A. Preliminaries

As I have noted elsewhere,Footnote ⁵ as with many issues in contemporary analytic jurisprudence, the story can be traced back (at least) to H.L.A. Hart:

Y’s mother is a third-party beneficiary under this norm. Now for Hart, a will theorist, for whom the function of rights is to protect the right-holder’s control over a duty, it is Y, and not Y’s mother, who possesses a right in this scenario: Y does, whereas Y’s mother does not, have control over X’s duty in the sense of power to demand/waive its performance. Additionally, by Hart’s lights, the interest theory, according to which rights function to protect the right-holder’s interests, is unable to account for Y’s having a right, and falsely decrees that Y’s mother has a right, in this scenario. Be this as it may, although Hart does not extend matters under this scenario to other agents who may benefit from X’s performance of his duty (e.g., alternative care-givers who would have to step in should X not perform, friends of Y’s mother who will enjoy her increased vitality as a result of the care, etc.), the implication is clear. For virtually all duties, many agents will stand to benefit from performance of the relevant duty. But it is an adequacy constraint on any theory of rights that not all such agents get to be right-holders. We can think of this delimitation challenge on a theory of rights most generally as the third party beneficiary objection. Insofar as the interest theory identifies having a right with such a benefit, the foregoing adequacy constraint will not be met. An explosion of rights ensues.

With the dialectic so configured, the interest theorist’s response should be obvious: viz. reject the identification of having a right with being a person benefiting from the performance of a duty as a component of the interest theory. It is only such an identification which leads to an interest theory explosion of rights. Indeed, such a rejection is present in the foremost contemporary formulation of the interest theory, as advanced by Kramer:

(The second thesis distinctive of the interest theory, for Kramer, is that control over the relevant duty is neither necessary nor sufficient for right-holding for the interest theory. By contrast, said control is necessary and sufficient for the will theory, meaning it does not face this delimitation challenge in any problematic sense. Leaving aside the slight oddity of including a negative condition like this in the analysis of the interest theory, it is not germane for present purposes.) Now straightforwardly, this formulation, (IT-1), does not identify being a person benefiting from the performance of a duty with being a right-holder: such a status is only a necessary, but not also a sufficient, condition on being a right-holder. As such, this formulation of the interest theory avoids mandating an explosion of rights.

But it achieves this at a cost of incompleteness. What IT-1 leaves us with, is not a theory of rights, in any full-blooded sense. It is not a theory of rights, insofar as such a theory must offer an analysis of rights (where an analysis would offer a set of necessary and sufficient conditions for right-holding). Of course, at least implicitly, Kramer is aware of this. While (IT-1) does not mandate an explosion of right-holders, for all it says, it leaves open such a possibility. Put vernacularly, without augmentation by a sufficient condition to close the loop, the interest theory is incomplete as a full-blown theory of rights and is open to charges of rights explosion. And in a sense to be made clear presently, Bentham’s test provides just that augmentation (for Kramer).

B. Example

So much by way of preliminaries. Let me now set out a core, running third party beneficiary example, which is an extension of Hart’s:

I find it helpful to represent this diagrammatically as:

The vertical line connecting X and Y represents the promise constituting the contract between X and Y; the diagonal arrow represents Y’s duty to pay Z 100 dollars; and the horizontal dotted arrow between Z and W represents Z’s plans—disposition—to spend that 100 dollars at W’s store. The set of dots after W indicates that we need not—though the example does—stop at W. The set of agents who stand to benefit from Y’s performance of his duty can ramify indefinitely; thus, representing the rights explosion issue.

Three quick points are in order as a gloss on the foregoing diagram. First, as the case is set up (partially in order for it to serve to distinguish between prominent competing theories of rights), it is stipulated that X alone has the relevant control over Y’s duty. Second, it may be helpful to flag that the fact that Y has a duty with the content ‘pay Z 100 dollars’ is not in doubt. What is in doubt is to whom said duty is owed. Third, a key figure, ‘Z’s GM’, will be omitted until our initial stage-setting is complete. That figure is Z’s grandmother, or granny. Suffice it to say for now, her role in the set-up is to, in some sense, mirror X’s: while X’s (the promisee’s) and Z’s GM’s (a close relative of Z’s) interests in seeing the promise constituting the contract performed seem radically different in kind at one level (X is, after all, a party to the promise whereas Z’s GM is not), at another level they are not. At this other level, both X and Z’s GM are in a sense to be fully explained presently, constitutively benefited by performance of the promise constituting the contract. Sreenivasan, as we shall see, trades on this similarity to the putative detriment of interest theorists.

C. The Interest Theory’s Response and the Bigger Picture

As foreshadowed, the response of the foremost contemporary advocate of the interest theory rests upon invocation of Bentham’s test. Rolling out Bentham’s test with respect to our running example (minus Z’s GM) will be relatively straightforward. I want to suggest that there is a reason for that. Our running example (minus Z’s GM) is an exemplar for a set of cases I shall dub non-constitutive third-party beneficiary cases, in the sense that the third party benefits non-constitutively from performance, and on the flip-side, suffers a detriment non-constitutively from non-performance.

