The ‘natural unintelligibility’ of normative powers

ABSTRACT

This paper offers an original argument for a Humean thesis about promising that generalises to the domain of normative powers. The Humean ‘natural unintelligibility’ thesis – prominently endorsed by Rawls, Hart, and Anscombe, and roundly rejected or forgotten by contemporary writers (conventionalists and non – conventionalists alike) – holds that a rational, suitably informed agent cannot so much as make a promise (much less a morally-binding promise) without exploiting conventional norms that confer promissory significance on act types (e.g., signing on the dotted line) that would not otherwise have such significance. The argument (like Hume’s) is action-theoretic in character; its central premise is the Contribution Condition on acting with an aim: roughly, that doing X in order to (bring about) Y requires that the agent believe that their having X-ed on the occasion might come to fully or partly account for Y’s coming to pass. The argument generalises to the broader domain of exercises of normative powers, i.e., to acts that modify normative properties by ‘fiat’ or ‘stipulation’. This stipulative act has been characterised in three different ways, each of which is given due consideration. Finally, two different ways of accommodating the ‘natural unintelligibility’ thesis are considered, as is the thesis’s normative significance.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Hobbes deployed the fiat metaphor in connection with promises: ‘[T]he pacts and covenants…resemble that fiat, or the let us make man, pronounced by God in the creation.’ Thomas Hobbes, Leviathan (first published in 1651, Hackett 1994) 3–4 (italics in the original).

2 Joseph Raz, ‘Voluntary Obligations and Normative Powers II’ (1972) 46 Proceedings of the Aristotelian Society, Supplementary Volumes 59; Joseph Raz, Practical Reason and Norms (Oxford University Press 1999); Joseph Raz, ‘Normative Powers’, in Ulrike Heuer (ed), The Roots of Normativity (Oxford Univerity Press 2022).

3 In introducing the category normative powers, Raz expressly aimed to rebut the claim ‘that norms which are not practices cannot be affected by human acts in a way which allows one to regard these acts as the exercise of normative powers’: see Raz, ‘Voluntary Obligations and Normative Powers II’ (n 2) 93; and Raz, Practical Reason and Norms (n 2) 53. ‘Moral rules are perhaps the clearest example of rules which are not practices’. Raz’s generalization proceeds as follows: while the notion of a legal power is defined in terms of the motivating reasons for which types of acts have been endowed with legal significance (see note 55 below) – and is therefore inapplicable to nonconventional norms that are not brought into existence in pursuit of any aim – the more general notion of a normative power is defined in terms of features of a norm’s explanation. According to Raz, while a legal norm is explained by the aforementioned motivating reasons (that is, the law’s intended design), nonconventional power-conferring norms are explained by the facts that ‘justify’ them. And just as a legal norm’s power-conferring status is owing to features of the law’s intended design, a nonconventional norm’s power-conferring status is owing to corresponding features of the facts that justify them. Raz’s account of the justifying features of power-conferring norms has evolved since his original 1972 paper, which appealed to the content-independent character of the justification of power-conferring norms. In subsequent work Raz modified his view, appealing to a different feature of their justification (see note 24).

4 Joseph Raz, ‘Promises in Morality and Law’ (1982) 95 Harvard Law Review 916, 927; see also Joseph Raz, ‘Promises and Obligations’ in PMS Hacker and Joseph Raz (eds), Law, Morality, and Society: Essays in Honour of H.L.A Hart (Oxford University Press 1977). Raz’s characterisation of promising has been widely influential in the literature on normative powers more broadly, and has been directly adopted by, e.g., David Enoch, ‘Giving Practical Reasons’ (2011) 11 The Philosopher’s Imprint; David Owens, Shaping the Normative Landscape (Oxford University Press 2012); David Enoch, ‘Authority and Reason-Giving’ (2014) 89 Philosophy and Phenomenological Research 296; Gregory Klass, ‘Promise, Agreement, Contract’ in Hanoch Dagan and Benjamin Zipursky (eds), Forthcoming, Research Handbook on Private Law Theories (Edward Elgar Publishing 2020); for illuminating synoptic discussions (with a focus on promising and consenting, respectively), see Ulrike Heuer, ‘Promising Part 2’ (2012) 7 Philosophy Compass 842; and Felix Koch, ‘Consent as a Normative Power’ in Peter Schaber and Andreas Müller (eds), Routledge Handbook of the Ethics of Consent (Routledge 2018).

5 David Hume, A Treatise of Human Nature (first published 1739, Clarendon Press, Oxford 2007).

6 Raz, ‘Voluntary Obligations and Normative Powers II’ (n 2) 94.

7 Mark Greenberg, ‘The Standard Picture and its Discontents’ in Leslie Green and Brian Leiter (eds), Oxford Studies in Philosophy of Law (Oxford University Press 2011) 45; cf. Ezequiel Monti, ‘Against Triggering Accounts of Robust Reason-Giving’ (2021) 178 Philosophical Studies 3731.

8 In due course I will relax the assumption that the description of the act in virtue of which it qualifies as an exercise of a normative power must be directly (normatively) significant (see note 43).

9 Noncausal explanation has, of course, been much investigated in recent years, and Selim Berker (among others) has emphasized the prominent role of the grounding relation in normative explanation. See Selim Berker, ‘The Unity of Grounding’ (2018) 127 Mind 729 and on grounding more generally Fabrice Correia and Benjamin Schnieder (eds), Metaphysical Grounding: Understanding the Structure of Reality (Cambridge University Press 2012).

10 HA Prichard, ‘The Obligation to Keep a Promise’ in Jim MacAdam (ed), Moral Writings (first published 1949, Clarendon Press, Oxford 2002). While Prichard’s discussion is framed in terms of ‘promising’ rather than ‘normative powers’, he notes that the relevant features of promising (for the purposes of his discussion) are common to a broader class of acts, citing gifting and exchanging as other relevantly similar examples of operations on obligations.

11 Of course, I make no claims about Prichard’s examples, and the reader can substitute her own.

12 Prichard sought such a characterisation of the promissory act and despaired that ‘we get into difficulties … as soon as we try to state what [it is] which renders us bound’ when we promise (id. 258). The source of Prichard’s ‘difficulties’, which extend to all exercises of normative powers, is that we typically promise by ‘making certain noises [or marks]’ – that is, by performing actions of (at least some) describable types that are of no immediate normative significance (id., 264). Of course, this latter fact alone does not serve to differentiate exercises of normative powers from other normatively consequential acts. For example, one may cause great harm (and incur various secondary duties) by moving a finger ever so slightly.

