Acts and Morals

1. Introduction: Acts as Embodiments

A theory about the nature and constitution of human acts should be constrained by a broader understanding of the sphere of human action, which includes expressions of moral agency and their moral assessment. This is a methodological requirement among other such requirements on metaphysical theorizing. It is likewise a requirement on our theorizing about the constitution of material objects that it should be constrained by a broader understanding of the sphere of materiality, attained perhaps through physical theory or the principles of material engineering.

In this paper I propose to examine Fine’s theory of embodiment through the lens of the aforementioned requirement and compare it with a couple of alternatives loosely based on Anscombean ideas.¹ The requirement that what we say about the nature and constitution of acts should be informed by a broader understanding of human action, including our thinking about moral matters, isn’t obvious or uncontroversial. Fine, for example, explicitly regards the concern with what acts are separately from the concern with what it is to act, and focuses on the former.² So perhaps for him, even if moral considerations might bear on how we should answer the question what it is to act, they need not bear on how we should answer the question what acts themselves are. But if this paper is on the right track, such separation of what we say about what acts are from what we say about what it is to act, and by extension of what acts are from matters of moral deliberation and assessment, might not be feasible after all.

Fine’s theory of embodiment brings his general framework of qua objects to bear on human acts. The general framework has been deployed most notably in the articulation and defense of pluralism in the metaphysics of material constitution.³ According to the pluralist position, an artwork can be damaged or expensive or rare without the underlying physical object being damaged or expensive or rare, due to the distinctness of two qua objects: the artwork qua being an artifact, say (in Fine’s notation o/μ, where o is the objectual “basis”, and μ is the property of being an artifact, the “gloss”), and the artwork qua being a molecular structure (o/ν, where the gloss ν is the property of being a structure of molecules). Insofar as o/μ and o/ν are distinct, we can ascribe being damaged to the one without ascribing it to the other, and similarly for the other cases. In the application of this framework to acts, we can account for Smith’s act of shooting Jones being intentional without the act of killing Jones being intentional by utilizing the same idea. Where a is the act’s core or basis, ϕ is being intentional, ψ is being a shooting, and χ is being a killing, we can have ϕ(a/ψ) and ¬ϕ(a/χ), which may obtain simultaneously insofar as a/ψ and a/χ are distinct.

This construal of qua as modifying the subject of predication isn’t standard. A more traditional take on qua is found in Anscombe’s work and can be traced back to Aristotle. Anscombe (1979) lists eight “misunderstandings” of her use of ‘under a description’ in the theory of action – which she explicitly identifies with qua – the first of which reads:

Anscombe endorses the Aristotelian view according to which qua modifies the predicate rather than the subject of predication.⁴ In response to a second “misunderstanding” of her use of ‘under a description’ (attributed to Kenny) she goes on to imply that apparently monadic predications are revealed by their susceptibility to qua modification to be relational:

Anscombe may have a monist account in mind according to which an apparent predicate or property applying to something is in fact revealed as a relation of the thing to a further predicate or property.⁵ In the language of properties we can say that the property of being intentional, for example, turns out to be a relational property Φ with act a as one relatum and property ψ as another relatum: Φ(a,ψ). (The upper-case ‘Φ’ is meant to connote the idea that the relational property is higher-order as per the second relatum.) This may seem like a merely notational variant of Fine’s ϕ(a,ψ), but in fact the contrast between the accounts runs deep. To begin to appreciate what is at stake here, let us go back to the larger context of the framework of qua objects and focus on the example of the artwork being damaged while the underlying physical object isn’t.

Letting δ be being damaged, μ be being an artifact, and ν be being an underlying molecular structure, we have Fine’s δ(o/μ) and ¬δ(o/ν) and the monist’s Δ(o,μ) and ¬Δ(o,ν). According to the former pluralist view we have two objects with qua modification in subject position; according to the latter monist view we have one object with qua modification in predicate position. Which account is to be preferred? There are certain “surfacey” advantages to the latter account that shouldn’t carry much weight. True, the monist version is more parsimonious when it comes to how many things are involved in the target predicational disparity damaged/undamaged while being on par with the pluralist version when it comes to how many properties are involved, despite differences in arity and order as per the second relatum. But such considerations of parsimony aren’t decisive when it comes to adjudicating among rival theories and are often outweighed by considerations of systematicity and overall explanatory power.

How about complexity in theoretical paraphrase? Consider the claim

The monist paraphrase is the tripley existential

The embodiment-theoretic paraphrase, on the other hand, is the somewhat less direct, quintupley existential

But this ideological contrast in directness, with the added detour via the identity clauses, is surely not a decisive disadvantage relative to the monist account, and for a parallel reason to the fact that ontological parsimony isn’t a decisive factor. We easily incur added costs of ideological complexity for the sake of clear benefits in theoretical virtues such as global simplicity, elegance, and scope.