Naturally, this class of cases is in contrast with constitutive third party beneficiary cases, which I want to say are exemplified upon introduction of Z’s GM. Focusing on detriment scenarios, a defining feature of constitutive cases is that the relevant third parties are implicated in non-performance via one act—they are one act cases.Footnote ⁹ By contrast, in non-constitutive cases, the relevant third parties are implicated in non-performance via two (or more) acts—they are two act cases. I want to suggest that Bentham’s test can be straightforwardly rolled out by interest theorists in non-constitutive, but not in constitutive, cases; that is, in two act, but not in one act, cases.

Let us then quickly roll out the interest theory’s (aka Bentham’s test) verdict on our basic running example. I have argued in detail elsewhere, and so shall simply state here: Bentham’s test, unlike the foregoing (IT-1), which is a mere necessary condition for right-holding, just is the interest theory of rights in the sense of stating necessary and sufficient conditions for right-holding.Footnote ¹⁰ This renders Sreenivasan’s Granny Attack on Bentham’s test all the more pressing: the interest theory stands and falls with Bentham’s test. Thus, we first need the test itself:

And, second, an account of a key technical term contained therein:

Although certain complications in applying this test will emerge as we come to introduce Gopal’s Granny, for the time being, its application to just X, Z, and W is tolerably clear.Footnote ¹³ X’s, qua promisee’s, interests as a moral agent in having promises made to him kept is set back in the event of non-fulfilment of the promise by Y.Footnote ¹⁴ That fact is, let us suppose, minimally sufficient for breach of the contract.Footnote ¹⁵ Z’s interests in getting the promised 100 dollars is also set back in the event of Y’s non-fulfilment of the promise. That fact is also minimally sufficient for breach of the contract. By contrast, though (i) Y’s non-fulfilment of the promise to pay Z, and (ii) Z’s resultantly not patronizing W’s store, is a jointly sufficient set of facts for breach of the contract, it is not minimally sufficient, as the element of the set including detriment to W is, palpably, redundant. Thus far, it is relatively straightforward. Moreover, this is all to be expected: W, as a non-constitutive third party beneficiary, suffering detriment in two acts ((i) and (ii)), is readily handled by Bentham’s test.

Before the introduction of problematic constitutive, one act third party beneficiary cases, I want to zoom out to the bigger picture, and to say a few telling words about some other theorists’ responses to our core, running third party beneficiary example, in particular Joseph Raz’s. It turns out that consideration of Raz’s interest theory in this context proves to be methodologically highly illuminating. Sreenivasan writes suggestively on this matter (transposing matters to our running example):

This passage is suggestive in the extreme. Let us start with what is clear. According to Sreenivasan, WFootnote ¹⁷ is ruled out as a right-holder for Raz on account of his interest ‘in the [$]100’ being insufficient to ground a duty on Y to pay Z $100 (regardless of to whom else we consider the said duty being owed). While this is true, it is, it seems to me, utterly irrelevant. To see why, note that the interest which Sreenivasan ascribes to W in this context is what I shall refer to as normatively neutral—in the sense that it is a bare sum of money and invokes no normative elements or statuses, here, that a promise to pay this sum of money has been made to a promisee.

This strategy of Sreenivasan’s of retreating to a bare, normatively neutral description of the relevant interest being set back is no accident and is repeated by him in his more recent consideration of cases in the realm of the third party beneficiary objection. In his treatment of such cases, we see Sreenivasan make two telling remarks on certain proscribed descriptions of the relevant interests:

Now while Sreenivasan is sparring with Kramer in this passage, we can see that in his application of Raz’s interest theory, he eschews what we can refer to as normatively loaded descriptions of the relevant interests. This refers to the sense of invoking normative elements or statuses here, that a promise to pay the sum of money in question has been made to a promisee. Thus, Sreenivasan’s methodological point is a general one and resultantly, our response to it is also of general importance. The basic point will be that restriction to bare normatively neutral descriptions of the relevant interests will lead to muddled results in teasing out a full Razian response to the core third party beneficiary case. Moreover, moving on to Kramer’s Bentham’s test, no vacuity or circularity issues of the kind Sreenivasan diagnoses arise if we are to allow certain normatively loaded descriptions of the relevant interests in our application thereof. Let me make these related points by turn.

Let us take Raz first. The problem arises when we turn (as Sreenivasan does not) to explain a Razian account of X’s and Z’s status as right-holders (or not) in this scenario. The basic point is, armed with a purely normatively neutral account of the relevant interests, we land in a muddle. Clearly, any account of X’s status as a right-holder here must appeal to X’s status as a promisee, to whom a promise by Y to pay Z $100 has been made. It is only by appeal to X’s normatively loaded status as such a moral agent that we can begin to explain X’s status as a Razian right-holder. So much, so staggeringly obvious.