13 In note 65, I distinguish between stronger and weaker versions of conventionalism, and clarify that what is at issue in this paper is the weaker thesis (‘count-as conventionalism’).

14 John Rawls, ‘Two Concepts of Rules’ (1955) 64 The Philosophical Review 3.

15 HLA Hart, The Concept of Law (first published 1961, Oxford University Press 2012).

16 I set aside compound descriptions built out of these three, since my argument generalises. For the same reason, I also set aside the variation on the ‘communication’ characterisation that substitutes ‘belief’ for ‘intention’. And I note that some may prefer substituting ‘expressing’ for ‘communicating.’

17 For those uneasy about the notion of a ‘fully informed’ agent, the condition can be recast so as to require that the agent’s satisfaction of D does not depend on any particular false beliefs (again, with the possible exception of the belief that figures in (i)).

18 With respect to iii (a), note that I make no assumptions about the ‘genuine’ normative significance of conventional norms. When they are not morally binding, then an informed agent would know this. As for iii (b), its motivation is articulated below (n 28).

19 A central theme of Rawls’ classic paper is that ‘in the case of actions [such as promising] specified by [rule-governed] practices it is logically impossible to perform them outside the stage-setting provided by those practices, for unless there is the practice, and unless the requisite proprieties are fulfilled, whatever one does, whatever movements one makes, will fail to count as a form of action which the practice specifies’, see Rawls (n 14) 25. Hart makes the same claim about promising at several points in The Concept of Law, see Hart (n 15) 43, 225, while Anscombe offered several original elaborations of Hume’s natural unintelligibility argument in GEM Anscombe, ‘On Promising and Its Justice, and Whether It Needs be Respected’ (1969) 3 Crítica: Revista Hispanoamericana de Filosofía 61; and GEM Anscombe, ‘Rules, Rights, and Promises’ (1978) 3 Midwest Studies in Philosophy 318.

20 Hume (n 5) 332. In their respective discussions of Humean theories of promising, Liam Murphy (a fellow traveler) and Seana Shiffrin (a critic) each mention only the second of Hume’s two claims, which they (somewhat understandably) misattribute to Anscombe, see Seana Shiffrin, ‘Promising, Intimate Relationships and Conventionalism’ (2008) 117 Philosophical Review 481; and Liam Murphy, ‘The Artificial Morality of Private Law: The Persistence of an Illusion’ (2020) 70 University of Toronto Law Journal 453. In the article cited by Shiffrin and Murphy, Anscombe defends only the first of Hume’s two claims, after carefully distinguishing between them, see Anscombe, ‘Rules, Rights, and Promises’ (n 20). Others take notice of Hume’s ‘natural unintelligibility’ claim only to reject it out of hand without bothering to reflect on its basis. For example, despite extensive engagement with Hume’s account of promising, Owens innocently asserts that ‘there is nothing unintelligible about such a [communicative act] – each party possesses the conceptual materials necessary to both make it and comprehend it when made’: Owens (n 4) 161.

21 As noted, Hume distinguishes between the two claims at the outset of the promising chapter in the Treatise, see Hume (n 5) 332, where he goes on to offer two distinct arguments in support of the first claim (i.e., the ‘natural unintelligibility’ claim), one in the main text (assuming his own sentimentalist metaethics), the other in the adjoining footnote (assuming moral rationalism for the sake of argument). Hume’s argument in the main text shares a central premise with my own (see infra n 33), while the spirit of his argument in the footnote is preserved by my argument (as indicated by my epigraph). My argument is also different from Anscombe’s mesmerizing ‘natural unintelligibility’ arguments (which themselves stand in an uncertain relation to Hume’s), though all the premises of my argument are endorsed by Anscombe in her earlier monograph Intention (see infra n 33).

22 As I note below (n 58) Hume’s second argument is strong in exactly the same sense as my own.

23 Quite apart from the natural unintelligibility of normative powers, the argument rules out so-called second-order reasons (that is, reasons to act for certain reasons), including Raz’s well-known category of exclusionary reasons. Similarly, the argument rules out the possibility of reflexive intentions (more specifically, reflexive intentions in action). I will leave it to the reader to draw out the former implication. The latter implication is briefly discussed below (n 41, 59).

24 Raz breaks the link between exercises of normative powers and the modification of normative properties by fiat by holding that ‘a person’s act is an exercise of a normative power if it brings about or prevents a normative change because it is, all things considered, desirable that that person should be able to bring the change about or prevent it by performing that act’: see Raz, ‘Normative Powers’ (n 2) 163, a reformulation of Raz, Practical Reason and Norms (n 2) 98. As Raz frankly acknowledges, for all that we know, this definition applies to the act of procreation as much as to promissory acts, see Raz, ‘Normative Powers’ (n 2). But, if we know anything, it is that any obligation incurred by procreation is not produced by stipulation, irrespective of whether one procreates with the aim of incurring the responsibility. Conversely, if one follows Raz in identifying the act of promising with the communication of an intention to thereby incur an obligation, then one does not need to accept Raz’s normative explanation of promissory obligations to conclude that any obligation immediately grounded in the fact that one has promised is the product of fiat or stipulation. Finally, while one can stipulate the meaning of a term like ‘normative powers’, it is not a matter of terminology whether one can perform the communicative act that Raz identifies with promising in the absence of social or conventional rules of promising.

25 According to the Will Theory, one exercises a normative power by ‘willing’ a normative change in the way one is sometimes said, in philosophical texts, to will so-called basic actions such as the intentional (and unaided) raising of one’s arm. In considering the Will Theory, it is worth remembering that, perhaps since Ryle, no proposition in the philosophy of action has commanded more assent than the one denying the existence of a private, mental action of willing (or trying) mediating between an intention, on the one hand, and a bodily movement, on the other, when somebody exercises the capacity to move their body at will. For illuminating discussion of Ryle’s actual arguments, see John Hyman, Action, Knowledge, and Will (Oxford University Press 2015) 20–23. And while some philosophers, e.g., Brian O'Shaughnessy, The Will: Volume 1, Dual Aspect Theory (Cambridge University Press 1980), have identified the act of (successfully) willing the movement of one’s body with the moving of one’s body, no analogous position is available to the Will theorist about normative powers: for a single act of willing cannot at once be identical with and serve as explanatory ground of the incurrence of an obligation. Moreover, while it may be tempting to update Will theories (of normative powers) by substituting ‘deciding’ for ‘willing’, this temptation should be resisted. (In the context of duty-incurring acts, such an updated version of the Will theory would hold not that one incurs an obligation by deciding to incur it – after all, nobody thinks I would incur an obligation merely in virtue of deciding now to make a promise next week – but rather by deciding to incur it by this very decision. Such a view would have a number of untoward features, however, including incompatibility with the principle that a decision to Φ must be able to give rise to an intention to Φ.) Finally, one must take care not to conflate the phenomenon of inner speech (e.g., silently uttering the English sentence ‘I forgive you Davida’) with the putative act of ‘willing’ a normative change. For these reasons, I believe that contemporary theories of normative powers that are cast in terms of willing – see, e.g., Ruth Chang, ‘Do We Have Normative Powers?’ (2020) 94 Aristotelian Society Supplementary Volume at 275 – are most charitably re-characterised, if possible, in terms of intentions (aims).

