Judgement and the role of the metaphysics of values in medical ethics

Abstract

Despite its authors' intentions, the four principles approach to medical ethics can become crudely algorithmic in practice. The first section sets out the bare bones of the four principles approach drawing out those aspects of Beauchamp and Childress's Principles of biomedical ethics that encourage this misreading. The second section argues that if the emphasis on the guidance of moral judgement is augmented by a particularist account of what disciplines it, then the danger can be reduced. In the third section, I consider how much the resultant picture diverges from Beauchamp and Childress's actual position.

A visiting ethics adviser at the start of a recent week focusing on values in medicine at the University of Warwick Medical School said

On balance, the four principles approach has done the teaching of Medical Ethics more harm than good.

Her objection was that, in practice, the four principles approach encouraged an assumption among medical students and doctors that medical ethics were algorithmic, robotic and required no element of judgement in the sense of having good judgement. This in turn risked abusive practice. But, given that the authors of the four principles approach explicitly argue against such an interpretation, why is it thought to be algorithmic and to preclude the exercise of judgement?

I will argue that this impression is encouraged by the absence of a proper account of the metaphysics of values. Beauchamp and Childress's Principles of biomedical ethics¹ stresses the practical use of the four principles for guiding judgements: the epistemology rather than the ontology of values, which is useful for medical ethics teaching. But in this case, a little more metaphysics would have a positive effect on practice.

In the first part of this paper, I set out the bare bones of the four principles approach, drawing out those aspects of Beauchamp and Childress's Principles of biomedical ethics that encourage the criticism mentioned above. My aim is not to give a balanced reading of that work so much as a typical reading. In the second section, I argue that if the emphasis on the guidance of moral judgement is augmented by a particularist account of the metaphysics of values, then the danger of an algorithmic interpretation can be reduced. In the third section, I consider how much the resultant picture diverges from Beauchamp and Childress's aim.

The four principles approach as an algorithmic method

Tom Beauchamp and James Childress in their Principles of biomedical ethics set out the four principles approach at length.¹ It is, to a first approximation, a principles‐based approach to ethics or a form of principlism. Ethical judgements are informed by principles in a “top down” manner (Beauchamp and Childress, p 385).¹ One of the contrasts to this is a bottom‐up or case‐based approach. I will return to this idea in the second section and to Beauchamp and Childress's view of it in the third section.

Beauchamp and Childress's principlism is indebted to the work of the early 20th century British philosopher WD Ross.² They say that he had “more influence on the present authors than any twentieth century writer” (Beauchamp and Childress, p 402)¹ and in particular that “Ross's distinction between prima facie and actual obligations is basic for our analysis” (Beauchamp and Childress, p 14).¹

According to Ross, moral duties are encoded in general principles. Each principle imposes a “prima facie” duty, a duty that would be obligatory, all other things being equal, that is, if no other principles were to apply. Although the principles encode prima facie duties, the obligation to act in a particular situation requires an actual or concrete duty: “the‐all‐things‐considered” duty imposed by the situation. This turns on the interplay of the principles—possibly a subset of them—that are relevant to the case. But because the different principles can pull in different directions, the actual duty depends on which duty is the strongest in the situation.

Ross himself proposed seven such duties: fidelity, reparation, gratitude, non‐injury, harm prevention, beneficence, self‐improvement and justice. Beauchamp and Childress's four principles aim at highlighting a smaller number of universal duties that are more relevant to medicine. Nevertheless, even with this smaller number, the principles can conflict.

As this is a form of principlism, if the principles conflict, then some method of determining the actual duty that they underpin should be available. Beauchamp and Childress describe two methods for coping with conflict between principles: specification and balancing. (Ross himself does not offer an account. This is Dancy's key criticism of him.³)

Specification is a way of deriving more concrete guidance from the fairly abstract higher‐level principles. It is described in outline thus:

Specification is a process of reducing the indeterminateness of abstract norms and providing them with action guiding content. For example, without further specification, “do no harm” is an all‐too‐bare starting point for thinking through problems, such as assisted suicide and euthanasia. It will not adequately guide action when norms conflict. (Beauchamp and Childress,¹ p 16)

This seems at first to be a kind of deduction. In much the same way as Kepler's laws of planetary motion can be (more or less) derived from Newtonian physics once particular assumptions are made, so a specified rule can be derived from a higher‐level principle. Just as Kepler's laws are useful in the specific context of planetary systems, so a specified rule—such as that doctors should put their patients' interests first—can be tailored to give concrete guidance to cases of euthanasia, for instance.