Things get more interesting as we transition to Z. Is Z a Razian right-holder? The correct answer here is, or should, it seems to me, be: I do not know. Put differently, I think it is very much an open question of whether or not Z is a Razian right-holder. In addition, perhaps, it is none the worse for a theory of rights for delivering such a verdict.Footnote ¹⁹ But here is the key point. If we restrict ourselves, as Sreenivasan demands, to a purely normatively neutral account of Z’s relevant interest—here ‘in the [$]100’—Z’s status as a Razian right-holder is in no sense an open question. Quite the contrary, so described, and absent any description of Y’s making of a promise to X to pay Z $100, Z’s interest in $100 is no different to any other bystander’s interest in getting $100 from Y. Clearly, neither Z nor any other bystander has such a right. By contrast, Z’s status as a Razian right-holder rightly becomes an open question when, and only when, we describe Z’s relevant interest in a normatively loaded fashion, here, in the agreed-upon $100 (that is, in the $100 Y promised X to pay Z).

That is why, upon consideration of Raz, muddled results can only be avoided by advancing beyond purely normatively neutral descriptions of the relevant interests. It remains to pinpoint why, pace Sreenivasan, no vacuity or circularity arises by allowing reference to certain normatively loaded descriptions of the relevant interests in application of Bentham’s test. In the excerpt from Sreenivasan immediately above, he countenances two kinds of descriptions of the relevant interests being set back, each of which he takes to be prohibited from being invoked in application of Bentham’s test on pain of circularity:

1. (1) detriment in breach [of the contract].Footnote ²⁰

2. (2) [detriment in] breach of the promise.

Why is there, for Sreenivasan, the spectre of (vicious) circularity here? Bentham’s test, we have seen, asks us in the legal case, to hunt for facts which are (minimally) sufficient to constitute breach of the contract in question. Invocation of (1) and/or (2) in such a set of facts does on the face of it, smell of circularity.

Now on any rendering of constraints on circularity, the following one-premise argument is circular: p, therefore p. That is to say, an argument is categorically circular if the premise just is the conclusion. So much is uncontroversial. However, it is an interesting feature of the literature on the bounds of circularity (or question-begging) within philosophical logic, that beyond this clear case of circularity, it is extraordinarily difficult to circumscribe the bounds of this epistemic (or dialectical) vice. In particular, one adequacy constraint that must be assiduously avoided is that the account must avoid decrying all valid one-premise arguments circular. Let me explain.

Consider any of the following quotidian valid one-premise arguments.Footnote ²¹ The ball is red; therefore, the ball is coloured. Similarly, there are over 40 students in the lecture hall, therefore, there are over 10 students in the lecture hall. The point is that these are perfectly kosher valid arguments, in which the premise entails the conclusion, and where no whiff of circularity arises. We have clearly stepped away from the domain of arguments where the premise just is the conclusion. However, as a general point, as valid arguments, there is some clear sense in which the premise presupposes the conclusion, or in which the premise already contains the conclusion. Rehearsing the argument is a legitimate way of teasing out these facts. The point is, unless we want to operate with an account of circularity decrying all valid one-premise arguments viciously circular, when we come to extend the bounds of vicious circularity beyond arguments where the premise just is the conclusion, we cannot rest content with unexamined invocations of presupposition or containment along the lines advanced above.

I trust that the carry-over to Bentham’s test should be obvious. For the same reasons as for our proscription on the uncontroversial circular argument, p, therefore p, any invocation of:

1. (1) detriment in breach [of the contract]

in applying Bentham’s test will be impermissibly circular. But it is not at all clear that invocation of:

1. (2) [detriment in] breach of the promise

suffers from a similar epistemic (or dialectical) vice. On the contrary, invocation of detriment in breach of the promise as a component of a minimally sufficient set of fact for breach of the contract looks very much like the foregoing (perfectly kosher) red-to-coloured transition, for example. Just as being red is a constitutive way of being coloured, so the promise in question is (partly) constitutive of the contract.

Now my purpose here is not to settle this intractable debate in philosophical logic once and for all. It is merely to counter Sreenivasan. Even if invocation of (1) vacuously drains Bentham’s test of content, he is not warranted in asserting that invocation of (2) has a similar result. Indeed, I have pressed that invocation of (2) in the context of Bentham’s test resembles some perfectly respectable valid one-premise arguments. As a final nail in the coffin, it is only, as we have seen, by invoking normatively loaded facts, such as (2), when applying a major test of right-holding such as Raz’s interest theory, that we can avoid muddled results.

A. The Example and Initial Skirmishes

It is helpful, I think, to reprise our diagram from section 1.A above:

The character whom we have assiduously avoided thus far is Z’s GM (Z’s grandmother). Sreenivasan’s example (implicitly) trades on the fact that, for all their obvious differences, both X, the promisee, and Z’s GM are constitutively benefited by performance of the promise constituting the contract.

Essentially, Sreenivasan is asking, what about Z’s grandmother (Z’s GM), who has an interest in seeing her grandchildren receive benefits? Sreenivasan says it is hard to see how any version of Bentham’s test classifying X as a right-holder can avoid also classifying Z’s grandmother.Footnote ²³ Moreover, in contrast, how any version excluding Z’s grandmother as a right-holder, can avoid also excluding X. Either way: counterintuitive upshots.