I am going to argue that where the monist utilization of the Aristotelian version of qua modification really comes into its own is in its applicability to human affairs in the practical sphere. Here the monist account can in fact be shown to be genuinely preferable to the embodiment-theoretic approach. The plan for the rest of the paper is as follows. Having introduced Fine’s pluralist theory of embodiment, and provisionally comparing it with a monist alternative, I consider in Section 2 how well each of these alternatives handles a familiar scenario of moral deliberation due to Thomson, a certain variant of the Trolley Problem. I argue that the pluralist theory saddles the agent with more moral infractions than its monist rival. Where the monist recognizes one impermissible thing done, the pluralist recognizes multiple coincident impermissible acts. If the agent’s liability is to be as of the monist account, then this discrepancy between the accounts becomes an objection to pluralism, a matter deferred to Section 4. In Section 3, however, I consider a response to the discrepancy on behalf of the pluralist according to which the discrepancy arises from a double-counting fallacy. According to this response, the badness of two coincident acts should no more compound than the weight of a statue and the weight of the lump of matter with which it coincides. In Section 4, I sharpen my critique of the pluralist theory by arguing that the defensive move of treating the badness of coincident acts as equaling the badness of each – assuming each is as bad as the other – runs afoul of the fact that moral assessment of multiple acts is a complex affair that doesn’t admit of a one-size-fits-all solution. Furthermore, the agent’s moral liability from the participant standpoint – the point of view of practical deliberation – should be as of the monist account. Having thus concluded that the monist account is preferable to its pluralist rival, Section 5 outlines a monist alternative preferable to both. Section 6 concludes.

2. The Moral Ledger Problem

We begin to see a significant contrast between the accounts once we attend to certain practicalities. When moral assessments are at stake, it matters which account of qua modification is chosen, or so I will argue. The problem I will raise for the theory of embodiment requires some setting up. Its initial formulation will prove ultimately unsatisfactory, but this will pave the way to a more satisfying formulation of the problem in Section 4. To fix ideas, let us consider the “Fat Man” version of Thomson’s (1985) Trolley Problem. Thomson’s original problem is how to account for the difference in permissibility between two cases: the case of a bystander throwing a switch to divert a renegade trolley to a sidetrack, thereby killing the one on the sidetrack to save the five on the main track; and the case of a surgeon harvesting and then transplanting vital organs from an unwitting “donor”, thereby killing the one to save the five who would otherwise die of organ failure. Both cases are of killing one vs. letting five die and both offer a net gain of four lives. But the act in the first case seems permissible while the act in the second case seems clearly impermissible. Why? Thomson offers a twofold conjecture:

1. in the first case, but not the second, there’s a preexisting threat to the five that gets minimized and redirected to the one;⁸

2. in the first case, but not the second, the means deployed to save the five do not themselves constitute an infringement of a stringent right of the one.

To illustrate the need for the second half of the conjecture, Thomson considers a variant (“Fat Man”) of the original trolley case where instead of throwing a switch and diverting the trolley onto the sidetrack, thereby killing the one to save the five, the agent pushes the one off a footbridge over the main track to stop the trolley, thereby killing the one to save the five. This variant seems clearly impermissible. The means deployed to save the five in this case, but not in the original case of throwing a switch, clearly infringe a stringent right of the one. Hence the second half of the conjecture.

But this talk of means opens up important action-theoretic questions: What are those means, exactly? What did the agent actually do? How are we to assess what was done in terms of whether or not it in fact infringes some stringent right of the one? Suppose the one on the footbridge were merely startled, or suppose the handrail the one was leaning on were merely wobbled, resulting in a fall off the bridge. Could wobbling a rail constitute an infringement of some stringent right of the one? Thomson’s response is: the act of wobbling isn’t just a wobbling. It is a wobbling-to-stop-the-trolley-with-the-body-of-the-one. Or perhaps a wobbling-to-topple-the-one-off-the-bridge-to-shut-down-the-track. And so on.⁹ No such act is permissible. All constitute an infringement of a stringent right of the one.

Let us now consider the situation from the embodiment-theoretic perspective and compare it with the situation from the monist perspective. Let b be the wobbling, ρ be toppling the one off the bridge to cause the track to shut down, and τ be toppling the one off the bridge to physically stop the trolley with the body of the one.¹⁰ From the embodiment-theoretic perspective we may consider b/ρ and b/τ, each impermissible due to infringing a stringent right of the one. From the monist perspective, on the other hand, we have act b, one and the same, impermissible qua ρ and impermissible qua τ – impermissible twice over.