26 In the contemporary literature, the Aim Theory is sometimes adopted in the context of consent of Larry Alexander, ‘The Moral Magic of Consent (II)’ (1996) 2 Legal Theory 165; although Alexander’s formulation is slightly different, it should be charitably recast in terms of the Aim Principle). For uniformity’s sake (and since the plausibility of the theory is not relevant to my purposes), I consider the version of the Aim Theory applied to the context of duty-incurring acts (rather than grants of permission). Finally, below (n 41) I will take note of an important ambiguity in the above formulation of the Aim Principle (pertaining to the construal of ‘thereby’).

27 This assumption (reflected in the earlier definition of ‘natural intelligibility’) is innocent: for it is no significant weakening of the original claim (namely, that one cannot incur an obligation by fiat without exploiting conventional promissory norms) to acknowledge an exception for certain forms of irrationality.

28 This assumption is also innocent (and is also reflected in the prior definition of ‘natural intelligibility’), since it is no significant weakening of the claim that one cannot satisfy the Aim Principle absent conventional duty-imposing norms to draw an exception for such unusual cases where one’s act triggers an independent moral principle (e.g., as when one inflicts harm with the aim of incurring compensatory duties).

29 The ascetic’s consolation is loosely adapted from a prayer attributed to Thomas Merton. And while everyone would agree that a person of Merton’s distinction would not be caught squeezing doorknobs (etc.), one might reply that the reason for this is due not to general conditions of rational agency but rather to the appreciation of ‘local’ considerations about what pleases God (viz., that such acts of ‘bootstrapping’ are exceptions to the general rule that God is pleased by acts performed in order to please God). While a conclusive rejection of this approach would require a deeper excursion into the philosophy of religion than I can here undertake, we may easily discern its limitations. If the ascetic were to exploit her conclusion to develop novel forms of worship, the sincerity of the ascetic’s motives would not be in question – after all, why else would she be squeezing the doorknob? And since the conclusion the ascetic has reached in effect says that God is willing to overlook mistaken judgments concerning what pleases God, why would God carve out an exception for sincere (and harmless) bootstrapping? I return to this motivating example below (n 44), where I consider an alternative principle (governing rational agency) that might explain why such bootstrapping is not licensed by the ascetic's consoling conclusion.

30 Some readers will recognise that the constraint causes difficulties for Evidential Decision Theory (EDT).

31 For a cutting-edge discussion of the various species of contribution, see Mikayla Kelley, ‘How to Perform a Nonbasic Action?’ forthcoming in Nous.

32 The truism: if S did X in order to do Y, and succeeded in her aims, then S Y-ed by X-ing or by performing a series of acts that includes X-ing. One may wish to qualify this thesis in light of a purported distinction between ‘mere preparation’ and ‘the act itself’. That is, one might wish to say that while I made the omelet by breaking eggs (inter alia), I did not make the omelet by walking to the kitchen (inter alia), even though I walked to the kitchen in order to make an omelet. We may ignore this complication, as all the cases of relevance to this paper fall on the ‘act’ side of any such divide.

33 Hume employs the Contribution Condition in the first of his two arguments for the Natural Unintelligibility thesis: ‘’tis certain we can[not] … by a single act of our will, that is, by a promise, render any action [obligatory] … It wou’d be absurd, therefore, to will any new obligation … nor is it possible, that men cou’d naturally fall into so gross an absurdity.’ Hume (n 5) 332 (emphases added). According to Hume, if I will an obligation then what I’m doing is [performing an act of will in order to render an action obligatory], and this is absurd because it’s obviously impossible to incur an obligation by a single act of will. (If one somehow resists this reconstruction, and denies that the Contribution Condition is strictly implied, one must at least concede that the argument can be generalized in that direction without loss of its essential logic). In Intention, Anscombe endorses the Contribution Condition (‘a vague and general formula’) and discusses it at some length: ‘the [remote] state of affairs mentioned must be such that we can understand the agents thinking it will or may be brought about by the action about which he is being questioned’: see GEM Anscombe, Intention (Harvard University Press 2000) 35; id. §22 more generally.

34 I assume that, when an agent does Y by doing X, either their act of Y-ing is not identical to their act of X-ing, or, at the very least, the fact that they are doing Y is distinct from the fact they are doing X (because, say, the facts are about distinct properties of a single act). For helpful discussion, see Maria Alvarez and John Hyman, ‘Agents and their Actions’ (1998) 73 Philosophy 219, esp. section 4. In any case, notwithstanding my occasional use of shorthand in this paper, the Contribution Condition should be construed in terms of (beliefs about) facts about actions rather than (beliefs about) actions.

35 I consider below (n 42) whether the Contribution Condition can be modified in ways that would undermine my argument.

36 The first and third conditions together ensure that the (property corresponding to the) more specific description is ‘subdeterminate under’ (the property corresponding to) the less specific description, to adopt the more general terminology of Chisholm. (I will not pause to consider whether my second condition is redundant). As Chisholm puts it, ‘eating lobster is a subdeterminate under eating,’ where this means that while the property eating lobster ‘falls under’ the property eating – that is, while instantiating the former property entails that one instantiates the latter property, but not vice versa – the property eating ‘is not one of two [independent] properties which are together equivalent to’ eating lobster: see Roderick Chisholm, ‘Adverbs and Subdeterminates’ in Ernest Lepore and Brian P McLaughlin (eds), Actions and Events: Perspectives on the Philosophy of Donald Davidson (Oxford: Blackwell 1985) 326–28. Adapting Chisholm, two descriptions are logically independent of one another (in the relevant sense) just in case satisfaction of neither description entails, or is entailed by, satisfaction of the other description (or of any description logically or conceptually incompatible with the other description). Id., 325. The appeal to independence is intended to screen out ‘contrived’ equivalences, such as that obtaining between ‘eating lobster’ and the conjunction of ‘eating’ and ‘eating lobster’. Ibid.