Now Sreenivasan considers three options for an interest theorist to distinguish Z’s GM from X (and Z). Let us set them out, as one of the options has been, in a sense, seized upon by Kramer:

Now whether or not these options of Sreenivasan exhaust the options available to the interest theorist is somewhat beside the point. The point is that they are a set of good options which Sreenivasan has descried. I think we can safely conclude with Sreenivasan, that the first and third distinguishing options for the interest theorist are, ultimately, no-goers. The second requires more scrutiny, not least because it was on one natural way of reconstructing the dialectic here, initially seized upon by Kramer, and developed as his response to this distinguishing challenge.

B. The (Wholly) Vicarious/Non-Vicarious Distinction

On one natural way of reconstructing the dialectic here, Kramer seized upon Sreenivasan’s notion of parasitic/non-parasitic interests to attempt to carve out a distinction between types of interest that could prise apart Z’s GM from X (and Z). Kramer asserts that ‘wholly vicarious’ interests are interests of the wrong kind in counting for whether an agent is, or gets to be, a right-holder. That is to say, ruled out in such a consideration are interests that ‘reside wholly in the furtherance of somebody else’s interest’.Footnote ²⁵

It must be acknowledged that, before, and in the course of application of the operative distinction to separate Z’s GM from X (and Z), Kramer does say some further things in elucidation thereof. Eligible non-vicarious interests do not have to be ‘selfish in any way’. Additionally, as a final word on the mechanics of the distinction in question, ‘An interest that underlies R’s status as a right-holder—which might be an altruistically oriented interest, of course—has to be R’s own interest rather than someone else’s interest’.Footnote ²⁶ The application then supposedly falls out: ‘[Z’s GM] [by contrast with X (and Z)] does not have any non-vicarious interest in the fulfilment of [Y’s duty].’Footnote ²⁷ As such, on this line, Z’s GM gets cut off as a right-holder, while leaving X’s (and Z’s) statuses as such intact.

Now Sreenivasan puts sustained pressure on the efficacy of this distinction.Footnote ²⁸ We can, very briefly, set out these pressure-points. First, Sreenivasan presses, does not provide an explanation of X’s, the promisee’s, damage as a moral agent in this scenario (impermissibly), who also needs to invoke Z’s detriment? Second, is it not the interest of Z’s GM (admittedly in Z’s welfare) in some deep sense her own interest, such that on any reckoning it would fail to be wholly vicarious. Finally, Sreenivasan presents a motley set of objections that put pressure on the sustainability of the operative distinction itself.

A. Preliminaries

We were able to be refreshingly brief in the preceding section in articulating the interest theorist’s initial response to the challenge of separating Z’s GM from X (and Z). This is for a number of reasons, but most notably because Kramer, the leading interest theorist, no longer thinks this response, viz. invoking the (wholly) vicarious/non-vicarious distinction, is necessary to repel Sreenivasan’s Granny Attack. That is to say, while the leading interest theorist maintains the cogency and importance of the distinction itself, and is eager to provide a more detailed account of the distinction in subsequent work, his latest view is that invoking the distinction to rule out Z’s GM as a right-holder (without, of course, ruling out X (and Z)) is unnecessary.

This is significant. We can roughly divide possible responses to Sreenivasan’s Granny Attack—efforts to prise apart Z’s GM (not a right-holder) from X (and Z) (right-holder(s))—into two kinds. Responses may be evaluatively neutral, which is structural or formal in nature, or they may be evaluatively loaded. Of course, this is hardly going to be a neat distinction, and it may be more accurate to array responses on a spectrum.

Thus far, in section 2.A above, when reviewing Sreenivasan’s conjuring of possible interest theory responses to his Granny Attack, we saw two (ultimately doomed) responses which may aptly be classed as evaluatively neutral, in the sense of resting on structural or formal properties to distinguish Z’s GM from X and (Z). First, Z’s GM is not a party to the promise (but nor—sadly for this response—was Z, an interest theory right-holder). Thus, this is a feature that simply appeals to a status of Z’s GM. Of course, as a caveat, it is putatively an evaluatively significant status, and so this is why we are in the realm of a spectrum rather than a neat distinction. The point remains that the status itself can be specified without recourse to a substantive moral argument. Likewise, moving onto the second relevant response in this context, Z’s GM’s arguable unimportance, where such unimportance is cashed out in terms of the modal notion of (non-)necessity for breach, is aptly classed as evaluatively neutral (with the foregoing caveat).

The key issue for present purposes is that any response to Sreenivasan’s Granny Attack invoking the (wholly) vicarious/non-vicarious distinction is clearly going to fall on the evaluatively loaded side of this distinction (or spectrum), in the sense that the distinction’s full specification clearly rests on substantive moral argument in charting some region within a broader theory of well-being. Now no claim ought to be made that invocation of evaluatively laden moving parts within one’s theory of rights is ruled out of court. In fact, Kramer himself has acknowledged that his own interest theory rests on a so-called thinly evaluative account of the good.Footnote ²⁹ Additionally, Kramer’s overall theory of rights is partly grounded on an ethical account of the range of potential right-holders.Footnote ³⁰ Moreover, of course other interest theories (e.g., Raz’s) are even more open about their extensive evaluative commitments. However, it is clear that all things equal, Kramer takes it as a methodological desideratum to make as few such evaluative commitments as possible (in construction of a theory of legal rights, at any rate).

