1. Introduction

1This paper concerns how the normative order achieves both justification and stability. Generalism argues that justification lies in generals that, despite Bernard Williams’s persuasive doubts, can be found to have essential qualities. The new particularists have shaken analytical ethics with an opposing view, but they appear to accept instability as inevitable. This standoff gives pragmatism an opening. The metaphilosophical position is that justification and stability are contingent and must be achieved together. Stable answers to normative questions result from what Dewey called an “empirical naturalism,” that often comes from resolving social conflict. The evidence from law and legal inquiry is that Dewey’s continuum of inquiry is more dependent than he recognized, or at least explained, on the outcome of battles among competing groups.

1.i. Analytical Theory and Pragmatism

2The author of the essay on “American Moral Theory” in Cheryl Misak’s 2008 Oxford Handbook of American Philosophy is Brad Hooker. Hooker has an impressive command of contemporary moral theory, but he is not a pragmatist. Pragmatism gets faint attention, and is mentioned only briefly, with little account of its moral theory. The essay begins with a summary listing of Dewey’s works; Hooker says, “His interest in human psychology, his hostility to dogmatism, and his commitment to instrumentalism and experimentation were deeply influential in American thought in general, and moral and social philosophy in particular.” Rather than elucidating that influence, Hooker continues, “the giants of early twentieth century moral philosophy were British, not American.” Of Dewey he concludes: “Whereas Dewey was a naturalist in moral metaphysics, an experimentalist in moral epistemology, and a contextualist about what questions need answering, G. E. Moore, H. A. Pritchard, W. D. Ross, and C. D. Broad followed Henry Sidgwick in espousing non-naturalism in moral metaphysics and foundationalist intuitionism in moral epistemology” (Hooker 2008: 578-94).

4Revival of interest in ethical particularism has drawn attention for its stark contrast to the dominance of generalism, lately criticized by Bernard Williams and other writers. Pragmatists may have themselves to blame for being left out of this conversation. They have failed to meet analytical theory head on, to address normative problems like the trolley dilemma in a fully naturalist and experimentalist vein, and to tackle the synchronic analytical account of normative logic for its unrealistic representation of actual moral experience.

5The rise of ethical particularism presents an opportunity. The perceived rivalry of moral particularism and generalism will prove unsustainable, and will open up space for pragmatist viewpoints. This essay contends that the key to pragmatism’s response to analytical moral theory is to be found in Dewey’s continuum of normative inquiry.

6Viewed as part of an experimental continuum, particularism and generalism are stages of inquiry into real, not imaginary, moral dilemmas, which analytical theory conflates in its static vision of ethical inquiry. The continuum of inquiry was not fully apparent until Dewey’s later works, especially Logic: The Theory of Inquiry, and it remains underappreciated even by many of his own advocates. It was anticipated by O. W. Holmes Jr. in The Common Law of 1881 (Kellogg 2021: 1, 3, 43, 48, 58-9).

7Real moral dilemmas represent actual conflicts, that cannot be understood through the analysis of cleverly constructed moral puzzles. Repeated over time, real problems drive the consensual formation and revision of social practices and the predication of general moral rules and principles. The real continuum of normative inquiry is unrecognized in the analysis of hypothetical dilemmas, like the trolley problem, where the natural context of a problematic continuum is entirely removed. It is only through actual normative engagement, such as can be studied in law, that the continuum is clearly on display.

8Early in the continuum, as new conflicts are presented for de novo judgment, opposing reasons are unprioritized. They are gradually compared through comparative reasoning from experience, as generalization emerges with practical usage, eventually fixed through adaptive practice. The eventual body of moral knowledge is thus a mixture of settled and unsettled inquiry, a patterned fabric of tentative as well as entrenched propositional domains. In this regard it presents a “pluralist” ontology.

9The paper proceeds in two parts. The first part (section 2) is an analysis of Jonathan Dancy’s argument for ethical particularism, hinging upon his claim of “reasons holism.” I will explore his use of examples, and compare them with the role of hypothetical dilemmas in analytical ethical theory, beginning with Philippa Foot’s example of the snake (“it is dangerous to pick up a snake, but it may be more dangerous not to pick up this one”). I distinguish the snake example from the emergence of adaptive practice in the traffic circle or roundabout, in order to demonstrate the retirement of reasons and show how constructed hypothetical dilemmas are isolated from the continuum of experience. The second part of the paper (section 3) is a discussion of the trolley problem in the context of the continuum of normative inquiry, in which I examine the process of normative induction, the creation of obligation from emergent practices, and the actual social path of inquiry that leads to normative predication.