Consider an agent who wobbles the rail to topple the one off the bridge to shut down the track underneath and wobbles the rail to cause the trolley to stop with the body of the one toppled off the bridge.¹¹ Insofar as we are assessing acts for their permissibility, and insofar as the embodiment theorist will recognize two acts undertaken concurrently, the two accounts are not morally speaking on par. The former account has the agent act impermissibly twice whereas the latter account has the agent act impermissibly only once. True, the act of wobbling according to the monist account is impermissible twice over. But as far as the moral ledger of the agent goes, insofar as we’re assessing what the agent did and assess the agent as a point of origin for such doings, the agent who commits both b/ρ and b/τ would seem to be worse off morally speaking than that agent according to the monist account. I call the discrepancy between the moral implications of the two accounts the moral ledger problem. The number of acts committed and assessed for rightness or wrongness matters morally speaking. And what goes for morality goes for law. Suppose acts infringing stringent rights of individuals incur punishments or fines at fixed rates. Then all else being equal, the punishments or fines incurred by the agent who commits both b/ρ and b/τ would seem to double those incurred by that same agent according to the monist account.

3. The Participant Standpoint

The pluralist framework of the theory of embodiment appears to increase our liability as compared with the monist framework. Insofar as morality is concerned with the rightness and wrongness of what is done and with our moral assessment as doers, the pluralist and monist accounts appear not to be on par morally speaking. We needn’t say at this point which account is to be preferred, the one according to which our agent acted impermissibly twice or the one according to which the agent acted impermissibly only once but what was done was impermissible in two ways. The moral ledger problem simply takes note of the moral discrepancy between the two accounts.

If the agent’s moral ledger is to be as of the monist rendering, as I will argue in the next section, the moral ledger problem constitutes an objection to pluralism. From the monist perspective, one act is being treated as two. And yet there is an obvious similarity between the moral ledger problem as stated above and a familiar objection to the pluralist view that a material thing and the lump of matter with which it coincides are numerically distinct things. The objection is general and is vividly illustrated with the special case of weight. If the statue has a certain weight X, then presumably the lump of bronze with which it coincides has the same weight X, in which case upon being placed on the scale the reading should be 2×X, which it isn’t. Other such examples of double-counting may be multiplied as needed. Here is how Fine (2006) responds to the charge:

Similarly in our case, we can imagine the embodiment-theoretic response to the moral ledger problem to be that the moral score for distinct acts adds up only for non-overlapping distinct acts, acts that do not share a basis. b/τ and b/ρ do share a basis and therefore should not count twice for the purpose of keeping moral score.

The problem with this response to the moral ledger problem is that it treats the issue as one that can be resolved metaphysically, independently of moral considerations. But it can’t, as we are about to see. The problem is a problem that arises from the perspective of our first-order moral thinking about acts as participants in moral practices, the agentive standpoint of moral deliberation. One can certainly draw an analogy to the case of weighing coincident material objects or counting members of overlapping sets, but such analogies can only go so far. At the end of the day, in the practical realm we have strong pre-theoretical notions that are far more recalcitrant and far less negotiable via metaphysical theorizing. If this is right, the monist approach is at a clear advantage over the embodiment-theoretic approach.

Let us now turn to examine some of the relevant details. For requisite background, consider a duality of standpoints emphasized by Nagel for many seemingly disparate areas of philosophy. The standpoints are sometimes labeled subjective and objective, or first-person and third-person, or sometimes even sympathetic and perceptual.¹² But because the terminologies of subjective/objective and first-person/third-person do not do justice to second-person aspects of the relevant phenomena, and because sympathetic/perceptual in the intended sense are too closely tied to matters of imaginability, I am going to label the duality of standpoints participant and spectator. In Nagel (1976) the contrast is utilized for a discussion of what has come to be known as moral luck, the details of which lie outside our present concerns. What is relevant here, however, is that the contrast is touted as crucial to how we approach the purview of moral assessment quite generally.

From a spectator standpoint, we may think, ‘This is a good (bad) thing to have happened’ of some act, or ‘It is good (bad) that this person exists’ of some agent. From a participant standpoint, on the other hand, when evaluating acts or agents as good or bad we think, ‘This is a good (bad) thing to do’ or ‘This is a good (bad) person to be’. The first standpoint is that of a surveyor of the moral scene from above; the second is that of a would-be participant in a practice aspects of which are being assessed. Importantly, morality from a participant standpoint is addressed to a deliberating agent choosing what to do next. Some vexing problems in moral philosophy emerge from the difficulty of reconciling the two standpoints.

Consider, for example, Williams’ (1973) famous discussion of the shortcomings of utilitarianism. Someone on a botanical expedition, “Jim”, stumbles upon a public square in a remote town where twenty natives, arbitrarily chosen, are lined up against the wall and about to be shot by a soldier so as to be made an example and sow fear in the local population. The person in charge, “Captain”, offers Jim the choice of killing any one of the twenty, thereby saving nineteen lives; if Jim refuses the offer, the soldier will kill the twenty. Those lined up against the wall and the families present are begging Jim to accept the offer. What should Jim do? From a spectator standpoint the case seems like a no-brainer: accepting the offer will result in a net gain of nineteen innocent lives (at some psychological cost to Jim, perhaps, but this seems relatively negligible). From a participant standpoint, however, by accepting Captain’s offer Jim will have killed an innocent person, a terrible thing to do. Considering the case in light of Thomson’s twofold conjecture delivers the verdict that Jim may not accept the offer. As per redirecting a preexisting threat from twenty to one, this is clearly not what’s going on here. The killing of the one by Jim is clearly not related in that way to the killing of the twenty by the soldier if Jim refuses the offer.