Accordingly it is natural to frame this part of the theory of rights debates in the foregoing terms, and to see Kramer’s motivation to dispense with invocation of the evaluatively laden (wholly) vicarious/non-vicarious distinction in answering Sreenivasan, in favour of a purely structural or formal response, in terms of the avoidance of evaluative commitments. Alternatively, if this is too strong, one can see Kramer’s motivation in terms of the simplicity of a response which purely invokes the structural or formal notion of minimal sufficiency (which is in any event integral and central to his account of right-holding), with the resultant downplaying of the role of the evaluatively laden (wholly) vicarious/non-vicarious distinction being a salutary side-effect.

Thus, taking stock, the foremost contemporary interest theorist initially responded to Sreenivasan’s Granny Attack by appealing to an evaluatively laden distinction to add to the essential structural or formal machinery of Bentham’s test. The latest set of responses, as we shall see presently, attempts to shake off this evaluative overlay, and to claim that the purely structural or formal notion of minimal sufficiency operative in Bentham’s test is alone adequate to handle Sreenivasan’s Granny Attack. However, my claim will be that this latest evaluatively neutral set of responses fails; and, moreover, that no evaluatively neutral response can work. Put differently, the claim will be that, for all that has been said so far in this debate, some kind of (problematic) evaluative overlay, along the lines of the (wholly) vicarious/non-vicarious distinction, must be injected into the essentially formal mechanism of Bentham’s test in order for the interest theory to be viable.

I also think a final important contrast must be drawn here. The extra evaluative overlay is necessitated here to respond to an example centred on a core, quotidian third party beneficiary contract. By contrast, one abovementioned evaluative commitment of the interest theory was necessitated by consideration of whether animals and dead people are within the range of potential right-holders. Now I do not for a moment seek to downplay or diminish the potential ethical importance of such a consideration. But I hope the contrast is clear: the evaluative commitments which I am insisting on here strikes at the core of the interest theory, and not at its margins.Footnote ³¹

B. Kurki’s Recent Interest Theorist Response

We ought to recall, from section 1.A above, Bentham’s test itself:

And, second, an account of a key technical term contained therein:

Thus, the latest interest theory claim is that Bentham’s test alone, and the associated condition of minimal sufficiency (absent invocation of non-vicarious interests), can readily handle Sreenivasan’s Granny Attack.

How so? Here is Kurki, crisply, on the matter (transposed to our running example):

The main point is, on purely structural or formal grounds, Kurki takes himself to have demonstrated the intuitive status of Z’s GM as a non-right-holder, on behalf of interest theorists, on account of the failure of setback to her interests to be minimally sufficient for breach of the contract.Footnote ³³ In contrast, as Kurki neatly establishes, this setback is redundant. Another key point for the purpose of this paper is that Kramer himself has given his personal seal of endorsement to the general approach of deflecting Sreenivasan’s Granny Attack purely by appealing to minimal sufficiency.Footnote ³⁴ Therefore, pinpointing its failure is highly illuminating and important.

C. Its Failure

The problem for Kurki here is that, on this (permissible) way of carving up the facts, we can run exactly the same line of reasoning, mutatis mutandis, to rule out X, the promisee, as a right-holder. Replace (b) above with:

1. (b*) X’s interests as a promisee are set back because of (a).Footnote ³⁵

On this (permissible) way of carving up the facts—to reprise and apply Sreenivasan’s thought that Z’s GM and X stand and fall together as right-holders—in excluding Z’s grandmother, as Kurki does, we have (impermissibly) paved the way, by means of (b*), for also excluding X. What is sauce for the goose is sauce for the gander: If Z’s GM’s setback is redundant, so too, highly problematically for the interest theorist, becomes X’s, the promisee’s.

A. Re-prising apart Z’s GM and X

Now, in itself, I take my diagnosis, in the previous section 3.C, of the failure of the recent effort of Kurki’s to give an evaluatively neutral answer to Sreenivasan’s Granny Attack, to be successful. But the debate can fruitfully be extended, drawing on a likely line of reply to my efforts to tie together Z’s GM and X, the promisee.Footnote ³⁶

Here then is a likely way to try to re-prise apart Z’s GM and X. It will be helpful here to set out on display the three key theses ((a) is duplicated below) discussed thus far:

1. (a) Y has not paid the agreed-upon money to Z.

2. (b) Z’s grandmother’s interests as a grandmother are set back because of (a).

1. (a) Y has not paid the agreed-upon money to Z.

2. (b*) X’s interests as a promisee are set back because of (a).

Kurki began by pointing out that (b) entails (a), rendering Z’s GM’s detriment in (b) redundant. I countered by pressing that, just as (b) entails (a), so too does (b*) entail (a)—rendering X’s, the promisee’s, detriment in (b*) redundant.