Although specification is some form of derivation, it cannot strictly be termed ‘deduction' because specified lower‐level rules have more content than the principles from which they are drawn. The validity of deductive inference, by contrast, is underpinned by the fact that a conclusion cannot contain information that was not already in the premises. As specification is ampliative, with more content generated, it cannot be deductive. Nevertheless, however precisely the four principles give rise to more concrete specifications in rules, the rules arrived at help generate actual duties when applied to concrete situations.

What is more, Beauchamp and Childress suggest that specification can dissolve an apparent conflict between principles by narrowing the scope of the principles in play. Two specific rules derived from principles which, in a particular circumstance, appear to conflict may themselves be consistent. Specification thus fits the general aim of principlism: to encode moral guidance in general norms (principles and rules in the terminology here). So far, it appears to underpin an algorithmic approach to moral judgements. Armed with suitable specifications, actual duties can be derived.

The second tool for generating an actual duty from apparently conflicting principles is, in principle, less algorithmic. It is balancing and complements specification:

Principles, rules and rights require balancing no less than specification. We need both methods because each addresses a dimension of moral principles and rules: range and scope, in the case of specification, and weight or strength, in the case of balancing. Specification entails a substantive refinement of the range and scope of norms, whereas balancing consists of deliberation and judgement about the relative weights or strengths of norms. Balancing is especially important for reaching judgements in individual cases. (Beauchamp and Childress,¹ p 18)

In fact, despite the talk of deliberation and judgement here, Beauchamp and Childress go on to specify six conditions that have to apply before one principle can be taken to outweigh another. In other words, they deploy a further codification or set of rules that plays down the role of individual judgement or sensitivity to the case. The suggestion is that the set of rules is needed to protect ethics from “purely intuitive or subjective judgement” (Beauchamp and Childress,¹ p 21).

Nevertheless, Beauchamp and Childress do suggest that balancing requires a degree of judgement (in the non‐algorithmic sense I intend) that goes beyond the rules they provide. They also think that genuine moral dilemmas will remain after the principles have been deployed, calling instead for an exercise of judgement. But it is all too easy to lose sight of the role of judgement in both the teaching of medical ethics and in discussions of the four principles approach in journals for the following reasons.

One problem is that standard cases come to dominate teaching and discussion. Thus, for example, the case of the Jehovah's Witness who competently refuses life‐threatening treatment is taken to exemplify the conflict of beneficence and autonomy, and on the standard solution, autonomy is rightly taken to dominate.⁴,⁵ (Things differ in the standard case of his or her young child.) The case is sketched in abstract and ideal terms (in contrast with the rich detail of real cases) and itself becomes a kind of rule to be applied to further actual cases. Competence in first solving such standard cases, by applying the principles and giving them standardly approved weight, and then further applying the standard cases as analogies for further cases becomes second nature to medical students keen to pass their ethics course. Thus the element of individual judgement and the need to inspect the details of real cases is played down. This is the concern highlighted by the visiting ethics adviser, with which I began this paper.

A second problem is that there is no clear account of what kind of skill is required in balancing principles. Beauchamp and Childress think that if balancing is not itself subject to further rules, it will be merely subjective. This seems to suggest that if balancing is not an expression of mere subjective whimsy, it must somehow be guided by principles in an inchoate way, which again suggests that it really is an algorithmic matter.

In summary, a cursory reading of Principles of biomedical ethics with the practicalities of and real constraints on the teaching of medical ethics encourages an algorithmic approach and hides the element of skilled uncodified judgement required. Beauchamp and Childress concentrate on a top‐down approach that seems to dictate actual duties almost deductively (“almost” because specification is ampliative). They believe that when principles cannot be given, there is a danger of subjectivism. This again highlights the concern expressed by the visiting ethics adviser. But it can naturally be eased by thinking a little more about what is being decided when conflicting principles are balanced.