37 Can this characterisation of ‘partial explanation’ be resisted? The only remotely plausible alternative of which I am aware involves holding, first, that the fact that Plum killed Boddy is a proper part of the fact that Plum killed Boddy with a knife, and second, that if one fact is a proper part of another, then the former partly explains the latter. However, this not only relies on the vexed idea that facts have proper parts (consider: is the fact that p a proper part of the fact that p or q?) but, worse still, on the rejection of the widely-held ‘weak supplementation principle’ (WSP), which ‘states that whenever an object has a proper part, it has another part that does not overlap – that is mereologically disjoint – from the first. This is a straightforward statement of a basic decomposition intuition, the idea that when a proper part is ‘removed’ from a whole, there must be another ‘supplementing’ disjoint proper part’: see AJ Cotnoir, ‘Is Weak Supplementation Analytic?’ in Massimiliano Carrara and Giorgio Lando (eds), Synthese (Special Issue: Mereology & Identity) (Springer 2018) 2. The rejection of WSP would be especially unfortunate in the context of the Contribution Condition, which is intended to capture the intuitive idea that an agent must believe that their action might go all or some of the way towards the realization of their goal. But where an agent knows that their having X-ed would (at most) go only some of the way, then surely they must believe that there might be something else (or some other things) that would, together with their having X-ed, complete the job.

38 The ‘highly intuitive’ claim follows because if the Explained is not overdetermined (in the above sense) and the Third Fact is not identical with the Explainer, then the third fact does not directly explain the Explained. And if the Explainer is the sole set of facts that does directly explain the Explained, then any other fact, or set of facts, explains the Explained only if it (directly or indirectly) explains the Explainer. Erasmus Mayr has suggested that the ‘highly intuitive claim’ may stand in need of qualification. If the Explainer fully and directly accounts for the Explained, this explanatory relation will in turn be associated with a Grounding Fact – namely, the fact that the Explainer accounts for the Explained. And while the Grounding Fact is identical neither with the Explainer nor with the Explained, it nonetheless arguably belongs to the direct explanation of the Explained, even where the Explained is not overdetermined. However, even if we were to qualify the ‘highly intuitive claim’ to exclude such Grounding Facts from its scope of application – and it is doubtful that we must, given that the Explainer would no longer fully explain the Explained if it needs assistance from the Grounding Fact – the argument that follows would not be impaired, as there is no such Grounding Fact to which my opponent can appeal that would blunt the force of the argument.

39 The above formulation assumes, controversially, that (non-conventional) moral principles do not, in general, figure among the explanatory grounds of their instances. Nothing turns on this simplifying assumption. If one rejects it, and thinks that the Aim Principle (if valid) figures in the full and direct explanation of the agent’s incurrence of obligation O, then we can reach the same conclusion – namely, the fact that S kicked a rock accounts for the fact that she incurred an obligation only if it accounts for the fact that she kicked the rock in order to incur an obligation – simply by generalising the ‘highly intuitive’ claim to reflect the possibility that the ‘Explainer’ can consist in a set of facts (rather than an individual fact). And since the fact that the agent kicked the rock in order to thereby incur the obligation does not belong to the explanation of the Aim Principle, it explains the Explainer – that is, the set consisting in the Aim Principle and the fact that the agent performed some act or other in order to thereby incur that obligation – only if it explains the latter fact.

40 The above claim constitutes the fourth and final plank of my argument. The claim is, of course, still controversial – indeed, Hume would not have accepted it – and should be rejected by one who both denies that ‘B caused S to do X’ is a specification of ‘S did X’ and holds that ‘S did X in order to Y’ is equivalent to ‘S’s intention to do Y caused S to do X’. Although this is not the place to adjudicate the dispute, it is worth flagging both that the causal analysis of ‘S did X in order to Y’ would (by all accounts) have to be supplemented by a condition requiring that the intention cause the action ‘in the right way’ and that (after decades of trying) the ‘right way’ has not, to my knowledge, been identified. Finally, it is worth relating my view to that offered by Anton Ford in a discussion to which I am indebted; see ‘Action and Generality,’ in Anton Ford, Jennifer Hornsby and Frederick Stoutland (eds), Essays on Anscombe's Intention (Harvard University Press 2011). While Ford defends the view that intentionally X-ing is a specification of X-ing, my specification claim concerns X-ing in order to Y.

41 Since the ‘thereby’ of the Aim Principle can be interpreted in either of two ways, it should be noted that the above claim about specifications – namely, that kicked a rock in order to thereby incur obligation X is a specification of kicked a rock – holds under either interpretation. On the first (‘non-reflexive’) construal, the description ‘kicked the rock in order to thereby incur obligation X’ is equivalent to ‘kicked the rock in order to incur obligation X by kicking the rock’; on the second (‘reflexive’) construal, it is equivalent to ‘kicked the rock in order to incur obligation X by kicking the rock with this aim’. Under either interpretation, ‘kicked the rock in order to thereby incur obligation X’ is a specification of ‘kicked the rock’. It bears noting that the reflexive construal faces further difficulties; after all, its condition is of the form ‘S does X in order to Y by doing X with the intention to Z’ and a rational agent can never satisfy sentences of that form without violating the Specification Constraint. (And if this rules out the possibility of a reflexive intention – as well as philosophical theories that rely on them – so much the better.) By contrast, the form ‘S does X in order to Y by doing X’ can be salvaged either by denying that ‘doing Y by doing X’ is a specification of ‘doing X’ or by equating it with ‘S did X in order to (bring about) Y, and S believed at the time of action that their having done X on that occasion will or may be (partly) constitutive of their having (brought about) Y.’ I will expand on this in a future paper.