Moreover, the means deployed to save the nineteen would certainly infringe a stringent right of the one. So accepting Captain’s offer would be impermissible on both counts according to Thomson’s twofold conjecture. This can seem puzzling because one would have thought that accepting the offer should at the very least be permissible. As mentioned above, from a spectator standpoint the case seems like a no-brainer. Here is what Thomson (1985) has to say about such conflicts:

Such passages make it clear that the problem of permissibility in the cases discussed by Thomson and Williams, specifically when it is permissible to sacrifice one to save many, are problems addressed to a deliberating agent from a participant standpoint. As Williams memorably puts the point:

The pertinent question for Jim isn’t whether accepting Captain’s offer to a net gain of nineteen lives is a better thing to happen than its refusal. It is surely a better state of affairs. The pertinent question for Jim, rather, is whether accepting Captain’s offer is a better thing to do than refusing it. What is Jim to do here? It is a question approached from the participant standpoint. To that question, Thomson’s answer is that Jim may not accept the offer, may not kill one to save twenty from being killed by the soldier to a net gain of nineteen lives. But how to reconcile the two standpoints remains a longstanding problem.

4. The Moral Ledger from the Participant Standpoint

Let us go back to the moral ledger problem. We now have the requisite backdrop to situate and sharpen our critique of the embodiment-theoretic approach. Where b is the wobbling of the rail, ρ is toppling of the one off the bridge to shut down the track, and τ is causing the trolley to stop with the body of the one toppled, we have b/ρ and b/τ, each a bad thing to happen (let us assume). Are things worse if both occurred than if only one did? Considering the acts involved from a spectator standpoint, we can extend to them the sort of treatment Fine provides for other cases of double-counting, such as the weight of the statue and its coincident lump and the number of elements in two partially overlapping sets. From such a perspective, insofar as b/ρ and b/τ share basis b, their badness does not compound. Perhaps we can say that their shared badness equals that of the worst of b/ρ and b/τ if one of them is worse than the other.

But b/ρ and b/τ are not just bad things to happen. Assuming Thomson’s conjecture, each of b/ρ and b/τ is impermissible due to infringing a stringent right of the one. Beyond being bad things to happen as viewed from a spectator standpoint, b/ρ and b/τ are bad things to do when viewed from a participant standpoint. And here it becomes harder to avoid the moral ledger problem. Consider moral education. We might teach our children not to wobble rails to topple those who lean on them off bridges to shut down the tracks running underneath, thus preventing trolleys from proceeding. Or we might teach our children not to wobble rails to stop trolleys with the bodies of those who lean on them. (The specificity here is obviously laughably exaggerated, a point to which I return shortly.) Such compound teachings take longer to impart than each of them taken on its own. Such compound teachings require more of those who are to follow them than what each of them requires on its own. In concert, they may require of the agent to deliberate in ways that aren’t reducible to the agent’s deliberation with respect to each teaching taken on its own. Indeed, how violations of such teachings are supposed to compound isn’t a straightforward matter from a participant standpoint.

To appreciate the complexity here, consider some hypothetical cases. Jay and June each wobbles a rail on a footbridge over a track with a renegade trolley headed towards five, toppling the one off the bridge. Jay violates the first teaching. The one leaning on the rail in Jay’s case is toppled off the bridge by Jay’s wobbling to shut down the track in advance, thus preventing the trolley from hitting the five. June violates the second teaching. The one leaning on the rail in June’s case is toppled onto the path of the trolley as in Thomson’s original “Fat Man” case to stop the trolley with the body of the one, thus preventing the trolley from hitting the five. What each of them does is bad. Toppling someone off a bridge is a bad thing to do in both cases.

Now consider Jane, who wobbles the rail to topple the one off the bridge and shut down the track in advance in case the trolley is early, and wobbles the rail to topple the one off the bridge and cause the trolley to stop with the body of the one in case the trolley is on time.¹³ Is what Jane does a worse thing to do than what either Jay or June does? Under the embodiment-theoretic approach, if Jane is really doing something morally equivalent to what each of the others does except concurrently, then assuming that what Jay does is just as bad as what June does, it would initially seem that what Jane does is twice as bad as what either of the others does. In violating two teachings, Jane is behaving in a morally unacceptable manner twice, albeit concurrently, whereas each of the others is behaving that way only once. And now the embodiment theorist would like to come back and say that nothing of significance turns on switching from the spectator standpoint to the participant standpoint. As long as we’re concerned with acts sharing a basis, evaluating such acts, even from the participant standpoint, isn’t the sum of the evaluations of each of the acts considered separately. What Jane does isn’t a worse thing to do than what either Jay or June does after all. But how exactly this is supposed to work in terms of morality telling the agent what to do or refrain from doing in practical deliberation is far from clear.

