Insofar as they are always recounted, they are not entirely shown, and they always involve an element of reconstruction. As a recent scholar of the history of rhetorical examples writes: “An act, entity or event is transformed from itself into ‘an example of.’ . . . Examples in short do not happen, they are made.”Footnote ¹⁷ Psychoanalysts have long reflected on the nature and uses of cases and have distinguished them from histories. Whereas histories remain secret and self-contained, cases are public and may serve to corroborate statements: “Inasmuch as it is the illustration of a reconstructed series of events, that get their sense from a theoretical perspective, the case is abstracted from individual history. Therefore, the question of whether it is true or false with respect the latter does not arise.”Footnote ¹⁸

We may summarize the lessons that we can draw from rhetorical examples in thinking about cases by saying that the fictional and iterative natures of cases are tightly linked to each other. Insofar as a case is always fictional to a certain extent, its nature is iterative rather than unique. The reverse is also true; the more a case is made to represent several other similar cases, the more it is a fictional representation of real events. As John Lantos writes in the introduction to a landmark neonatology book devoted to the analysis of a single case, The Lazarus Case,

These characteristics of cases—their fictional nature and their iterativeness—suggest a solution to the paradox involved in their use as an empirical basis for ethical reflection. Cases are useful insofar as they are tamed, namely, when they are recounted in a way that allows them to relate to a more general perspective and to represent other similar cases. Tamed cases are simplified for the purpose at hand; they highlight the common rather than the unique features of events, and they are depleted of their immediate emotional impact. As such, they can be used for a variety of purposes that go beyond inductive generalizations. They can serve as illustrations, as tests and potentially refuting instances of a given principle of conduct, and as arguments in favor either of a particular decision or of a more general outlook. It is worth describing these functions briefly, because they play such an important role in medical ethics.

First, as illustrations, cases have a strong rhetorical and pedagogical import. If general principles are anchored in the presentation of cases, they may appear more plausible and can be more easily considered as relevant for practice. Second, cases also serve to guarantee the applicability of principles. This more substantial use of cases is well represented by the examples cited in T. Beauchamp and J. Childress’s manual of medical ethics. In this sense, cases help fine-tune principles and thus ensure that they are more adjusted to the practice of medical decisions. In order to give some indication about the application of the principle of respect of autonomy, the authors cite the imaginary, but plausible, case of a woman who suddenly decides to discontinue dialysis after years of accepting to fight her disability. In a case like this, we may have doubts about the autonomous character of the decision. The authors conclude, “Actions are more likely to be substantially autonomous if they are in character . . . but acting in character is not a necessary condition for autonomy.”Footnote ²⁰ Charles Larmore has explored another aspect of the use of cases for the application of principles. He has argued that cases can serve as an exercise for moral judgment, which in turn is essential for applying principles to practice: “I would suggest that the important role of moral examples lies in their suitability as just such reasons [to justify specific actions], and that they are useful precisely to the extent to which they are examples of the exercise of moral judgment.”Footnote ²¹ Third, cases can also contribute to test theories, principles, and maxims of actions and thus constitute potentially refuting instances for them. Cases, especially when they are plausible and common, can thus play a crucial function in reductio ad absurdum arguments and are used to show that certain principles are unacceptable. Medical paternalism can be refuted precisely in this way, among others; in instance after instance, one can show that it may lead to abuse of power.

Lastly, cases can also be used as positive indications in favor of a proposed approach and thus can play a more constructive role; they can serve as arguments that justify either a particular decision or a general ethical approach. In this context, the selected cases are usually commonplace and uncontroversial. A casuistic methodology uses paradigmatic cases as a help for justifying a decision in more controversial cases. Paradigmatic cases are simplified cases about which a certain consensus exists; they can thus suggest a clear and standard solution about related cases that are sufficiently similar.Footnote ²² Intuitions about central cases can equally be used in a reflective equilibrium methodology as arguments in favor of a given theoretical approach. Cases here are general cases and concern types of ethical decisions that correspond to robust and uncontroversial opinions, what Rawls calls our “considered judgments,”Footnote ²³ rather than particular idiosyncratic situations. Intuitions about key moral decisions on crucial cases serve as a double-checking strategy for our ethical principles and can help us achieve a broader coherence between ethical principles and accepted social norms.Footnote ²⁴ Cases can also serve as metatheoretical tool. The editors of a recent book used a case explicitly described as “constructed”Footnote ²⁵ in order to show how different methods (principlist, hermeneutical, etc.) could deal with it: do conclusions converge or diverge? “Does it actually matter which method one uses? Or are some methods better for some purposes?”Footnote ²⁶ They concluded that the exercise was methodologically useful: “While consensus was not reached on what to do in the case, something like agreement was reached on the idea that conversation between the various perspectives added something to each of them.”Footnote ²⁷