Be all this as it may, here, it could be pressed, is a significant remaining asymmetry between Z’s GM and X. Not only, it might be said, does (b*) entail (a), but, moreover (a) entails (b*). That is to say, it might be pressed that (b*) and (a) stand in an equivalence relation. Put less formally, it might be pressed that a setback to X, the promisee, and Y’s non-payment to Z, are intimately connected (in a way that contrasts with the relationship between a setback to Z’s GM and Y’s non-payment to Z).

To complete this effort at asymmetry, while (b) entails (a), it is not, it might be said, the case that (a) entails (b). This is to say, it might be pressed that (b) and (a) do not stand in an equivalence relationship. Put less formally, it might be pressed that a setback to Z’s GM and Y’s non-payment to Z are not intimately connected (in a way that contrasts with the relationship between a setback to X, the promisee, and Y’s non-payment to Z).

To tell a story about this claimed asymmetry: (a) entails (b*), in the sense that the promisee (putatively, of necessity) exists in this scenario; but (a) does not entail (b), in the sense that Z’s GM need not even exist. There is, according to this response, something contingent, and hence redundant, about Z’s GM (her very existence) which is decidedly different with respect to X, the promisee.

If this claimed asymmetry can be vindicated, then the Kurki effort to respond to Sreenivasan’s Granny Attack, armed with minimal sufficiency alone, will also be vindicated: we can, put vernacularly chop off Z’s GM’s detriment from the relevant detriment to Z, in a way in which we cannot chop off X’s, the promisee’s, detriment from the relevant detriment to Z.

B. Waiver as the Deontic Analogue of Non-existence

The problem with this line of response is that, while it is true that Z’s GM can go out of existence, so too—distinct scenario—can X qua promisee, on account of the possibility of waiver.Footnote ³⁷ We might say that waiver is the deontic analogue of non-existence on the part of Z’s GM. This means that just as (a) does not entail (b) (and so there is non-equivalence between (a) and (b)), nor does (a) entail (b*) (and so there is also non-equivalence between (a) and (b*)). Let me explain.

To quickly dispense with it, let me first say one thing that the (non-)existence line of response is surely not contending. There is, or should be, no contention here that it is somehow impermissible to invoke the existence of a grandmother of Z’s in setting up one’s counterexample. A philosopher is perfectly at liberty to generate a counterexample of his or her choice, invoking the existence of whatever character he or she wishes. That is not the point at issue.

The point at issue is rather that this line of response is contending that Z’s GM can, modally, (very easily) drop out of the picture; counterfactually, she (very easily) does not have to exist. Insofar, as is the case, as this is true, her detriment might be said to become redundant (a modal notion), and not, thus, minimally sufficient for breach of the contract.

However, the issue is that we can, by mimicking my initial reply to Kurki, run an exactly similar line of reasoning to this possible comeback from Kurki vis-à-vis X. X, qua promisee, can, modally, also (very easily) drop out of the picture; counterfactually, there can (very easily) be waiver. X need not exist qua promisee.Footnote ³⁸ If such a waiver occurs, there is no detriment to X, qua promisee, since there is no longer an operative promise.Footnote ³⁹ In the circumstances of Y’s non-payment of the agreed upon money to Z in the context of waiver, the cancelling effect of waiver means that there is no breach of duty and thus no relevant detriment to X qua promisee (at the hands of a duty-bearer). This is the sense in which waiver is the deontic analogue of non-existence.

Now all sides to this debate are agreed that the existence of a duty borne by Y is presupposed to exist.Footnote ⁴⁰ But the key point at stake here is that no-one should allow some stipulation, by diktat, by the interest theorist, of a duty’s continued existence across all relevant possible worlds. Any such stipulation would radically impoverish the explanatory power of the interest theory. As such, the possibility of waiver is a relevant one; and the actualisation of it is the deontic equivalent of Z’s GM going out of existence. Any move, that is, to render Z’s GM’s setback to interests (physically) redundant can be reproduced (deontically) in the realm of X, the promisee: X need not exist qua promisee.

C. A Concluding Caveat

Before concluding this section, one caveat must be lodged. I take myself to be in the process of establishing that, with respect to bare appeal to the formal notion of minimal sufficiency (contained within Bentham’s test), the status of X and Z’s GM do indeed stand and fall together as right-holders. It might seem like I just did something more, namely, to establish that they both fall, on account of the possibility of waiver etc. But I have not established (and nor did I aim to establish) this further thing.

For present purposes, in sparring with Kurki, I have focused on defences of the interest theory starting from the cutting off direction. That is to say, more specifically, defences of the interest theory aimed at cutting off Z’s GM as right-holder by appealing merely to the formal notion of minimal sufficiency. My natural response was to mobilise the exact machinery used in such a defence and to turn it against the interest theory. That is, in section 3, I used the machinery which cut off Z’s GM as a right-holder to also cut off X as a right-holder. In this section, the first natural extension of the debate stimulated by Kurki’s proposal, I closed off at source any effort at resuscitating X as a right-holder (without so resuscitating Z’s GM). Prospects for the interest theory would be bleak indeed if X, the promisee, were unavoidably and ineluctably eliminated as a candidate right-holder. But, I’ve argued no such thing. I’ve simply turned a move in the cutting off direction against itself: Z’s GM and X, the promisee, fall together as right-holders.