2. Particularist “Holism” vs. Pragmatic “Retirement” of Reasons

I suspect […] that there is no way of cashing out propositionally the ways in which non-evaluative properties contribute to the evaluative natures of situations, actions, characters. […] The particularist’s claim is that the good-making relation cannot be cashed out in propositional form. (Margaret Little 2000: 283, 285)

10If this characterization of moral particularism by Margaret Little is correct, and if the statement is true, how, then, is moral knowledge communicable? How is it motivational? How is it even possible?

11This first section will explore the context of Little’s claim, ethical particularism, and the proper response of pragmatism. Her claim is grounded in reasons holism: as Jonathan Dancy claims, “holism holds that every consideration is capable of having its practical polarity reversed by changes in context.” Dancy’s error lies in conflating the difficulty of de novo judgment, in which the scope of relevant reasons is broad, with the late stage of a problematic situation in which practices have gradually emerged, restricting the scope of relevant reasons and rendering deduction from a general proposition definitive.

12Moral knowledge is of course communicated in general propositions. The latter can be seen as derived from the repetition of particular experience. Thus, I characterize particularism and generalism as stages of normative inquiry, rather than rival accounts of moral knowledge and motivation. The generalism-particularism debate fails to recognize a crucial difference between individual moral choice in unique situations (it is dangerous to pick up a snake, but it may be more dangerous not to pick up this one), and those in common problematic situations (such as entering a traffic circle), where emergent practice drives generalization; hence both approaches are imperative for social order.

13The example of the snake came from Philippa Foot. She had already invented the precursor of the now-famous trolley problem in demonstrating the reversible polarity, or valence, of intention: “Beside this example [the judge faced with an unruly mob that can be appeased by convicting an innocent man] is placed another in which a pilot whose aeroplane is about to crash is deciding whether to steer from a more to a less inhabited area. To make the parallel as close as possible it may rather be supposed that he is the driver of a runaway tram which he can only steer from one narrow track on to another; five men are working on one track and one man on the other; anyone on the track he enters is bound to be killed” (Foot 1978: 23).

15Such hypotheticals, as well as Dancy’s everyday dilemmas (having to return a book to someone who stole it from the library, or break bad news to his sister, 1993: 60, 116), serve to highlight aspects of individual moral conflict, showing that they cannot be reduced to or controlled by a universal general rule or principle; hence they appear in Dancy’s work to support a global particularism. But in cases of successive experience with similar conflicts, as in medically assisted death, the practical resolution of real dilemmas takes precedence.

16My argument begins with an analysis of “reasons holism,” a central tenet of particularism. The holism of reasons is characterized as the thesis that what functions as a reason in one case may not, in another case, be a reason at all, or may even be a reason with the opposite valence: “[A]ll reasons are capable of being altered by changes in context – […] there are none whose nature as reasons is necessarily immune to changes elsewhere.”

17I claim from examples in law that while the holism of reasons can readily be found in the unique or de novo individual case, it will prove temporary in the case of public problems, where of necessity reasons must be retired for the emergence of general practices and the rules that embody them. I argue that this illuminates how the “good-making relationship” is in actuality “cashed out,” eventually finding consistency inductively from experience, rather than deductively from principles.

18Why does ethical particularism take the view, opposing moral generalism, that the evaluative can never be “cashed out” in propositional form (Little 2000: 283), that it is epistemically “shapeless” with respect to the descriptive (ibid.: 279), and that there is no possible descriptive pattern unifying the class of right acts (Jackson, Pettit & Smith 2000: 99)? It is a reaction against a form of generalism that makes an equally sweeping general claim that a limited set of reasons can always dictate consistent results. As contrasted with that claim, robust moral particularism emphasizes the multiplicity and variety of reasons for a particular ethical decision, reasons that can shift polarity from case to case (Dancy 1993: 57).