Wild Cases, Heuristics, and Moral Imagination

As we have seen, if properly used, tamed cases serve a crucial function with respect to moral deliberation and moral theory. Indeed, tamed cases offer a solution to the paradox implicit in their use; by being tamed, cases can be used despite their uniqueness, their emotional impact, and their disturbing nature. This is so because they are controlled and so carefully crafted as to highlight only certain morally relevant features of the situation; as such, they cannot be used as the basis of hasty generalizations, faulty slippery slope arguments, and justifications of shallow intuitive judgments. In other words, they do not lend themselves to what has been called the “fallacy of misleading vividness.” In this respect they can be likened to a chastened form of case reporting in medicine.Footnote ²⁸ Thus, to the extent that they establish a bridge between the principles that inform our decisions and practice, they are a viable instrument of empirical ethics.

However, as we have seen, they can only do so insofar as they are fashioned in such a way that they can easily respond to a certain theoretical perspective, and therefore their use involves a double risk. First, insofar as they are fictional and iterative, a question arises as to their nature: are they genuine cases or simply disguised arguments? At best, they can be seen as sketchy simplified situations that define a broader class of instances. But, one may argue, if tamed cases are deprived precisely of their defining features, namely their uniqueness and originality, what is so special about them, as opposed to other forms of empirical data? Second, and more importantly, tamed cases are so dependent on a given theoretical approach that, although they may provide empirical evidence for a claim, they may not offer us as direct an access to reality in all its dynamic nature and complexity as one might expect. To the extent that they highlight certain features of the situation at the expense of others, and those features are precisely the ones corresponding to our perspective and purpose, cases used in this way—tamed cases—risk begging the question: cases cannot serve as independent evidence for, or against, a given normative claim, and we might well discover what we had previously assumed during the process of taming the case at hand, that is, selecting its morally relevant features.

More importantly, tamed cases cannot not perform a heuristic function with respect to our concepts and theories: in order for a case to challenge our current views and offer us some new insights that will allow our concepts and theories to evolve, a case must be defined in terms that are as much as possible independent of a theoretical and argumentative context. Onora O’Neill, a contemporary Kantian philosopher who has written widely in applied moral philosophy, has given an authoritative voice to this concern. She has argued that in order for examples to be of value in applied ethics, they have to help us specify our ethical problem in a way that takes disagreement about possible solutions seriously. It is only by comparing different cases, and identifying what is problematic in all of them, that we can formulate a moral problem in a way that allows for an open discussion and leaves several options open:Footnote ²⁹ “The reflexive activity that is needed if moral problems are to be specified in a serious and non question-begging way is no more dispensable than a theory of practical reasoning.”Footnote ³⁰

Quite independently of O’Neill’s systematic and constructive approach, we can argue that, like rhetorical examples, individual cases can only serve their specific purpose if they are used in a wild, as opposed to a tamed, way. Wild cases are described in a way that stresses their uniqueness, rather than the normal and common aspects of the situation. Also, they are described in a way that emphasizes certain idiosyncratic details, rather than their main features, although their characterization does not—and indeed cannot—aim at completeness and closure. Lyons describes two important features of rhetorical examples: their undecidability—they are always somewhat open ended—and their excess; as cases become more complex, “the number of other concepts that can be illustrated by the narrative begins to threaten the control of the generality.”Footnote ³¹ These two characteristics are true of wild rather than tamed cases. The difference between tamed and wild cases, therefore, is not one of nature but one that concerns the way cases are described and used. First, the description of wild cases stresses details that make them unique and particular. Such—albeit trivial—details encourage a first-person perspective and thus contribute to challenging received views: like exemplars in rhetorical theory, wild cases can have a stronger emotional impact on their potential audience. Second, the description of wild cases emphasizes moral dilemmas, uncertainties, and potential conflicts, rather than conformity with an already-established claim. To that extent, they inspire not so much particular actions—on the contrary, this is a possible misuse—but rather novel perspectives and questions. An example will highlight the difference between wild and tamed cases.