42 Jacob Rosen has suggested (noncommittally) a way of modifying the Contribution Condition that would undermine my argument. Instead of requiring, as a condition of doing X in order to (bring about) Y, that the agent believe that their having X-d will or may come to belong to the explanation of Y’s coming to pass, the Modified Condition would require that the agent believe that their having X-d in order to (bring about) Y might come to belong to the explanation of Y’s coming to pass. Although I cannot fully argue the point here, it is unlikely that the Modified Condition delivers plausible verdicts about cases. When I heaved a brick at my neighbor’s window in order to break it, the fact that I broke the window is partly explained by the fact that I heaved a brick at it (from a certain distance, with a certain velocity, etc.), not by the fact that I did so in order to break it, see Stephen Yablo, ‘Mental Causation’ (1992) 101 The Philosophical Review 245. (Of course, the fact that I intended to break the window partly explains why I heaved a brick at it; however, the fact that I intended to break the window is not the same as the fact that I heaved the brick in order to break it). Felix Koch has observed that the counterexample can be accommodated by replacing the Modified Condition with a Disjunctive Condition consisting of the disjunction of the Modified Condition and the original Contribution Condition. Furthermore, he suggests that the Disjunctive Condition can be given a principled defense by appealing to the intuitive idea that the agent who performs X in order to (bring about) Y must believe that their performing X in the current circumstances may fully or partly account for Y. Insofar as the motive for which an agent acts belongs to the ‘circumstances’ of their action, the Disjunctive Condition can claim to capture the truth that lies in the vicinity of the Contribution Condition. My response to this formidable challenge has two parts: first, the aforementioned principled defense supports not the Disjunctive Condition but rather the following More General Condition: S does X in order to (bring about) Y only if S believes, at the time of action, that (their) Y(-ing) will or may come to pass either (partly) in virtue of their doing X on the occasion, or (partly) in virtue of their doing X’ on the occasion, where X’ is a specification of X that will be realized if S does X on the occasion. As far as I can see, there is no principled reason for proponents of the Disjunctive Condition to stop short of the More General Condition. Second, the More General Condition is false. Consider: Like Neo in the Matrix, when I am offered the choice between the red pill and the blue pill – knowing that the former would reveal a painful truth while the latter would leave in place a blissful illusion – I opt to swallow the red pill. In such a case, did I swallow a pill of any color in order to learn the truth? If the lefthand side of the italicized formula specifies a description under which the action is intentional, then clearly I have not, even though ‘swallow a red pill’ is a specification of ‘swallow a pill of any color’ that was realized on the occasion. To take another example, consider ‘I swallowed a red or blue pill in order to learn the truth.’ It sounds OK only if the context invites an extensional construal – e.g., years later, I cannot recall which pill was which, and mean only to say that the pill I swallowed was either red or blue, and that it was chosen in order to learn the truth. Likewise, if an utterance of ‘I swallowed a colored pill in order to learn the truth’ passes muster this is only because it is naturally interpreted as ‘there is a pill of some color, such that I swallowed it in order to … ’ In reply, Koch observes (correctly) that the More General Condition is (at most) only one necessary condition of acting with an aim, and proposes other conditions that might account for the absurdity of the utterances at issue. However, I do not think that the other conditions he mentions are fit for purpose in accounting for the absurdities. (In my example, I did believe that the red pill was colored; and although I did not believe that I was taking a pill of any color in order to learn the truth, this is only because such a belief is ruled out by my conditions.) Koch also suggests that pragmatic implicatures may be at work, but I do not think that the tests for implicatures are satisfied in these cases. (In this connection, it is also worth reiterating that the relatively unspecific ‘there was a colored pill such that I took it in order to learn the truth’ is not absurd at all). Finally, I note that while Koch refers to the More General Condition as ‘maximally general,’ there is in fact one condition that is more general still. The Most General Condition (as I will call it) weakens the More General Condition by replacing the requirement that the two act descriptions (X’ and X) stand in the specification relation with the weaker requirement that they refer to the same act token. For example, when one succeeds in making a bid by raising a paddle at an auction house, both descriptions refer to the same act, and yet neither is a specification of the other. As before, I do not know how a proponent of the Disjunctive Condition could have principled grounds for resisting the Most General Condition. And yet the Most General Condition is surely false. Consider a person of means at an auction who decides that scratching a hard-to-reach back itch with their auction paddle is worth paying an exorbitant sum for an unremarkable painting. While they raised their paddle in order to scratch their back, surely they did not bid on the painting in order to scratch their back; after all, they did not bid for any reason at all – rather, their bidding was a side effect of their raising their paddle.

43 As noted earlier, I have been assuming that the fact that the agent performed an act of the stipulative type figures in the direct explanation of the fact that the agent incurred the obligation. Let us relax the assumption and consider the following alternative explanatory structure:

Let ‘did A’ stand for ‘kicked a pebble’ and let ‘special sauce’ refer to the agent’s intention-in-action – on this alternative proposal, what directly explains my incurrence of obligation O is the fact that I kicked the pebble together with the fact that I intended to be there-and-then incurring an obligation. If this structure is to constitute a genuine alternative that requires separate treatment, we must assume (what is by no means obvious) that the latter fact is distinct from the fact that I performed some act or other in order to thereby incur an obligation. Even granting this assumption, we may proceed to dismiss this alternative structure on two grounds: First, intuitively, the bare fact that I kicked the pebble is not of direct normative significance. At most, it is the vessel or vehicle for what really matters (that is, for the facts that are of direct normative significance) and therefore should not figure in the direct explanation of the normative change. Second, it is mysterious how a proponent of this alternative structure would go about explaining why the intention-in-action (the special sauce) is not sufficient to produce the normative change. For the fact that I kicked the pebble does not even ensure that the intention-in-action was successfully executed, as one may intend to kick a pebble with their left foot only to trip and (inadvertently, in the course of regaining their balance) kick it with their right. Thanks to Jacob Rosen for urging me to consider this alternative structure.