Adding an account of the metaphysics of values

In this section, I suggest that a particularist view of the metaphysics of values helps ease the worry that the four principles approach is algorithmic in practice. I will attempt to motivate the idea of adding a particularist ontology of values to the four principles here and examine how close this actually fits the view taken in Principles of biomedical ethics, in the final section of this paper.

Beauchamp and Childress introduce the four principles primarily as a tool to offer guidance for medical ethics. They say, “These principles can… function as guidelines for professional ethics” (Beauchamp and Childress,¹ p 12). The emphasis of the book is practical: on specifying duties in the light of the principles. But in moral philosophy in general, guidance is only one role for principles in a broadly principlist approach to moral values. (Ross himself does not suggest that principles have a guiding role.)

In addition to asking what guides moral judgement, we should also ask what disciplines it. What, in other words, makes the difference between a judgement being correct or incorrect or at least more or less correct? Whereas the question of guidance is an epistemological one, that of discipline is ontological and helps characterise what sort of judgements are in play.

Putting aside the subtleties of recent moral philosophy, there are three broad answers to the second question.

1. Nothing: If nothing disciplines value judgements then there is no difference between right and wrong, correct and incorrect. Beauty, on a popular view, is like this. If beauty is merely in the eye of the beholder, then what is in the eye answers to nothing and there is nothing that disciplines aesthetic responses. Beauty is nothing substantial.

There may be philosophical reasons for adopting this view. Some philosophers are driven to it because they find neither of the other answers I will set out plausible. Some neo‐Humean philosophers consider the surface form of ethical judgement to be systematically erroneous; others think that it at least requires reinterpretation. Such approaches may or may not reconstruct an ersatz notion of discipline for ethical judgements. But they are revisionary in spirit and for that reason I will ignore them here. There is a practical justification for my cavalier attitude. If one of the aims of medical ethics teaching is to foster respect for genuine evaluative constraints on clinical judgement, it will hardly help if one promotes an error theory of ethics.

2. Principles: What makes the movement of a chess piece—its bare displacement along a vector, a chess move—a legitimate action within the game? A necessary (but not sufficient) element of the answer is its accord with the rules of chess. Those rules or norms set the standard of correctness for moves and thus partly constitute the very idea of a chess move. Chess is merely a human institution but a similar approach is plausible in the case of arithmetic judgements whether the rules are taken to be the informal ones in which elementary students are drilled or the formalisation in Peano's axioms.

In both these cases, rules or principles function as epistemological guides. To determine whether to believe that a movement is a legitimate chess move, or what the answer to an arithmetic operation is, we can look to the rules. But the rules also determine what correctness consists in and thus the “ontology” of chess and maths: chess or mathematical facts. The facts can be marshalled into deductive structures.

According to principlism in ethics, ethical principles serve both these functions. While guiding moral judgements, they set the standard of correctness and incorrectness for those judgements. But this raises a substantial challenge. If the principles discipline judgements, then they should meet two conditions: they should give the right sort of guidance to resemble ethical or moral principles and they should be consistent. The problem is that these two conditions pull in different directions.

A single universal principle such as Kant's categorical imperative can be consistent in itself, but it does not seem to capture the whole of moral reasoning. In particular, it does not give practical guidance. Beauchamp and Childress, following Ross, argue that the only realistic approach is to accept that there are several principles that govern moral judgement and which thus offer guidance. But, as they realise, the principles conflict, and if they are to serve the role of disciplining moral judgements, such conflict is more than just an epistemological complexity. It threatens the idea that all that makes a moral judgement correct or incorrect is in accord with a set of principles. If principles were the only answer to the question of what disciplines value judgements, then the threat would be one of subjective whimsy (akin to Beauchamp and Childress's worry about “purely intuitive or subjective judgement” in response to which they deploy further rules to codify balancing the four principles). Something else is needed to explain how to balance the principles in the face of a particular situation and thus explain what the aim of balancing is.

3. Particulars themselves: If not ‘‘nothing'' and not ‘‘accord with principles'' what answer can we give to the question, what disciplines a judgement? A third candidate is “the facts”. Now, given the general contrast of fact and value, that answer may not seem promising here. Nevertheless, the idea is that, rather than thinking that a particular situation merely prompts a balancing of conflicting principles which themselves discipline a judgement, one should think that the aim of the judgement is just to get the situation itself right.