19Accordingly, “reasons holism” refers to the “context sensitivity” of relevant non-moral or factual properties. According to the holism of reasons, the list of such properties is long, and a consideration that is a reason in one context may not similarly be a reason in another, due to differences in the presence or absence of defeating and enabling conditions. This holism of reasons is claimed as an essential feature of all ethical choice. By comparison, as a classical generalist position, R. M. Hare’s doctrine of universalizability holds that when a particular action is judged morally wrong, this is so on account of a discrete set of properties. Consistency demands that any action that shares these properties is also wrong:

Universalizability can be explained in various equivalent ways; it comes to this, that if we make different moral judgments about situations which we admit to be identical in their universal descriptive properties, we contradict ourselves. (Hare 1981: 21)

20Ethical particularism has consistently contended, on numerous grounds, that the generalist’s “list of relevant properties” required to support universalization cannot coherently be limited.

21Generalists like W. D. Ross have recognized the problem of exceptions to every universal: no general can adequately account for all possible future cases. Particularism has more difficulty with Ross’s more moderate view, that ethical generals offer only prima facie support for moral decisions. Jonathan Dancy has responded that this pro tanto (or “for the most part”) position fails to account for defeated reasons in moral conflict, and for the “rationality of regret” that recognizes the unreduced epistemic status of the “defeated ought” (Dancy 1993: 111).

2.i. The Snake vs. the Roundabout

22An early example of the holist dimension of reasons was Philippa Foot’s snake example: she noted in 1983 that it is dangerous to pick up a snake, but it may be more dangerous not to pick up this particular one. Dancy comments, “The dangerousness of not picking up this one is not reduced by the dangerousness of picking it up,” and claims this demonstration of the “variable valence of factual properties” shows that the opposing reason “retains its full force” (Dancy 1993: 111, citing Foot 1983).

23However, consider Foot’s example in the context of circus snake charmers, who regularly gather venomous snakes for training and feeding. At first it is equally dangerous to pick one up, or not to; but through repeated experience a practice is established in which the parity of opposing valences is retired. The practice develops when the situation is repeated sufficiently within a given community to become a common problem. Circus snake performers will adopt, from experience, strict rules that govern when and how snakes are to be picked up. The equal dangers (valences) of picking up, or not, are removed by the practice. It is no longer the case, then, that (in Dancy’s words) “the defeated non-comparative ought can remain true” (1993: 111). The example suggests other occasions in which retired reasons do not retain their full force in defeat. Note, however, that in this instance the retirement of reasons takes place within a discrete community.

24Accordingly, we may consider the case of entering a traffic circle or the English “roundabout” – but before any rule regarding rights-of-way. The rules of the road have derived for an entire community in a fashion comparable to the snake charmers’ practice: from necessity driven by experience, by practices emergent from the experience of potential injury. In an early particularist world, there was no settled rule for when and how drivers first entered what might have been the first English roundabout. This is a world that was (as a phenomenologist might say) “pre-predicative” with regard to traffic circles. The holism of reasons may be characterized here as an initial condition.

25We can accept that the list of relevant properties governing normative conduct can be shown to be initially long, demonstrable by a hypothetical early roundabout accident requiring inquiry into the right, good, or just result. Imagine yourself as the barrister representing Ms Quickly, who was struck by a lorry when she blithely dashed into an early roundabout. Damage and injury resulted, and a judgment must now be made as to a just accounting. As there was no governing rule, your client has sued the owner of the lorry, arguing that the lorry driver should have taken proper action to avoid her. The lorry driver has counter-sued, claiming that she plunged recklessly into his path. But absent any rules of the road governing “plunging” or “avoiding,” that is not all that the two drivers can argue regarding justice or rightness. Both Ms Quickly and the lorry driver might extend the inquiry to any claim conceivably bearing on a better result.

26As her advocate you may bring up everything favorable to the moral balance of her case, regarding the question of “who should pay,” including her being late to church, her unblemished driving record, and her unimpeachable character, and likewise the character of the lorry driver and the fact that the lorry was carrying a cargo of scandalous literature. Counsel representing the lorry driver also has ample arguments from which to draw, including the economic importance of his client’s trip, effects on his career and livelihood, indeed the fact that your client is wealthy and he is poor by comparison. Moreover, the opposing advocate may “reverse the valence” of your own arguments regarding Ms Quickly, claiming that they weigh against rather than in favor of her case (“she should have known better”). Indeed, it is common among trial lawyers to reverse the valence of opposing arguments in close cases not controlled by settled rule.