Morality from the participant standpoint is addressed to a deliberating agent who faces decisions about what to do next. There is a practical limit on how specific moral requirements can get. We just imagined two moral “teachings”: one not to wobble rails to topple those who lean on them off bridges to shut down the tracks underneath; the other not to wobble rails to stop trolleys with the bodies of those who lean on them. Both are obviously implausibly specific. And yet such implausibility is compounded significantly if we are told not to wobble rails to topple those who lean on them off bridges only to shut down the tracks underneath, and not to wobble rails to topple those who lean on them off bridges only to stop trolleys with the bodies of those who lean on them, and not to wobble rails to topple those who lean on them off bridges only to both shut down the tracks underneath and stop trolleys with the bodies of those leaning on them. And yet such implausible specificity would be the effect of trying to avoid the moral ledger problem by making moral requirements from a participant standpoint specifically tailored to the respective situations of our three actors so that each of them can be construed as violating only a single moral requirement.

Perhaps the situation can’t quite be described as morality telling us not to wobble a rail to topple someone leaning on it in a way that is sensitive to who the actor is. But according to the present embodiment-theoretic strategy of avoiding the moral ledger problem, morality would be telling us not to wobble a rail to topple someone leaning on it off the bridge if we are specifically Jay-like, and not to wobble it to topple someone leaning on it off the bridge if we are specifically June-like, and not to wobble it to topple someone leaning on it off the bridge if we are specifically Jane-like. Such directive specificity is unmotivated at best. On the other hand, construing the moral requirements here as more generic, say a general prohibition against wobbling the rail tout court, or a general prohibition against wobbling the rail to topple whoever leans on it off the bridge, not only saddles Jane with two violated requirements in her conduct where the others are saddled with only one violation each, but ignores eventualities in which wobbling the rail, even wobbling the rail to topple someone off the bridge, might be permissible.¹⁴

Beyond the unmotivated directive specificity, the embodiment-theoretic strategy of avoiding the moral ledger problem is rather simplistic. To see this, consider why Jane might act as she does. She wobbles the rail to topple the one off the bridge and shut down the track in advance in case the trolley is early while wobbling the rail to topple the one off the bridge to cause the trolley to stop with the body of the one in case the trolley is on time. Let us suppose that Jane is being more vigilant than the other two in securing the desired outcome of saving the five to a net gain of four lives.¹⁵ With such emphasis, it seems compelling that what Jane does is not only not a worse thing to do than what either of the others does, but is in fact a better thing to do than what either of the others does. Jay and June are each securing only the death of the one by wobbling the rail but without the extra vigilance in securing the outcome of saving the five. In this respect, what each of them does is worse than what Jane does. Jane is at least being extra vigilant in saving the five, which is essential to securing the shared desired outcome of a net gain of four lives. We are thus disinclined to extend to Jane’s case the simplistic embodiment-theoretic treatment according to which she is morally indistinguishable from the other two (under the assumption that they are morally indistinguishable from each other).

From a participant standpoint, the diagnosis of double-counting seems like the sort of theoretical consideration that fails to meet the practicalities of the situation at the right level. A judge wants to know how many violations were committed and the number of acts clearly matters here. To be told that there were multiple violations committed despite the apparent singularity, sharing a basis without sharing a gloss, is a bit of theoretical overspeak that can easily mislead parties to moral or legal disputes into thinking that mundane cases of doing things for multiple purposes or serving multiple outcomes are rather more rarified than they really are. There is room enough in law and morals for the kind of multiplicity in concurrent acts that isn’t of the embodiment-theoretic variety: the single bullet killing multiple victims, the reckless driving killing multiple members of a single family, and so on.¹⁶ There is no need or wish to recast acts to fit the embodiment-theoretic mold in our everyday dealings with acts from a participant standpoint. What we ordinarily regard as individual acts undertaken perhaps with multiple purposes in mind or serving multiple outcomes should be distinguished from those special cases that we have specific legal or moral reasons to regard as involving multiple concurrent acts.