On this third view, the situation itself contains evaluative features—values—and ethical judgement aims at describing ethical values. I will call this view “particularism” and return in the final section to discuss the kind of particularism it amounts to.

I do not wish to play down the number of potential objections to a particularist view of ethical judgement. It requires accepting, for example, that the world itself is partly made up of features that can be understood only from a particular standpoint and that appreciation of these features can have a role in motivating action without an additional desire (contrary to the longstanding view of the practical syllogism as containing beliefs and desires). Nevertheless, these objections can be dealt with.⁶,⁷,⁸

It is, however, worth saying a little more about the first problem. This is a characteristically philosophical problem. The intuition behind it is this: real or genuine properties are those that can be described from no particular perspective. Modern science has been successful partly because it has moved away from local or subjective accounts of the world to an increasingly objective, or perspective‐independent, account. A vivid philosophical articulation of this is Bernard Williams' absolute conception: the conception of the world arrived at by a completed science.⁹ The basic properties of contemporary physical sciences will perhaps form part of the eventual absolute conception of the world, but secondary qualities like colours and smells will not. These are mere artefacts of our local perspective, including our specific biological constitution. Values are even less likely to be part of the absolute conception as their detection requires not just a particular biological constitution but also a particular local form of life. Only those who broadly share our interests, practices and customs will share our values.

Although the idea that what is real requires no particular perspective, mind or form of life to conceptualise or detect it is attractive, it is not mandatory. Resisting it requires partly distinguishing the methodological efficacy of natural science from its role as the touchstone of all that is real. Although the assimilation is attractive because of the success of science, it is a non sequitur. Although the natural sciences successfully describe aspects of the world, especially those aspects that fit within a more or less deductive system of natural laws, this need not exhaust reality, which may also contain the truths of history and literary theory, for example, that cannot be fitted into any such system.¹⁰

But what is the alternative view of values? How can values be part of the fabric of the world? Again, much more would need to be said. A first step is to resist the philosophical initial move that makes this seem fishy in the first place: thinking that those features of a social situation that are real are those that are describable in natural scientific terms. Without a prior philosophical bias, however, rejecting this is the natural way of viewing social situations. A full account is not just a matter of bodies in space describable in natural scientific terms but of people: people with their own interests, perspectives, values, concerns and expectations. These features of the social world impose obligations and duties on those with eyes to see them, and especially, in the case of medical ethics, on doctors. The obligations logically come before any attempts to regiment them into the deductive structures: the project of principles‐based normative ethical theory.

This emphasis on the central role of values realised in situations rather than on principles‐based theories fits (and augments) an approach to clinical judgement called values‐based practice, which also emphasises the importance of a range of different values for clinical judgement in addition to specifically moral and ethical values.¹¹ It also dovetails with an aspect of their education that medical students find natural. Clinical judgement requires the integration of medical scientific concepts into observation and perception. The skilled clinician sees more in an x ray film than a lay person. Similarly, a skilled practitioner of values‐based practice sees more in a social situation than a neophyte.

The advantage of this particularist approach to the ontology or metaphysics of values is that it helps provide a context for balancing principles in the four principles approach. These can now be seen merely as guides to judgement, which itself aims at getting the evaluative features of concrete situations right. The situations, rather than a combination of principles, make the judgements right or wrong. As mere guides, principles can be more or less useful, depending on the situation at hand. But as reminders of the kind of factors that it is useful to take note of (think of the autonomy of the subjects involved, avoid unnecessary harm, etc), they have a role both in practice and in education.³

Why would this emphasis help avoid the risk that the application of the four principles is, in practice, algorithmic? Because it reduces the idea that value judgement is a closed game with a limited number of factors to be considered. By making real situations central to the actual practice of ethical skill, and principles secondary, it encourages a realistic view of the complexity of judgement. Situations are not merely constituted by the interplay of the four principles, but instead independant of and serve to test those principles.

It also encourages a critical approach to the four principles themselves. Perhaps there are better tools to chart the space of values than just those principles. Although Beauchamp and Childress are open to this possibility—they, for example, see the need to defend their view—it is apt to be lost in medical ethics teaching and subsequent practice.