27However, once the rule establishing a presumptive right-of-way for any vehicle already in the roundabout has taken hold, none of this is relevant; Ms Quickly’s action, and the class of all similar actions, is now clearly predicable as wrong, at least prima facie, and a judicial decision in her favor would ceteris paribus be considered unjust. Indeed, the various reasons Ms Quickly might entertain to justify her haste in entering the already occupied roundabout, such as her being late to church, are discounted by the combination of rule and established practice governing other drivers, and their reciprocal expectations, including that she will not enter the roundabout in such a manner as to require their emergency action. The ethical general derives motivational force from the entrenched rule and practice, and will override the individual reasons that she and other drivers may yet entertain about their own priorities.

28I should note for later reference that the very nature of Ms Quickly’s putative intent is transformed by the emergent practice. In the early instance, the accident might at worst be viewed as unintentional, or possibly just “indirect” (in the sense of Foot’s brakeless tram driver’s running down the workmen on the track ahead). Once the rule of yielding is established, however, it will more readily be seen as intentional on her part.

29We may also note for later reference that predication of Ms Quickly’s hasty entry at the later stage as “wrong” or “bad” is not deductively related to a universal theory of rightness or goodness. This elucidates Little’s observation that “the good-making relation cannot be cashed out in propositional form.” The “thin” moral properties of good or bad, right or wrong, are revealed here as resultant properties, originating independently of the properties from which they result. I will return to this point below, where we can anticipate particularism’s response.

30How and why did the roundabout rule come about? A solution had to be found lest crashes in the roundabout continue unabated (Holmes 1881: 113-23). The process was likely one in which cautious drivers tended to wait before entering the proliferating roundabouts, gradually establishing a practice. Meanwhile there would have been disputes to be resolved, in which arguments regarding specific litigants and their actions, purposes, relative wealth, etc., were increasingly ignored or rejected, until eventually legal authorities adopted a rule based on the preferred practice.

31In the early or novel legal case, just as in the contemplated individual act, there are aggravating and mitigating factors everywhere you look, on both sides, and moral particularism does not accept restrictive rules of relevant evidence. The various properties available as reasons and advanced as arguments are narrowed by the emergence of cautionary practices, and the process of normative induction compares them in a gradual, rather than immediate, process, seeking Hume’s essential ingredient of similarity.

32Hume wrote “[w]hen we have found a resemblance among several objects, that often occur to us, we apply the same name to all of them, whatever differences we may observe in the degrees of their quantity and quality, and whatever other differences may appear among them” (Hume 1896: 16). Normative generalism, we might then observe, is a product of establishing a commitment to finite and predicable (and “namable”) expectations.

33This is not to deny the illustrative value of particularist dilemmas. The famous trolley problem has demonstrated the variable valence of individual intentional states. A comparable real dilemma occurred with the prosecution in 1999 of physician Jack Kevorkian, convicted of second degree murder for providing a device to a patient who pushed a button releasing fatal chemicals. Kevorkian placed himself in the trolley position of tragic choice among alternative individual destinies. His case was initially unique, but the problem has demonstrated sufficient repeated practical similarity for inductive comparison and predication, such that relevant properties have been subject to the process of inductive inquiry. Kevorkian’s act initiated the now well-established social continuum of inquiry into assisted death (Sunstein 1999: 76 f.).

2.ii. Salience and the Retirement of Reasons

34Having illustrated what I have called the retirement of reasons, what reply would particularism give to this account? Jonathan Dancy insists that defeated reasons always retain their epistemic status, and emphasizes instead how the “salience” of certain reasons guides moral decision. He writes, “defeated reasons are the normal result of moral conflict, where we face reasons of some strength on both sides of a disputed question, and so the question becomes what sense can be made of moral conflict by the sort of theory of moral reasons that I have been beginning to outline” (Dancy 1993: 109). Here Dancy draws on an observation by Bernard Williams that formed the latter’s own critique of generalist theory:

It seems to me a fundamental criticism of many ethical theories that their accounts of moral conflict and its resolution do not do justice to the facts of regret and related considerations: basically because they eliminate from the scene the ought that is not acted upon. (Williams 1973: 175)

35Dancy then observes, “the defeated non-comparative ought can remain true […]. Ross tried to capture this in his theory of prima facie duties, as Williams acknowledges, but he failed to retain the defeated comparative ought in its full vigour. The general idea is that the defeated ought has made its contribution by diminishing the overall rightness of the action that we in fact choose. That this is not sufficient as an account can be seen in Philippa Foot’s excellent example of picking up a snake” (1993: 111).