44 Having presented the central argument, I would like to return to the earlier example involving the ascetic. Although I only claimed that my arguments would account for the relevant intuition (namely, that a rational ascetic could not perform an arbitrary action with the aim of pleasing God), it is tempting to make the stronger claim that the Contribution Condition is necessary to account for the intuition. However, in order to make this stronger claim we must rule out an alternative constraint on rational agency that would, if valid, also account for the same intuition. To explain the intuition one may appeal not to the Contribution Condition, but rather to the following principle (‘Selection Condition’): For S to do X in order to (bring about) Y, S must believe that her intention to (bring about) Y will or may (come to) belong to an explanation of her doing X on the occasion. This principle accounts for the intuition provided that it is understood to incorporate a notion of ‘explanation’ that has the following implication: if an agent believes that any action performed with the intention to please God would achieve this aim, then the agent’s intention to please God cannot explain why she did X in particular. While there are two interpretations of the Selection Condition that carry this implication, neither one is valid. On the stronger interpretation, for an agent’s intention Y to explain why the agent did X, the intention must explain why the agent performed X as opposed to each alternative to X that was available to her. However, given that a rational agent can sometimes choose arbitrarily between several available alternatives (e.g., buying chocolates or flowers) that would equally realize a given goal (e.g., pleasing one’s host), the strong interpretation of the Selection Condition can hardly be taken seriously. On the weaker interpretation, for an agent’s intention Y to explain why the agent did X, that intention must explain why the agent did X as opposed to at least one alternative to X that was available to her. However, it is difficult to discern a compelling rationale for the weaker interpretation of the Selection Condition: if an agent can choose arbitrarily between several acts any one of which would equally realize her aim, then why can she not similarly choose arbitrarily when any of her available options would similarly realize her aim? Furthermore, this weakened Selection Condition, even if valid, would not explain the rational constraints on the ascetic’s mode of worship if the principle that the ascetic accepts, concerning what would please God, carves out an exception for a certain class of acts (e.g., injurious or sacrilegious acts). For in that case, her aim to please God could explain why she squeezes the doorknob rather than perform an action belonging to the excepted class. And yet, for all that, her squeezing of the doorknob would be no less irrational. Accordingly, while I have no quarrel with the Selection Condition as such, none of the interpretations of the principle that would account for the irrationality of the ascetic’s arbitrary act pass muster.

45 Raz defends this principle only insofar as the ‘relationships [of obligation] between people’ that the principle enables are ‘held to be valuable’: see Raz, ‘Promises and Obligations’ (n 4) 228. The reference to ‘other defeating conditions’ in my gloss should be interpreted by reference to this limit. As noted earlier (n 4), Raz’s characterisation of the promissory act (incorporated in the Communication Principle) has been widely influential in the literature on normative powers generally.

46 Whether the Gricean should adopt the parenthetical alternative is discussed at length in Wayne A Davis, Meaning, Expression, and Thought (Cambridge University Press 2003). The Gricean framework is made explicit in Searle’s well-known analysis of promising, which informed Raz’s account, see John R Searle, Speech Acts: An Essay in the Philosophy of Language (Cambridge University Press 1969) 57–61. I do not myself rely on the Gricean framework. The leading competitor is a normative theory, according to which one communicates (in the relevant sense) in virtue of undertaking conversational commitments of various sorts. To reduce promising to communication in this latter sense would be to reduce one species of interpersonal commitment to another, and all the problems related to the identification of the normatively significant characterisation of the undertaking would recur with respect to the latter commitment.

47 I note that while the first simplification eliminates the reflexive element of the communicative aim, it is nonetheless compatible with three reflexive (as well as one non-reflexive) interpretations of the ‘thereby’ occurring within the scope of the communicated aim. According to the reflexive interpretations, the communicated aim is an intention to incur obligation O by uttering U with either the communicative aim (interpretation one), the communicated aim (interpretation two), or the communicative and communicated aim (interpretation three). According to the non-reflexive interpretation (interpretation four), the communicated aim is simply an intention to incur obligation O by uttering U. (In his commentary, Koch questions my argument as applied to interpretation one, and I will have to take up this formidable challenge elsewhere. However, I note that, as a normative matter, interpretation one of the Communication Principle is unsatisfactory. For it implies that one cannot effectively promise unless one is known to have the correct normative theory of the source of promissory obligations.) Finally, I note also that the simplifications of the principle in the main text are not essential – that is, the argument offered against the simplified principle can be extended to the more complex (unsimplified) principle.

48 While such alternative features may be formalistic (e.g., involving signs or seals), they need not be, and may relate to the more fluid, pragmatic properties of an utterance (including presuppositions). For example, Raz contends, plausibly enough, that ‘[i]n our culture, communicating an intention to act in a certain way to a person who is known to be interested in the action is conventionally regarded as [a promise]’: see Raz, ‘Voluntary Obligations and Normative Powers II’ (n 2) 100. Relatedly, while Anscombe remarked that ‘it was absurd of Hume to write as if there had to be a special sign of promising,’ the charge is misplaced: see Anscombe, ‘Rules, Rights, and Promises’ (n 19) 320, since Hume explicitly allows for the possibility of wholly tacit promises several chapters later in his discussion of allegiance. (‘A tacit promise is, where the will is signify’d by other more diffuse signs than those of speech; but a will there must certainly be in the case, and that can never escape the person’s notice, who exerted it, however silent or tacit’): see Hume (n 5) 351.

49 In recent decades, several legal commentators have pushed back against the so-called ‘objectivist’ interpretation of contracts by emphasizing such ‘subjective elements’, see Melvin Eisenberg, ‘The Responsive Model of Contract Law’ (1984) 36 Stanford Law Review 1107.

50 This reflects the position of the Restatement (Second) of Contract (1981), §20 (illustration 1). The example is adapted from the famous Raffles v Wichelhaus [1864].

51 In the Anglophone contract literature, committed ‘objectivists’ often fail to properly distinguish between an objective approach that privileges ‘objective intent’ and one that privileges a notion of objective meaning equivalent to ‘what was said’. This is likely owing to the traditional ‘plain meaning rule,’ which excludes extrinsic evidence of the parties’ intentions when there is a sufficiently clear integrated document embodying the agreement; given such a rule, there is usually little practical difference between the two approaches. While lip service has sometimes been paid to the latter of the two approaches – for example, Holmes famously wrote that ‘no one will understand the true theory of contract or be able even to discuss some fundamental questions intelligently until he has understood that … the making of a contract depends not on the agreement of two minds in one intention, but on the agreement of two sets of external signs – not on the parties’ having meant the same thing but on their having said the same thing’: see OW Holmes Jr., ‘The Path of the Law’ (1897) 10 Harvard Law Review 457 – the objective intent approach remains the dominant one among objectivists. Indeed, Holmes elsewhere formulates his view in terms of objective intent, see, e.g., OW Holmes Jr., ‘Agency II’ (1891) 5 Harvard Law Review 1. For helpful discussion see: William F Young, ‘Equivocation in the Making of Agreements’ (1964) 64 Columbia Law Review 619; George Palmer, ‘The Effect of Misunderstanding on Contract Formation and Reformation Under the Restatement of Contracts Second’ (1966) 65 Michigan Law Review 33; Eisenberg (n 49) 1107; and Timothy Endicott, ‘Objectivity, Subjectivity, and Incomplete Agreements’ in Jeremy Horder (ed), Oxford Essays in Jurisprudence: Fourth Series (Oxford University Press 2000). (While Endicott maintains that a certain subjective intention is required of an offeror – namely, ‘the intention to act in a way that, understood in the context, counts as agreeing’ – Endicott does not equate this required intention with the intention (aim) to incur an obligation. Ibid 162).