Turning now to the parallel situation from the monist perspective, things look rather different. Jay wobbles rail, one falls and dies, trolley prevented from hitting five. June wobbles rail, one falls and dies, trolley prevented from hitting five. Jane wobbles rail, one falls and dies, trolley prevented from hitting five. Letting I stand for impermissibility, a relational property, and a^X stand for the act committed by X, we have three acts and four infractions: I(a^JAY, ρ), I(a^JUNE, τ), I(a^JANE, ρ), and I(a^JANE, τ). Assuming Thomson’s conjecture as background, what Jane does is a bad thing to do in different respects. Perhaps we could say that in being more vigilant about securing a net gain of four lives, what Jane does isn’t quite as bad in the final analysis as what the others do. But we need not commit in advance to any particular way of amalgamating the overall badness of what Jane does as compared with what each of the others does, an issue that arises specifically for the pluralist in light of the moral ledger problem. An ontology of coincident acts requires a principled solution to this problem, but there is little reason to suppose that such a solution is available in advance of actual cases.

Jane would not be held liable for two committed violations. Here is some evidence from California state law:

The discount in punishment articulated in the California Penal Code is straightforwardly compatible with a monist approach. We have an acknowledgement of the idea that a single act may be punishable under multiple provisions of law. Assuming the legal situation reflects the moral situation here, this reflects the idea that a single act may violate multiple moral requirements, much in the way that what Jane does is a bad thing to do in multiple respects. While the compatibility of the relevant portion of the law with a monist approach is straightforward, its compatibility with a pluralist approach is more complicated. Distinct legal provisions r and r′ give rise to distinct glosses, “punishable under r” and “punishable under r′”, which give rise to distinct acts by pluralist lights. It is then incumbent on the pluralist to offer some principled way of amalgamating those acts into individual act-bundles for such legal purposes as applying the discount in punishment articulated in the California Penal Code. On the other hand, care must be taken that such pluralist bundling won’t over-generate. There might still be compelling legal and moral reasons for regarding a case of reckless driving killing multiple members of a single family as multiple concurrent killings punishable separately. A principled bundling of acts that share a basis would pre-judge these cases in ways that might be legally or morally indefensible.¹⁷

This isn’t the end of the story, of course. The law isn’t infallible. But the law does often provide a fairly straightforward indication of where our everyday moral thinking tends to go when conducted from a participant standpoint. It seems fairly clear that the number of acts committed carries significant weight in our everyday moral deliberations. And when it comes to the matter of sheer numbers, it seems fairly clear that the monist approach enjoys greater compatibility with our thinking about morality from a participant standpoint than the embodiment-theoretic approach. Greater compatibility with our thinking about morality isn’t the end of the story either. Perhaps theoretical benefits elsewhere can tip the balance even when it comes to such practical matters as assessments of right or wrong from a participant standpoint. But given the availability of a monist alternative to the embodiment-theoretic account, the motivation for developing an effective embodiment-theoretic answer to the moral ledger problem is significantly reduced.

5. A Simpler Monism

I would now like to sketch a monist alternative that seems even more promising than the version discussed so far. This improved version of monism is simpler and regards our pre-theoretical, everyday judgments of predicational disparity at face value while staying true to the appearance of simple monadic predication – pace the relational analysis of the original monist account – without multiplying entities in subject position to account for the disparity – pace pluralism. Going back to the example of the artwork and the property of being damaged, we might say of the artwork that it is either damaged or not, full stop. The painting is damaged as befitting paintings because its canvas is torn, let us suppose. But it isn’t damaged as befitting molecular lattices, a matter of not having undergone radiation damage, say. The simpler monist will insist that there’s only one thing there, the painting as the singular subject of those predications, and that the ascribed property isn’t relational but monadic. How can such claims be reconciled? The present strategy is to appeal to a divergence in a contextually specified factual background. More specifically, the idea is to let the qua property remain in the factual background without structuring the relevant facts in the foreground. That the item in question is an artifact or a molecular structure isn’t itself part of the fact that it is or isn’t damaged. It is rather “metaphysically presupposed” in a sense to be explained below.

To fix ideas on what this might look like, we turn briefly to the original site of the notion of presupposition in Strawson’s (1950) famous critique of Russell’s (1905) theory of descriptions. Russell’s theory, recall, analyzes an apparently subject-predicate sentence of the form ‘The F is G’ as an existential quantification of the form ‘Something uniquely Fs and is G’. The theory is clearly revisionist regarding the pre-theoretical subject matter, specifically regarding the pre-theoretical subject-predicate structure of the descriptive sentence. Immediate implications of the theory are that (i) the apparent subject-predicate form of the sentence of the form ‘The F is G’ is merely apparent – logically speaking, the sentence is existential and not about anything in particular, and (ii) ‘Something uniquely Fs’ is logically entailed by ‘The F is G ’. Strawson’s denials of (i) and (ii) are well known and have had a momentous impact on the subsequent study of natural language. Regarding (i), Strawson claims that ‘The F is G’ is of subject-predicate form, just as our pre-theoretical judgment would have it, and is specifically about whatever the subject ‘the F’ refers to. Regarding (ii), Strawson claims that ‘Something uniquely Fs’ is not a logical entailment of ‘The F is G’. Rather, it is presupposed. To assert ‘The Queen of England is bald’ is not to say something that logically entails that there is a unique female sitting on England’s throne. Rather, having a truth-value to begin with depends on there being a unique female sitting on England’s throne. The truth or falsity of the original then turns on whether or not such a person in particular, the person referred to, is bald.