36As I have emphasized, missing from Dancy’s vision is the pervasiveness of adjusted practice, which populates the landscape of reasons supporting Dancy’s own perception of “salience.” Dancy observes that “some of the properties of a situation are relevant to the question of what one should do, and some are not […]. These relevant properties are salient; they stick out or obtrude, and should catch our attention if we are alert” (ibid.: 112). When moral decisions are explained, according to Dancy, salience is the critical measure: “When we come to give a description of the situation, the various saliences (i.e. the shape of the situation) make a difference to how we should go about it” (Dancy 1973: 112).

The father who tells his child not to take the flowers from the next door garden because that would be stealing should not be seen as subsuming this action under the general principle “Stealing is wrong” (or perhaps “Do not steal”), but rather as pointing to the most salient feature of the situation (that the flowers belong to somebody else), which in this case gives sufficient reason for the child not to do it. (Dancy 1993: 113)

37Salience, I conclude, consists inter alia in the acceptance of already existing patterns from prior normative induction. In coming upon a roundabout, dominating the holism of reasons would be the rule and practice of yielding to traffic already there. Particularism accepts the patterns of normative induction without recognizing that they have themselves emerged from a stage of indeterminate particularism. Surely, the landscape of normatively derived patterns is pluralist, and ill-suited to universal description, supporting Margaret Little’s observation regarding the difficulty of “cashing out the good-making relationship.” But that does not imply a dominant and essentialist particularism.

38Holistic particularism rests on a classification error, that of considering the unique individual dilemma as typical of all moral experience. Particularism has challenged the very notion of identicality in real normative life, what Dancy terms “the rich multiplicity of lived situations” (ibid.: x). In conventional explanations of the inductive process, similarity among objects of inquiry is commonly presumed. Hume’s remark, “when we have found a resemblance among several objects, we apply the same name to all of them, whatever differences we may observe in the degrees of their quantity and quality, and whatever other differences may appear among them,” implies that an important part of the inductive process is the finding of similarity. The question to be addressed is whether, and how, similarity among normatively defined situations comes to be recognized and named.

39The trolley problem is illustrative but hardly representative in demonstrating the valence of intention. As mentioned above it originated with Foot in 1967 (as an improvement on two older chestnuts, the explorers trapped by a fat man stuck in the cave entrance, and the judge faced with the mob threatening mass violence if the innocent man is not executed) in order to critique “the doctrine of the double effect” (Foot 1978: 21-3). “The doctrine of double effect offers a way out of the difficulty, insisting that it is one thing to steer towards someone foreseeing that you will kill him and another to aim at his death as part of your plan” (ibid.: 23).

40Medically assisted death illustrates a further aspect of normative induction hidden from reasons holism, and hence missing from the generalism/particularism debate: the transformation, by repeated experience, of the reasons environment – of the character of available reasons. In 1999 Dr. Kevorkian’s tragic decision took place when the risk of imprisonment was great, because criminal prosecution was the only option in a legal environment that recognized few of the medical, legal, and social considerations that accompany current debate. Philippa Foot’s 1977 paper, “Euthanasia,” treated the issue as an indeterminate dilemma, but anticipating the eventual questions – regarding living wills, better medical procedures, and patient protections – that would have to be, and gradually have been, addressed (Foot 1978: 40, 48-51). These have influenced the course of practice and legislation in the forty years since Foot’s essay.

41General rules are products of experience and necessity, and of a common situation that is explored in repeated judgements. In focusing only upon the defeated ought, particularists ignore that new reasons are forthcoming as old ones are retired. The distinction between the retirement and continued truth of defeated reasons indicates a critical difference in perspective with particularism. Retirement implicates successive points of experience within a community. Dancy’s perspective is synchronic and individual, whereas normative induction is diachronic and social, treating moral decision as a process of inquiry, with distinct stages of reasoning.

3. Can the Trolley Problem be Naturalized?

42This background may elucidate where the trolley problem, like other hypothetical dilemmas, falls short in solving the problem of reversible intentional polarity raised by Philippa Foot. The predicament of the trolley, in its various guises, lies in understanding the role of intention in moral judgment. A redirection of the trolley, whether by the driver or a bystander, implicates a conflict between the opposing principles of blame for direct, but not for indirect, intention. The question raised by Foot and pursued by contemporary commentators like Frances Kamm and Judith Jarvis Thomson is whether a single overall general principle can be found that will serve to explain all of the many and varied instances of reversible polarity that can be imagined.