52 Even intent conditions in the rules governing the execution of wills (where the interest in fulfilling the actual wishes of the deceased is presumably quite strong) are construed objectively. As a leading casebook puts it, ‘the plain meaning of the words of a will cannot be disturbed by evidence that the testator intended another meaning.’ A corollary to this rule is that courts ordinarily will not correct ‘mistakes’ in wills even whether there is extrinsic evidence showing that the testator intended something other than what the words of the will instrument convey. Of course this leaves open the question of what counts as ‘plain meaning,’ but the basic idea seems to be that courts are looking for objective indicia of the testator’s intent, so much so that these indicia, where clear, control the disposition of the estate even when there is strong evidence that the indicia do not reflect what was actually going on in the testator’s mind, see Jesse Dukeminier and Robert H Sitkoff, Wills, Trusts, and Estates (Aspen Publishers 2013). It bears noting that an intent condition that singles out the intent imputable to the testator on the basis of evidence available to the executor during probate, say, would be equally objective as one that privileges the evidence available (to the witnesses, say) when the will is executed.

53 With respect to the former proposition (that we must distinguish between rhetoric and reality), I do not deny that Savigny – to take a prominent example – attributed to the Roman Law a subjective-intent requirement with respect to informal contracts, but I merely question the correctness of the attribution (by all accounts influenced by his Kantianism), see MD Howe, Justice Oliver Wendell Holmes: The Proving Years (Harvard University Press 1963) 232, 233. Similarly, it is worth noting that the widely held view that French contract law is a modern-day subjectivist outlier has been refuted in the scholarly literature, see Wayne Barnes, ‘The French Subjective Theory of Contract: Separating Rhetoric from Reality’ (2008) 83 Tulane Law Review 359. More generally, we may note that judicial techniques such as conclusive presumptions can serve to conceal the objective character of a rule. As Williston observed, ‘[d]oubtless the law is generally expressed in terms of subjective assent, rather than of objection expressions, the latter being said to be merely ‘evidence’ of the former, as for example in the so-called parol evidence rule; but when it is established that this is no rule of evidence the whole subjective theory falls to the ground. Under the guise of conclusive presumptions of mental assent from external acts, the law has been so built up that it can now be expressed accurately only by saying that the elements requisite for the formation of a contract are exclusively external’ (Williston 1920, §1536). With respect to the second proposition (that such false understandings may have an effect on legal doctrine), it is likely that, at various points in its development, the law’s treatment of topics such as the revocability of offers has been influenced by lofty subjectivist rhetoric about the meeting of minds. But such influence does not entail that the intent conditions were in fact ever construed by the law subjectively. Additionally, it is perhaps worth mentioning that avowedly objectivist understandings of intent conditions in power-conferring rules date back much further than Holmes. Talmudic law, for example, offers an avowedly objectivist understanding of the law governing interpersonal agreements, reflected in its application of the dictum ‘devarim she-balev einan devarim’ (‘undisclosed intentions are not words’) (Babylonian Talmud, Tractate Kiddushin 49b). Finally, it is worth noting an important contrast in the treatment given by Holmes (far and away the most influential objectivist in modern legal thought) to subjectivist approaches pertaining to criminal law and tort law, on the one hand, and contract law, on the other. Although he championed objectivist approaches in all three areas, with respect to the first two areas, where the rules under consideration are not power-conferring, Holmes maintained that the law itself evolved from embodying a thoroughly subjectivist approach (interested in the actual mental states of the offender) to a largely objectivist one. By contrast, in his treatment of the law of contract (where the intent requirements at issue do figure in power-conferring rules) there is no indication that, in his view, the law ever embodied a subjectivist approach, see OW Holmes Jr., The Common Law (Little, Brown & Co. 1881); and Howe (n 53) ch 9.

54 Armed with the distinction between subjective and objective intent, one might object that I have previously mischaracterized the Aim and Communication principles: properly construed, the intention condition (of either principle) should be construed objectively and not subjectively. Now, even if I were to concede the point and weaken my conclusion accordingly (by restricting it to putative non-conventional power-conferring norms with subjective intent conditions), the weakened conclusion would still be worth drawing, given the prevalence of subjective interpretations of the relevant domain of nonconventional morality (not to mention the independent significance of the constraints on agency that figure in my argument). However, I do not believe that the point should be conceded. I shall set aside the delicate question of whether an agent, in the absence of conventional norms, could perform an observable act that would give another person sufficient reason to impute to him the relevant intention. I also defer (until the next section) consideration of the extensional difficulties that beset even the objective versions of the Aim and Communication principles. Instead, I simply observe that the rationales that have been put forward for the objective construal of intent requirements in power-conferring norms are distinctly institutional in character and have no ready application to a non-conventional domain that neither presupposes a centralized enforcement mechanism nor assumes that agents lack intrinsic motivation to comply with the principles of true morality. Consider again a norm that confers a power on an agent to incur an obligation; let us suppose that such a promissory norm (as I will now call it for the sake of simplicity) has an objective intent condition that includes what I have called an uptake requirement on the part of the promisee. According to this norm, a promissory obligation is not extinguished even if the promisee learns, sometime after hearing the promissory utterance but before relying on it, that his prior imputation of subjective intent to the promisor was mistaken. Most of the justifications that have been given for such a rule are premised on the putative facts that, first, a speaker’s claim that they lacked the subjective intent to incur an obligation that they appeared to possess ‘would be hard [or merely costly] for the addressee to refute [in an enforcement proceeding], since it would relate to the addressor's own state of mind’; second, that it may be onerous for the addressee who does rely on misleading appearances prior to learning the truth to prove such reliance in court, see Eisenberg (n 49) 1119. Given these facts, the subjective requirement would undermine the policy of ‘security of transactions’ and would thereby dampen the wheels of commerce to the detriment of society at large. Ibid. For individuals would be less likely to rely on promissory undertakings (or would take more time before doing so, in order to more thoroughly collect evidence concerning the subjective attitudes of the addressor) if they thought that enforcing the promise would require rebutting the above claim or proving reliance. Again, considerations such as these are of doubtful relevance for a non-institutional morality.