I now propose to offer a Strawsonian refinement of the original monist position, reaping the benefits of monism while respecting the apparently monadic character of the relevant predicated properties. The relational monist treatment of predicational disparity includes qua modification in predicate position. For the painting to be damaged qua being an artifact and undamaged qua being a molecular structure is for the relational property of being damaged to relate the painting to being an artifact while not relating it to being a molecular structure. This is but one monist alternative to the pluralist rendition in terms of a monadic property of being damaged applying to the painting qua being an artifact while not applying to the painting qua being a molecular structure, two distinct qua objects sharing a basis. But within the metaphysical dispute between monism and pluralism, we can treat the qua properties as analogous to the descriptive material in the Strawson-Russell debate. The original monist position includes the qua property as part of what is being predicated of the subject. The apparently monadic first-order predication in re of being damaged is revealed as a relation in re to the qua property.

Two aspects of this monist position are useful to view against Russell’s revisionist approach to ‘The F is G’, one pertaining to structure or form of the putative fact and the other pertaining to order or level of predication. The first is that the structure of an apparently monadic simple fact that some object has a property is revealed as only apparently monadic but deeply relational, by analogy to the way an apparently subject-predicate structure of a descriptive sentence is analyzed by Russell as merely apparent and deeply quantificational. The second is that the predicated property in re is revealed as higher-order as per the second relatum, by analogy to the way the apparently first-order predication in ‘The F is G’ is revealed as the higher-order predication of existential quantification.

These analogies allow us to steer a middle course between the Scylla of the pluralist framework of qua objects and the Charybdis of relational monism. The cost of the former is a certain proliferation in subject position and the cost of the latter is a certain added complexity in predicate position. It would be better to have an alternative that avoids both costs. The challenge is to articulate a metaphysical analog of Strawsonian presupposition so that we can say that the qua property isn’t in fact structuring the fact at issue as a relatum but is rather metaphysically presupposed. Being an artifact isn’t a relatum in the painting being damaged but is rather metaphysically presupposed. Being a molecular structure isn’t a relatum in the painting being undamaged but is rather metaphysically presupposed. Talk of presupposition here is clearly metaphorical insofar as there is no agent to do any presupposing in the metaphysical case. (The world itself is not an agent.) What the notion of metaphysical presupposition is meant to point to is a kind of contextualization according to which predication in re is sensitive to a factual background, a range of background facts that bear on foreground factuality.

A vivid illustration of such contextualization is found in Lewis’ (1976) influential treatment of the paradoxes of time travel. We have the case of Tim who builds a time machine and goes back in time to kill his own grandfather. Assuming there’s no question whether Tim does in fact kill Grandfather – he doesn’t – the question Lewis considers concerns ability: Can Tim kill Grandfather? Tim, we assume, has everything it takes to kill Grandfather, the best gun, training, conditions, what have you. So it seems that Tim can kill Grandfather, he can kill him qua being a skilled shooter under optimal conditions. On the other hand, killing Grandfather in the past would undermine the conditions for Tim to exist later, specifically for the later killing to occur. So it seems that Tim cannot kill Grandfather, he cannot qua being a traveler into his own past. So he can and he can’t. Contradiction! Lewis remarks:

When considering whether the killing can happen, we need to take into account a contextually determined background consisting of various facts. Lewis compares this to other cases of ability: Can I speak Finnish? I have what it takes: facts pertaining to my larynx, my nervous system, my natural endowment broadly construed are compossible with my speaking Finnish and determine a positive answer: I can speak Finnish qua being a normally endowed human.¹⁸ But facts pertaining to my background, my upbringing, my personal history broadly construed are not compossible with my speaking Finnish and determine a negative answer: I can’t speak Finnish qua being someone with a personal history that includes no experience in the language. So I can and I can’t. The apparent contradiction is resolved by relativizing the ability to speak Finnish to a contextually specified background. Relative to some range of background facts I can speak Finnish; relative to another range of background facts I can’t speak Finnish. And what goes for my ability to speak Finnish goes for Tim’s ability to kill Grandfather. He can kill Grandfather relative to one range of facts and can’t kill him relative to another range of facts.