In this section, I have suggested that the four principles approach is usually explicitly introduced as an answer to the question of what guides, or should guide, ethical judgement, but that in the wider area of moral philosophy principlism is also supposed to answer the more fundamental question of what disciplines such judgement. Principlism has difficulty answering this question because of the dual challenge of showing that the principles selected capture sufficiently what we prephilosophically mean by ethical judgement and thus can offer practical guidance, but also that they are consistent. By giving a different answer to the discipline question we can sidestep this problem and at the same time reduce the temptation to construe the four principles approach as circumventing the need for judgement.

Is particularism consistent with Beauchamp and Childress's four principles approach?

In the previous section, I distinguished between two questions to which moral particularism may be an answer. One was, what guides ethical or value judgement and the other what disciplines them. I outlined the attractions of a particularist answer to the second and suggested that it also helps emphasise the central role of good judgement in medical ethics and value judgement more broadly. In this section, I will return to the question of guidance to clarify the nature of particularism and to explore its relation to what Beauchamp and Childress themselves say about “bottom‐up” approaches.

According to Smith, in a recent Journal of Medical Ethics paper,

Moral particularism can be understood as propounding a “case‐by‐case” methodology in dealing with moral questions …

What motivates particularism … is hostility to the manner in which agents tend to deliberate about a particular case as though it must be considered always and only with reference to a relevant principle … The moral particularist complains that this “intellectual” way of reaching a moral conclusion can serve to obscure the potential complexity of the particular case.¹³

This summary emphasises guidance rather than discipline. Value judgements, according to particularism, should be guided by close attention to the case at hand, rather than by appeal to principles. According to the most influential recent exponent of particularism, UK philosopher Jonathan Dancy,

The primary focus of particularism is the particular case, not surprisingly. This means that one's main duty, in moral judgement, is to look really closely at the case before one. Our first question is not “Which other cases does this one best resemble?”, but rather “What is the nature of the case before us?” (Dancy,³ p 63)

In other words, in addition to the idea that judgements are made right or wrong by reference to the situation rather than by reference to principles, principles also have (and should have) little role in deliberation. How different is this approach from the combination I suggested in the previous section, a situation‐based account of discipline that allows some use for principles in answering the question of guidance?

Dancy argues for a radical difference. He argues, against Rossian generalism (of which, as we have seen, the four principles approach is an instance), that principles do not succeed even in capturing prima facie duties. In appropriate circumstances, a principle that apparently encodes a prima facie duty can invert its normal force or “valency”. Consider the idea that all other things being equal, promoting happiness is an ethical good. Now imagine two equally competent dentists who both scrupulously minimise patient discomfort and cause only the same small unavoidable amount. One of them, however, takes great pleasure in the patients' residual pain. To whom should their practice manager favour sending patients? According to some ethical intuitions, we should not promote the sadistic dentist's happiness. If so, in this situation, promoting happiness has the valency opposite to normal. Dancy claims, therefore, that Ross's idea that actual duties are forged from prima facie duties that always have the same “all other things being equal” force is mistaken.

Dancy's emphasis on looking very carefully at the particular case rather than comparing it with others, however, may become too radical. To take that injunction at its most radical would preclude the description of particular situations using any general descriptions and concepts that also apply to other cases. But without at least conceptual connections between one case and another, every case looks like an instance of what Wilfred Sellars terms the Myth of the Given.¹³ It looks as though experience of a particular case may be supposed to comprise a foundation for judgement that is fully independent of any other judgement on other cases. But Sellars has convincingly argued that the foundationalism of the Myth of the Given is untenable. There can be no such foundational judgements. So the question is, is there a middle ground between the rejection of moral principles encoding prima facie duties and the rejection of moral concepts that enable us to judge whether one case is relevantly similar to another? I will not attempt fully to adjudicate that debate here.¹²,¹⁴,¹⁵ But examination of the motivation for Dancy's particularism suggests that there need not be a problem.

Dancy's main argument for particularism is what he calls “holism in the theory of reasons”. He argues:

The leading thought behind particularism is the thought that the behaviour of a reason (or of a consideration that serves as a reason) in a new case cannot be predicted from its behaviour elsewhere. The way in which the consideration functions here either will or at least may be affected by other considerations here present. So there is no ground for the hope that we can find out here how that consideration functions in general, somehow, nor for the hope that we can move in any smooth way to how it will function in a different case. (Dancy,³ p 60)

Although this is not Dancy's way of putting the argument, an analogy with a different area of philosophy helps illustrate the main point.