This segues nicely to our final extension of the debate. In the second and final extension of the debate, I focus on a defence of the interest theory starting from the other direction—the vindication direction. That is to say, more specifically, a defence of the interest theory aimed at vindicating X, the promisee, as right-holder by appeal merely to the formal notion of minimal sufficiency, without thereby vindicating Z’s GM. My response will be predictable: X and Z’s GM stand together as right-holders.

After all is said and done, X and Z’s GM do indeed stand and fall together as right-holders, at any rate, by the lights of the bare formal strictures of Bentham’s test.

A. Re-re-prising apart Z’s GM and X

In the previous section, I delineated a natural extension of the evaluatively neutral effort to prise apart Z’s GM and X, the promisee, recently voiced by Kurki. Moreover, as noted, the original effort in particular attempts to prise apart Z’s GM and X, the promisee, by cutting off Z’s GM as a right-holder without also cutting off X.

It has now become clear that Kramer’s endorsement of the Kurki strategy in prising apart Z’s GM and X by appeal to the purely structural, or formal, notion of minimal sufficiency, is more steadfast in terms of its overall strategy, than in its precise execution. That is to say, whatever exactly Kramer thinks of the foregoing debate between myself and, in the first instance, Kurki, Kramer has an alternative means of prising apart Z’s GM and X purely by appeal to the evaluatively neutral machinery of minimal sufficiency. This alternative means works in the opposite direction from Kurki’s original effort: it is a vindication strategy. Whereas Kurki attempted to cut off Z’s GM as a right-holder without also cutting off X, Kramer’s alternative means attempts, more positively, to descry X as a right-holder without also descrying Z’s GM as one.Footnote ⁴¹

B. Vindicating X as a Right-holder: Detriment to X, the Promisee, Without Going Through Detriment to Z

When discussing Kurki’s response on behalf of the interest theorist I pointed out that Kurki’s means of carving up the operative facts was perfectly permissible. The clear implication being that there are multiple permissible ways to describe the operative facts. What follows is another such permissible act of carving. Indeed, this is all perfectly in keeping with the notion of minimal sufficiency lying at the heart of Bentham’s test: All the interest theorist needs to vindicate an agent as a right-holder is to specify one minimally sufficient set of facts including detriment to that agent (compatibly with which many other permissible carvings of the facts including detriment to that agent may fail to disclose minimal sufficiency).

Let us then set out Kramer’s latest effort to establish X, the promisee, as an interest theory right-holder (without establishing Z’s GM as such). To fully understand it, we need to begin with a brief detour to Kurki. In my foregoing debate with Kurki, I argued that his way of carving up the facts to cut off Z’s GM as a right-holder could be turned against X, the promisee, to also cut off X. My claim rested on this proposition:

1. (b*) X’s interests as a promisee are set back because of Y’s non-payment of the agreed-upon money to Z.Footnote ⁴²

(It is helpful here, and in what follows, to put portions of the relevant propositions in bold-face.) My claim, then, against Kurki was that his proposition (b), which cut off Z’s GM as a right-holder,Footnote ⁴³ could be mirrored by this cognate proposition (b*), which also problematically cuts off X, the promisee. The shared property of these two propositions ((b) and (b*)) is their specification of detriment to the candidate right-holder which—the bold portion—goes by way of specification of detriment to Z. Non-payment of the money to Z—the bold portion—is minimally sufficient for breach of the contract. Thus, any detriment piggy-backing on something already minimally sufficient, will not itself be minimally sufficient. So much, so straightforward.

The natural—Kramer’s—alternative response on behalf of the interest theorist is to specify detriment to X, the promisee, without going through detriment to Z. As noted, the fact that one specification of detriment to X, viz. (b*), fails to be minimally sufficient, is perfectly compatible with other such specifications which are minimally sufficient. In terms of the overall dialectic, all the interest theorist needs, we have noted, is one serviceable minimally sufficient set involving detriment to X in order to vindicate X’s status as a right-holder. Here is Kramer’s (latest) chosen proposition:

1. (a*) X’s interests as a promisee are set back because of the non-fulfilment of the promise made to X by Y.Footnote ⁴⁴

This is a reprisal, or more detailed specification, of X’s, qua promisee’s, interests as a moral agent suffering a setback in virtue of the non-performance of the promise. Crucially, in (a*) we have no specification or reference to Z’s detriment (indeed, no reference to Z at all). The bold portion of (b*), contrastingly, contained such a reference. Given, we noted, that Z’s detriment is already minimally sufficient for breach of the contract, the detriment to X, the promisee, in (b*) which piggy-backs on Z’s detriment cannot itself be minimally sufficient.