We can utilize this idea of contextual background, or relativization to a range of background facts, for the Strawsonian middle course we’ve been steering. The question whether the painting is damaged or not will receive the same sort of contextual determination. Assessing damage as befitting paintings is relativized to a range of background facts that include the item being an artifact; assessing damage as befitting molecular structures is relativized to a range of background facts that include the item having a certain molecular constitution. The painting with the torn canvas is damaged qua being an artifact to the extent that the range of facts in the contextual background includes facts pertaining to the item being an artifact. The painting with the torn canvas isn’t damaged qua being a molecular structure to the extent that the range of facts in the contextual background includes facts pertaining to the item being a molecular structure. Indecision may arise when the contextual background includes facts of both kinds.¹⁹

This account respects apparent singularity in subject position: no plurality of objects adduced to resolve the apparent conflict between being damaged and not being damaged for what seems pre-theoretically to be one and the same item. The account also maintains simplicity in predicate position: no departure from the apparently monadic character of being damaged to something relational (and higher-order as per the second relatum). It is a simpler, contextual version of monism. And it allows us to say more clearly how Smith’s act of shooting Jones is intentional while the self-same act of killing Jones isn’t. Against a background that includes certain facts pertaining to the act being a shooting, such as Smith’s focused determination in targeting Jones and his cold execution, Smith’s act of shooting Jones is intentional. Against a background that includes facts pertaining to the act being a killing, such as causing death due to the victim’s unknown hemophilia, Smith’s self-same act of killing Jones isn’t intentional. No need to multiply acts here. And no need to construe being intentional as a relational property.

6. Conclusion

We’ve been considering the theory of embodiment as a special case of a more general pluralist framework of qua objects. A cursory glance at the theory’s construal of qua modification and the initial comparison with the theory’s initial monist rival failed to reveal what is really at stake in the contrast between the two views. True, according to the pluralist position the qua property modifies the subject while according to the monist position it modifies the predicate. But this contrast doesn’t itself disclose which account is to be preferred. I’ve been arguing that the advantage of monism over pluralism in the theory of acts can be truly appreciated once we compare their respective implications for moral deliberation and assessment.

The initial presentation of the moral ledger problem suggested that the number of acts committed can have moral significance. This could be appreciated even before we considered which account is to be preferred, the pluralist one or the monist one. All we could see at first is that the moral ledger of the agent differs according to which account is adopted. We then examined a defensive move on behalf of the pluralist that effectively treats multiple coincident acts as single acts for the purpose of moral evaluation, thereby attempting to level the moral ledger of the agent across the two accounts. We considered this particular pluralist strategy in more detail in light of the familiar distinction between morality from a spectator perspective (sub specie aeternitatis) and morality from a participant perspective (the deliberative standpoint). This facilitated a fresh take on the pluralist response to the moral ledger problem. We saw that the monist account better accords with our verdicts on moral liability.

Finally, we looked at a simpler and more satisfying version of the monist position that better coheres with our pre-theoretical judgments about predicational disparity, including such cases as the shooting eventuating in a victim’s death being intentional while the killing isn’t. According to this preferred monist alternative, the single act is intentional against a factual background that includes such facts as the focus and skill with which the agent pulled the trigger and unintentional against a factual background that includes such facts as the death of the victim being unforeseen.

We can easily imagine how analogs of the moral ledger problem might arise in other areas of practical concern. An insurance policy on a disconnected sideways installed urinal labeled “Fountain” – an influential work of art by Duchamp – offers coverage for the sort of damage appropriate for artworks. Would there be any need for a separate policy covering damage to the item qua being a urinal? Or coverage for the item qua being a porcelain artifact? We can easily imagine such questions becoming pressing and handled differently by the pluralist and monist frameworks. The number of insured items clearly matters. It would seem, for example, that a pluralist framework of qua objects might offer the insurance company a loophole for advancing the position that a significant visible crack in the porcelain isn’t covered by the existing policy, which was only taken for the item qua artwork and not for the numerically distinct item qua being a porcelain artifact. The motivation to develop a satisfying monist alternative to a general pluralist framework of qua objects seems to extend beyond the special case of human acts.

I began the paper with the methodological claim that a theory pertaining to the nature and constitution of human acts should be constrained by a broader understanding of the sphere of action, including moral action, much in the way that a theory of material constitution should be constrained by a broader understanding of materiality as afforded by the physical sciences. The theories thus constrained are metaphysical theories, and while the analogy between the metaphysics of acts and the metaphysics of material constitution is apt as far as being constrained by a wider understanding of the relevant domain goes, there is another respect in which the cases seem rather unlike. In the latter case, the methodological requirement becomes in effect an inter-theoretical one. The metaphysics of material constitution should remain answerable to what more established theoretical endeavors have to say about the shared subject matter of materiality. In the former case, however, the situation is different. Our metaphysics of acts should be constrained by our broader understanding of human action, yes, and that includes moral action. But such understanding pertains in the first instance to the practical sphere, and here the question of how we should be theorizing about the domain is itself hotly debated. A sense of what the controversy is about can be gleaned from attending to the deliberative standpoint in ethics and its ramifications for ethical theory. I leave a more thorough exploration of the interplay between ethical and metaphysical theorizing for future work.

Competing Interests

The author has no competing interests to declare.