Consider the more familiar debate between theory theory and simulation theory in response to the problem of other minds.¹⁶ According to theory theory, we have knowledge of other minds through a theory of mind. In much the way that a theory mediates between observation of patterns in a cloud chamber and a judgement of the underlying particles causing them, a theory of mind mediates between observed behaviour and underlying mental states. According to simulation theory, by contrast, things are not so intellectual. We need to imaginatively put ourselves into another person's predicament and see what mental states we would form. On the basis of this view, we need merely possess a mind, not a theory of mind, to know another's mind. The debate, however, is not as clearcut as this because, to be plausible, the theory proposed by theory theorists needs merely to be tacitly known. If so, then perhaps the ability to simulate can be described in just such a theory. Thus the two sides may be closer than they at first appeared.

Nevertheless, there remains one reason to hold that the approaches are distinct and thus that the ability to simulate cannot be encoded in a theory. Imagining ourselves in another's situation requires, in addition to thinking of their motives and desires, assessing what it would be rational to believe in the light of their other beliefs. This requires assessing the interplay of reasons in context. Although there may be guidelines that help govern beliefs in some areas, they will in general be sensitive to other factors. A cow‐like visual appearance may usually indicate the presence of a cow. That may approach the status of a principle. But in the context of an antecedent reason—perhaps a well‐founded general zoological theory—to believe that there are no cows present, but there are other similar animals or replicas, that principle will no longer hold.

Simulation theory is thus a form of particularism. It denies that reasons can in general be governed by principles. But it need not require an epistemological foundation, an experiential intake that does not also depend on our other beliefs. Imaginatively putting ourselves into someone else's position—both literally and figuratively—requires imagining how the world looks from there. That requires an exercise of judgement on the interplay of reasons available, but it does not imply any form of Given. By analogy, there seems no reason to think that particularist judgement in the ethical case need subscribe to that myth either.

How does such a moderate epistemological particularism fit with what Beauchamp and Childress themselves say about bottom‐up approaches? They focus on Albert Jonsen and Stephen Toulmin's casuistry, which they characterise as a focus on an “intimate acquaintance with particular situations and the historical record of similar cases.”¹ This fits the sketch of particularism given above. But unlike radical particularism, casuistry at least relies heavily on reasoning by analogy. “The leading cases—often called “paradigm cases”—become the most enduring and authoritative sources of appeal in the underlying consensus … Decisions reached about moral rights and wrongs in these seminal cases serve as a form of authority for new cases”(Beauchamp and Childress,¹ p 394). They go on to criticise it saying that it is a “method without content”(Beauchamp and Childress,¹ p 395) and that it requires ethical principles to serve as the initial premises for analogical reasoning. As a result, they suggest a combined approach: “Principles need to be made specific for cases, and case analysis needs illumination from general principles” (Beauchamp and Childress,¹ p 397). (This is why their position is principlist only to a first approximation.)

What do they mean by this? Normally, by combining opposing philosophical positions we run the risk of contradiction. But, given the distinction between the questions of guidance and discipline sketched above, it is possible to suggest in more detail than Beauchamp and Childress themselves provide what this may require. A particularist answer to the latter can be combined with a principlist answer to the former. As the questions are independent, the result is consistent.

The position, however, need not be made up in quite that neat way. A particularist answer to the latter question can be combined with a position that is neither strongly particularist nor principlist with regard to the former. The analogy with the problem of other minds helps fill this out while reinforcing the need for good judgement. Being responsive to evaluatively complex situations requires being sensitive to the demands of evaluative reasons. This includes partly working out the combined effect of different factors that are in play at the same time. Principles may help in this. They can serve as reminders of factors that can often be important. But, just as we do not form reasons for belief in accord with an explicit or implicit theory of best evidential practice, so good value judgement, in what we can call values‐based practice, turns on a skilled intellectual appreciation of the case at hand. The price of understanding the central importance of good judgement—in both values‐based practice, broadly construed, and more specifically with respect to medical ethics—is distinguishing between the questions of guidance and discipline for value judgements and adopting a particularist ontology of values realised in situations in response to the latter.