So, things must, for Kramer, be different when it comes to (a*). The bold portion of (a*) cannot itself, for Kramer, disclose a minimally sufficient detriment to an agent for breach. Given that (a*) is being sold as a minimally sufficient condition for breach it cannot be that the following obtains:

1. (a*) X’s interests as a promisee are set back because of (minimally sufficient).Footnote ⁴⁵

It cannot be that Y’s non-fulfilment on the promise made to X is a minimally sufficient set of facts for breach of contract, including a detriment by any agent—X, Z, or otherwise. If, contrariwise, non-fulfilment on the promise were minimally sufficient, (a*) itself could not be minimally sufficient—the very property it is being proposed by Kramer as having. Indeed, on the face of it, a case can be made that Y’s non-fulfilment on the promise made to X does not itself disclose detriment to any agent, X, Z, or otherwise. It is a bare description of an event.Footnote ⁴⁶

The problem for Kramer is that if Y’s non-fulfilment on the promise made to X can be piggy-backed on by X—if it is not minimally sufficient in the required sense, that is—so too can it be piggy-backed on by Z’s GM. We can think of all this in the form of a dilemma for an interest theorist wielding (a*). Either the bold portion is minimally sufficient or it is not. If (horn 1) it is minimally sufficient, then piggy-backing on it via (a*) to vindicate X, the promisee, as a right-holder, is not permissible. On this horn, this vindication response of Kramer’s does not even get off the ground. If, by contrast (horn 2), as Kramer seems to want to be the case, the bold portion is not minimally sufficient, then Z’s GM can also permissibly piggy-back on it. On this horn, X and Z’s GM stand together as right-holders. Either way, it is a severe problem for the interest theorist. It remains to vindicate the claim made in horn 2.

To take stock, this revised vindication strategy of Kramer’s to prise apart X, the promisee, and Z’s GM via (a*) is to forego specifying detriment to Z, as in Kurki’s (a). It foregoes, that is, going by way of, or through specification of detriment to Z. The thought on this revised strategy, is that a similar move cannot be made with respect to Z’s GM. Z’s GM’s setback to interest, so the claim goes, inevitably goes through—is parasitic on—specification of detriment to Z. Contrary to this thought, though, we can now run a parallel premise to (a*) instead involving Z’s GM:

1. (a**) Z’s GM’s interests as a grandmother are set back because of the non-fulfilment of the promise made to X by Y.

Now, a disanalogy with (a*), there is a reference to Z in (a**). But, so what? It is the fact that Z’s GM is a grandmother of Z’s which connects her to this contract. The disanalogous reference to Z in (a**) is merely a reflection of the fact that Z’s GM’s interests are relevant in this context by dint of her blood relationship to Z. By contrast, X’s interests are relevant here in a way specifiable without reference to Z—i.e., X’s status as a promisee. Again, though, so what?

Crucially, although there is reference in (a**) to Z, there is no specification or reference in (a**) to Z’s detriment. To be sure, Z suffers detriment; it is just not specified in (a**). It is only such a specification of Z’s detriment which would threaten (a**)’s status as minimally sufficient for breach of the contract. Put differently, the bold element of (a**) makes no reference to Z’s detriment and, moreover, no reference to Z at all. This bold element is the important bit in establishing (a**)’s status as minimally sufficient for breach. That is, in this context, only insofar as this bold element were to make reference to Z’s detriment, would we have (a**) as redundant, and therefore, not minimally sufficient. It was this aspect of (b) and (b*) which led to their failure to be minimally sufficient. That the non-bold element of (a**) makes reference to Z (albeit not Z’s detriment) is irrelevant. To be sure, invoking Z’s GM entails the existence (at some point) of Z. But, again, so what?

To complete the picture, the bold element of (a**) does indeed make reference to X, the promisee. Moreover, to be sure, the precise reference to X in this portion of (a**) does come very close to disclosing a detriment to X. To return to our dilemma, though, this whole vindication line of Kramer’s response only got off the ground, in construction of (a*), by (implicitly) denying that the bold element common across (a*) and (a**)—viz the non-fulfilment of the promise made to X by Y—disclosed a minimally sufficient condition for breach, including detriment to any agent, X or otherwise.

To conclude, on this (permissible) way of carving up the facts, both X and Z’s GM are right-holders, via (a*) and (a**) respectively. Sreenivasan’s—our—puzzle remains.

To close this section, we can note that even if (a*) does not explicitly specify Z’s interests, as a matter of fact here, X’s, the promisee’s, detriment is, to use Sreenivasan’s term, in some intuitive sense parasitic upon Z’s setback. To further complete this thought, given the content of the promise to X on which Y defaults is the prosaic payment of a sum of money to Z; and given the less prosaic setback to X, qua promisee, as a moral agent, is only to be fully explicated by appeal to the prosaic financial setback to Z, worries, at the very least must remain regarding the interest theorist’s efforts, of any stripe, to prise apart X and Z’s GM. Thus, we have seen that structurally we cannot prise apart X and Z’s GM. To come full circle, the thought is that a non-structural distinction along the lines of the vicarious/non-vicarious distinction is going to need to be wheeled out. Moreover, the worry is that such a distinction is going to have its work cut out in keeping X on the non-vicarious (and Z’s GM on the vicarious) side of that divide. It should be stressed, however, that since my focus in this paper has been on structural efforts to prise apart X and Z’s GM, this closing note of scepticism is far from the last word on non-structural efforts to prise them apart.