43As noted above, the trolley problem came from Foot’s 1967 article “Abortion and the Doctrine of Double Effect.” Her purpose there was to examine the opposing valences of direct or indirect action causing harm to an innocent party. At first, the doctrine was applied to cases of fetus abortion where the life of the mother was at stake. That situation was and remains a real dilemma, and has indeed operated within a continuum of inquiry, one that has led to a series of judgments affecting the adoption and revision of medical and legal practices. It will remain controversial while the inquiry continuum is yet ongoing, even as particular judgments continue to highlight and evaluate salient factors, such as the stage of development of the fetus, the nature and degree of danger to the mother, and available medical technology.

44We may recall the changing intentional perspective of our imaginary roundabout driver, Ms Quickly. In the original roundabout situation, it is harder to argue that she intended to cause a collision between her car and the lorry. Her direct intentions lay elsewhere, in hastening to the church service for which she was late. The roundabout was then simply a circular avenue through which she must pass. But the same act in a contemporary roundabout gives rise to a different perception of her intention. The traffic circle is now a controlled thoroughfare, where rules have emerged to prevent repeated instances of injury.

45A similar transformation is found in the field of bioethics and medical practices. While fetal abortion per se may remain controversial indefinitely, nevertheless in the context of choice between the death of a fetus and of the mother, the line between direct and indirect action is not a static one, and what was at first considered blameless for indirection may acquire the perception of blame for failure to use available precautions. Nothing could be clearer than the radical transformation in the public perception of Dr. Jack Kevorkian’s intention following his imprisonment over 20 years ago.

46Foot, rather than recognizing the progression in abortion cases of case-specific inquiry into intent, sought instead to identify an overall principle consistent with any hypothetical example of double effect, including her famous tram, concluding that “sometimes it makes a difference to the permissibility of an action involving harm to others that this harm, although foreseen, is not part of the agent’s direct intention” (my italics). She thus questioned whether there is any uniquely consistent treatment of direct and indirect intent covering all possible real or imagined situations. Her 1967 article extended to cases in which no continuum is conceivable:

Suppose that a judge or magistrate is faced with rioters demanding that a culprit be found for a certain crime and threatening otherwise to take their own bloody revenge on a particular section of the community. The real culprit being unknown, the judge sees himself as able to prevent the bloodshed only by framing some innocent person and having him executed. Beside this example is placed another in which the pilot whose aeroplane is about to crash is deciding whether to steer from a more to a less inhabited area. To make the parallel as close as possible it may rather be supposed that he is the driver of a runaway tram which he can only steer from one narrow track on to another; five men are working on one track and one man on the other; anyone on the track he enters is bound to be killed. In the case of the riots the mob has five hostages, so that in both the exchange is supposed to be one man’s life for the lives of five. The question is why we should say, without hesitation, that the driver should steer for the less occupied track, while most of us would be appalled at the idea that the innocent man could be framed. (Foot 1978 [1967]: 23)

47A clear distinction in principle is found to be less clear in discrete hypothetical cases. Foot first defends the doctrine of double effect, as helpful in some cases, but then recognizes it as unhelpful in others, leading her to suggest another approach toward the end of her article:

• 3 Compare Kamm’s “principle of permissible harm,” n.1 above.

At one time I thought that these arguments in favor of the doctrine of the double effect were conclusive, but I now believe that the conflict should be solved in another way. The clue that we should follow is that the strength of the doctrine seems to lie in the distinction it makes between what we do (equated with direct intention) and what we allow (thought of as obliquely intended). (Ibid.: 25)3

48Foot reaches a surprising conclusion here, that the very conflict she intended to explore through varied examples may not be the same as that with which she began. What indeed is the essential difference between “doing” and “allowing”? Where widely varied adaptive practices are involved, as in medical technology, the line between doing and allowing is plainly subject to movement.