55 The above formulation tweaks Raz’s account of legal powers and extends it to the conventional domain. (‘Let us agree, then, that an act is the exercise of a legal power only if it is recognized in law as effecting a legal change and if it is so recognized because, among other things, it is an action of a type which it is reasonable to expect to be performed for the most part only when the person concerned wants to bring about the legal change’): see Raz, ‘Voluntary Obligations and Normative Powers II’ (n 2) 82, 83. Although I find Raz’s account of legal powers attractive, I am not wedded to it. In particular, a case can be made that the class of (conventional) power-conferring rules should be defined with reference to whether their conditions in fact reliably screen for intent (to a reasonable degree), rather than with whether they are designed to screen for intent. Given the availability of this alternative, I do not ultimately rely on the assumption (reflected in the main text) that all conventional norms have intended designs (i.e., that they are brought into existence for reasons).

56 I digress to note a significant upshot of this account of normative powers. As a general matter, just like a duress (or knowledge) condition, an intent condition in a rule – power-conferring or otherwise – is outside the scope of an agent’s choice: that is, an agent never faces the bare choice (i.e., one where everything else is held fixed) of whether to fulfil the norm’s other conditions with or without an intent to incur an obligation, see TM Scanlon, Moral Dimensions: Permissibility, Meaning, Blame (Harvard University Press 2008) 60. Again, this species of deliberative insignificance of intent requirements is common to power-conferring rules and non-power conferring rules alike. For example, if a criminal statute provides that one is criminally liable for a certain act only if one performs the act intentionally, the provision carries no deliberative significance for an agent who is considering whether to perform that act, see GEM Anscombe, ‘Two Kinds of Error in Action’ (1963) 60 Journal of Philosophy 398, 399. However, given the above definition of normative powers, an intent condition of a power-conferring rule is deliberatively insignificant in a second respect. Assuming that conventional ‘power-conferring rules’ are designed to be, in Hart’s (avowedly oversimplified) formulation, ‘recipes for creating duties,’ an intent condition is trivially satisfied by whomever turns to the rule with the aim that the power-conferring rule is (by definition) designed to serve, see Hart (n 15) 33. (In formulating this last point, I set aside the issues concerning subjective intent considered in the previous section.).

57 The terminology will be defined below.

58 I note that Hume’s position on this point was similar. Hume argued, first, that anyone who holds that the act of promising is possible in the absence of a promising convention must identify that act with a certain ‘act of the mind’ (namely, the willing of an obligation); second, that due to principles governing agency this act cannot be performed either in the absence or the presence of a promising convention. That Hume held that the act cannot be performed even against the backdrop of a promising convention not only follows from his two arguments for the ‘natural unintelligibility’ thesis but is an implication that he appears to embrace. For example, it is notable that the second clause of the following sentence (unlike the first clause) is not modified by the term ‘naturally’: ‘A promise, therefore, is naturally something altogether unintelligible, nor is there any act of the mind belonging to it’: see Hume (n 5) 332. Similarly, when he maintains that nobody will be able to show that there is a ‘peculiar act of the mind annext to promises,’ he does not restrict his claim to promises that are made in the state of nature, see Hume (n 5) 333.

59 Two points of clarification: First, the argument below can be easily modified to reach other variants of the Aim Principle (dealing with other sorts of normative change). Second, I assume here, and for the remainder of this section, the non-reflexive interpretation of the Aim Principle (see n 41). This assumption is warranted, as it is not possible to incur an obligation by satisfying the condition of the reflexive version of the principle while remaining ignorant of the source of one’s obligation.

60 The point that such a case does not involve the exercise of a normative power is often made in the normative power literature, e.g., Raz, ‘Voluntary Obligations and Normative Powers II’ (n 2) 81; and Victor Tadros, ‘Appropriate Normative Powers’ (2020) 94 Aristotelian Society Supplementary Volume 301.

61 To be sure, it is possible that a failed attempt to humiliate or harm someone may generate secondary duties, but even here the ground of the secondary duties is the attempt to harm or humiliate, not the attempt to incur an obligation. Additionally, I note that the argument in the paragraph is not diminished if the Aim Principle is revised to require apparent (i.e., objective) intent rather than actual (i.e., subjective) intent. And the same is true of the argument against the Communication Principle in the next paragraph.

62 I digress to observe that communication (in the broadly Gricean sense) is not, in general, necessary for the exercise of conventional powers. Consider someone who deliberately exercises a power to vote on some important matter by following some mechanical procedure (e.g., pushing a button in a private booth). While such a voter intends to cast a vote, they often have no reason to care whether this intention is recognised by anyone – for example, it may be no part of the vote-counting procedure to determine the intentions of a voter – and hence will not aim (or appear to aim) to achieve such recognition.

63 Paul Grice, Studies in the Ways of Words (Harvard University Press 1989) 218.

64 To be clear, my point is that it is ad hoc to restrict the Communication Principle to ‘tellings’ rather than ‘showings’ as those notions are understood by broadly Gricean accounts (where the distinction is drawn at the level of communicative intentions). For incisive critical discussion of this aspect of the Gricean program, see Richard Moran, The Exchange of Words: Speech, Testimony, and Intersubjectivity (Oxford University Press 2018) ch 6.

65 For present purposes, we may define conventionalism about promising as the view that identifies promising with the triggering of a social rule of promising (a species of social rule that we were able to characterize without appealing to a prior notion of promising). According to a weak version of Conventionalism (‘Count-as Conventionalism’), while the prevailing social rules of promising determine whether a given utterance qualifies as a promise to perform some course of conduct, once it is determined that such a promise was made, nonconventional morality takes over from there to determine the (genuine) normative upshot of the promise. By contrast, a stronger version of conventionalism (‘Thoroughgoing Conventionalism’) holds that the normative upshot of one’s promise to Φ is (subject to moral constraints) also determined by the conventional rule that one has triggered – for example, the particularities of the rule’s provisions related to outsourcing, delegation, defeasing conditions, etc. may be binding in morality (subject to moral constraints). I defend the stronger version elsewhere, see Jed Lewinsohn, ‘By Convention Alone: Assignable Rights, Dischargeable Debts, and the Distinctiveness of the Commercial Sphere’ (2023) 133 Ethics 231.

66 As my aim here is schematic, I set aside the possibility (worthy of investigation) that such notions as the owner’s acquiescence can be appealed to in lieu of their grant of permission.