In a recent article, Hilde Lindemann uses what we have called a wild case in order to deal with end-of-life decisions and advance directives, in a case in which proxy opinion and medical judgment are in conflict. The case is described in detail: the presence of a cat, for example, is highly relevant; dialogues between the proxy and the patient are reported verbatim; and the story is long winding, citing events that might seem secondary with respect to the final decision of whether to stop or continue resuscitation. More importantly, all the options are analyzed in a way that leaves many doubts about what the right solution is: “So should Edmond be waked?” the author writes—”I honestly can’t say.”Footnote ³² However, an original justification for giving more weight to a proxy’s opinion slowly emerges at the end of the analysis, adding a new element to the ongoing debate on the justification of advance directives. It is the intrinsic value of the intimate and long-term familiarity between the patient and the proxy who testifies about the advance directive that makes the directive binding, and not the fact that the directive expresses the patient’s autonomy. Here too the author expresses a tentative conclusion: although it is true that the proxy’s job, “maintaining another identity, is morally valuable work,” it can also give a “tremendous amount of power over the other.”Footnote ³³ Insight, and not outright decision, is what wild cases can offer us. We may contrast the use of this case with another more classical one. In a recent article, Stephen Bonner describes a case in which doctors rightly did not abide by the patient’s advance directives to withhold treatment, because they were not certain of the life-threatening nature of the condition; the reasons for the suicide attempt were not sufficiently clear, and her prognosis was good. It is not my purpose to discuss the case but only to note that it was sufficiently sketchy and free of emotional details (tamed) to allow for a definite conclusion: “We thought that this justifies treatment despite the advance directive.”Footnote ³⁴

As we have said, in the rhetorical tradition, examples are not only argumentative tools but also models for action; they are not only examples of but also exemplars. In this sense, cases are often understood either as personal models worthy of being emulated or as effective figures of speech designed to render an argument more persuasive. I would like to suggest that the importance of wild cases is even more substantial than the rhetorical discussion of exemplars would have us believe. Wild cases become the object of strong attitudes, emotions, and thoughts, and therefore they focus our attention on the unseen features of a given situation that deserve to be scrutinized further. Indeed, the rhetorical function of wild cases goes beyond the rational persuasion that tamed cases, as arguments, help us achieve. Insofar as they are vivid and perspicuous, they produce a rhetorical effect of amplification that, in turn, provokes an emotional identification with the character who is the subject of the case. This identification does not only produce imitation but provides a keener understanding of the lived experiences of those involved in an ethical dilemma and thus permits a first-person perspective on an ethical issue.

As a consequence, wild cases have both a debunking effect and a heuristic function: they can raise doubts about received views, allow us to see ethical problems in a new light, and encourage us to take seriously different normative perspectives, in a way that is not question begging. The evocation of idiosyncratic detailed cases can cast doubt on the most entrenched ethical beliefs. Even on such a controversial issue as reproductive cloning, emotions conveyed by the lived experiences of hypothetical (and still improbable) candidates might prompt a new examination of the issue.Footnote ³⁵ In justifying a reasonable role for emotion in ethical reasoning, Gregory Kaebnick writes: “We may also find that our initial misgivings diminish as we think more about cloning and how it might be used—with what hopes and goals, and to what effects on human lives. Spinning out some plausible stories might change our feelings about it.”Footnote ³⁶ This is not to say that the evocation of one striking and emotionally charged case should be sufficient as an argument for changing our current views and should thus lead us to adopt an ad hoc principle in favor of the practice.Footnote ³⁷ Rather, such a case would have the effect of opening the debate to new questions, of allowing our moral imagination a little more leeway, and of enabling us to take more seriously other points of view.Footnote ³⁸ It has been argued that examples in the moral domain contribute to establishing what is preferable in general, precisely by creating a preferring attitude concerning a single action in the audience.Footnote ³⁹ Eventually, from careful consideration of an emotionally gripping wild case, new concepts and approaches may emerge. Interestingly, today some advocates of the rehabilitation of case reports as a legitimate method of medical research, on par with randomized control trial and observational studies, advance a similar reason: “One hallmark of case reporting is to recognize the unexpected. Rather than representing the bizarre, the unexpected is where discovery begins.”Footnote ⁴⁰