49The inquiry posed by the trolley problem, as well as other extreme hypotheticals, entails avoidance of the real, or natural, relation of particular instances to general rules. Shifting the question to carefully constructed hypotheticals only serves to strain any practical attempt to generalize, thereby reinforcing the appearance of a global and essentialist ethical particularism. The history of actual trolley accidents has already changed prevailing practices affecting track-bound vehicles, reducing the danger to pedestrians or workmen. Early legal decisions placed significant responsibility upon workers and pedestrians to get, or stay, out of the way of moving trams, while later practice came to demand institutional precautions and safety devices. Precautionary design and practice has now all but prevented the infamous hypothetical trolley situation from occurring.

50Meanwhile, awareness of accepted practice will affect “moral intuition,” or the public perception of particular moral dilemmas. When Foot asks, “The question is why we should say, without hesitation, that the driver should steer for the less occupied track, while most of us would be appalled at the idea that the innocent man could be framed,” she is drawing on “moral intuitions” that have their own origin in prior normative induction. She is making the same assumption as Jonathan Dancy makes in his defense of particularism, when he defers to the element he calls “salience.” Salience is Dancy’s own intuitive ranking of relevant reasons, and consists in his acceptance of already existing patterns of cautionary practice from prior normative induction. Foot’s respect for the justice system, in her response to the exemplary punishment of an innocent person, is a similarly entrenched inductive response supported by vast experience in a society of laws, as is her sense of the pilot’s or driver’s duty to minimize damage and injury in the event of sudden tragic malfunction.

51The trolley puzzle seeks to disturb the realm of common moral intuition through unrealistic posits, such as a surrealistically fat man on a bridge or stuck in the mouth of a cave. If such posits achieve anything, it is to provide a dramatic illustration of Margaret Little’s observation that “there is no way of cashing out propositionally the ways in which non-evaluative properties contribute to the evaluative natures of situations, actions, characters.” This may convince the particularist of the global essentialism of particularism, but for the pragmatist, her statement requires a crucial revision: there is no way of cashing out all of the ways that can be imagined in which non-evaluative properties contribute to the evaluative natures of situations; such a cartoon-like realm of reasoning must remain inaccessible to practical analysis. Little’s particularist claim “that the good-making relation cannot be cashed out in propositional form” must thereby be pragmatically amended. Particularist indeterminacy can be, and is, resolved by adjustment of practices, in which the good-making relationship can be, and is, cashed out propositionally, albeit in piecemeal fashion.

4. Conclusion

52Hooker’s article on American moral theory, after examining recent contributions, concludes with this summary:

Connected closely to ideas about responsibility and blame are questions about whether one is more responsible, and to blame, for bad consequences that were part of one’s intention (i.e. part of one’s aim or one’s means to achieve that aim) than one is for bad consequences that were not. A similar question arises about whether one is more responsible for bad consequences that result from what one does than one is for bad consequences that one could have prevented but didn’t. (Hooker 2008: 593)

53Answers to such abstract questions will not be found by a priori armchair inquiry into the concepts of intention and responsibility, drawing on established intuitions as they might apply to extreme imaginary dilemmas. The future of moral theory lies in gaining a deeper understanding of the actual continuum of normative induction through which intuitions are constantly in transformation. The field demands a full recognition of naturalist experimentalism through understanding the social dimension of the emergence of generals through practice. As Dewey noted in 1938,

The notion that a moral judgment merely apprehends and enunciates some predetermined end-in-itself is in fact, but a way of denying the need for and existence of genuine moral judgments […]. The position which holds that moral judgment is concerned with an objective unsettled situation and that ends-in-view are framed in and by judgment as methods of resolving operations is consistent with the fact that, because of recurrence of similar situations, generic ends-in-view, as ways of acting, are built up and have a certain prima facie claim to recognition in new situations. But these standardized “prepared” propositions are not final; though highly valuable means, they are still means for examining the existing situation and appraising what mode of action it demands. The question of their applicability in the new situation, their relevancy and weight with respect to it, may and often does lead to their being re-appraised and re-framed. (Dewey 1938: 168)

54Bernard Williams wrote in 1972, “If there were such an activity as deducing substantial moral conclusions from a priori premises, trained philosophers might reasonably be expected to be rather good at it; but there is not” (1972: x-xi). This comment has lately been read by particularists in support of their extreme position. To the pragmatist, it signals the need to escape the rivalry of generalism and particularism, now a fundamental dead end in analytical moral philosophy, and to examine “genuine moral judgments” drawing on actual experience.