How Is the Paradox of Blackmail Paradoxical?

Blackmail presents a perennial problem to economists, lawyers, and philosophers alike. The economic analysis of blackmail is far from conclusive, in that blackmail seems to be efficient in one sense and inefficient in another. Specifically, as noted by Joseph Isenbergh (1993, 1907), “the prohibition of blackmail leads to relatively greater disclosure of a smaller body of information, because on the one hand B is more likely to disclose information about A to C when prevented from bargaining with A, but on the other hand, not being able to extract the full return from bargaining with A, B has less incentive to seek out information about A in the first place.”^[1] This may suggest that permitting blackmail may be only allocatively efficient. That is, once the embarrassing information about the blackmailee has already been discovered, permitting blackmail—rather than banning it—would maximize social utility. For then the interested parties could freely bargain over the information at stake and thereby reach a privately efficient outcome.

However, permitting blackmail would seemingly create perverse incentives to seek out the embarrassing information in the first place. And, crucially, the blackmailer is normally keen to keep the information secret in exchange for money. Thus, the blackmailee’s payment to the blackmailer for silence appears to constitute a mere money transfer, if not worse.^[2] Given this, the decision to permit blackmail might well turn out to be unproductive in the long run (Nozick 1974, 84–87), thereby rendering blackmail dynamically inefficient. A similar point is made by Richard Posner (1986, 630), who argues that even though blackmail involves a threat “to do something completely lawful,” it should still be punishable, as it is plainly socially wasteful. After all, blackmail is just a species of all those threats that the threatening party has no intention of executing “except to maintain credibility.” Or, in Posner’s (1986, 630) words, the blackmailer doesn’t “want to turn you in.” Rather, he “want[s] your money.” So, in the end, typical blackmail activity involves spending economic resources to discover information that the blackmailer is ultimately willing to pay money to suppress. As observed above, this is even worse than a mere transfer of money and therefore cannot by any means increase social utility.^[3]

Still, first and foremost, blackmail presents a moral problem. And it is in the realm of morality that the paradox of blackmail looms large. Since our next section is dedicated to stating precisely what the paradox in question consists in, we will at this point simply lay out what is prima facie paradoxical about blackmail. Normally, for the paradox to get off the ground, we need to assume the blackmailer’s (1) liberty to reveal the blackmailee’s secret and (2) liberty not to reveal the blackmailee’s secret.^[4] Once we assume those two liberties, it appears all the more natural that the blackmailer is permitted to give up his liberty to reveal the secret in exchange for money. But to offer to give up one’s liberty to reveal the victim’s embarrassing secret for money is precisely to issue a blackmail proposal.^[5] Or, to put it differently, what is intuitively paradoxical about blackmail is that it seems permissible to threaten to do what one is independently permitted to do. And yet such threats are impermissible. In fact, they are criminal. Hence the contradiction and the paradox.

This article aims to deepen the sense of the paradoxicality involved in blackmail. The burden of the first section is to formalize the paradox. In particular, we will argue tentatively for a relevant closure principle that permits ending up with contradictory verdicts on the permissibility of blackmail. The second section illuminates a normative question: What counts as a dissolution of the paradox of blackmail? In this section we also heavily rely on the treatment of paradoxes in logic. The third section, on the other hand, illuminates the question of what it takes to solve the paradox of blackmail. In particular, we demonstrate why our previous writings on blackmail should indeed count as a solution to the paradox. Moreover, we suggest an alternative, coercion-based approach which not only remains true to the letter of the paradox but also explains the apparent contradiction it generates.

Formalizing the Paradox

The task of the present section is to try to logically derive the conclusion that blackmail proposals are permissible. This conclusion, coupled with an otherwise known verdict to the effect that blackmail proposals are impermissible, breeds a contradiction. To prove that blackmail proposals are permissible, we will tentatively suggest a disjunctive closure. But before we show the prima facie plausibility of said closure, let us dismiss another closure which bedeviled our previous writings on blackmail.

In Dominiak and Wysocki (2026), we attempted to take the sting out of the paradox of blackmail, submitting that the fact that blackmail is a combination of two otherwise permissible acts does not render blackmail itself permissible. And indeed, there are cases (some of which we cited) clearly showing that this conjunctive closure is implausible.^[6] The classic example is drunk driving. Getting drunk in your car is permissible. Driving your car is permissible too. And yet drunk driving is impermissible. Similarly, sitting in a public park is permissible, and undressing is allowed too. And yet undressing in a public park would count as indecency and could therefore be banned.

Hence, as demonstrated, the assumption of a conjunctive closure would not enable us to derive the conclusion that blackmail proposals are permissible, as the closure itself is indubitably false. And so the contradiction would disappear, together with the paradox. However, here we propose a different principle of closure, the one which prima facie appears not only to be independently plausible but also allows for the derivation of the conclusion. The closure in question is what we shall label disjunctive closure. This closure has it that if it is permissible for agent x to do A and it is permissible for x to do B, then it is permissible for x to do either A or B (disjunctively). Let us represent the above piece of reasoning formally:

Let

1. A be some action performed by agent x;

2. B be some other action performed by agent x;

3. P(A) denote the permissibility of A; and

4. P(B) denote the permissibility of B.

Now, the disjunctive closure states the following principle:

\(\forall A \forall B\, P(A) \wedge P(B) \rightarrow P(A \vee B)\)^[7]

Let us now demonstrate how the disjunctive closure would strengthen the paradoxicality of blackmail; later we will argue that the closure is quite plausible on its face,^[8] especially in comparison to the previously adopted conjunctive closure. That our initial premises (i.e., two liberties on the part of the blackmailer), coupled with the disjunctive closure, allow us to derive the contradiction can indeed be represented formally.

Let

1. action A′ denote agent x’s asking the blackmailee for money, while

2. B′ denotes x’s revealing the blackmailee’s secret. Further, let

3. P(A′) denote the permissibility of action A′ and

4. P(B′) denote the permissibility of action B′.

Given the above, we can derive the contradiction as follows:

1. P(A′) (premise)

2. P(B′) (premise)

3. \(P\left(A^{\prime}\right) \wedge P\left(B^{\prime}\right)\) (from conjunction of 1 and 2)

4. \(\forall A \forall B P(A) \wedge P(B) \rightarrow P(A \vee B)\) (the principle of disjunctive closure)

5. \(P\left(A^{\prime}\right) \wedge P\left(B^{\prime}\right) \rightarrow P\left(A^{\prime} \vee B^{\prime}\right)\) (by universal instantiation from 4)

6. \(P\left(A^{\prime} \vee B^{\prime}\right)\) (by detachment from 3 and 5)

7. \(\neg P\left(A^{\prime} \vee B^{\prime}\right)\) (an otherwise believed proposition)

Now, lines 6 and 7 are contradictory. Moreover, the above formal representation shows the resemblance of the paradox of blackmail to paradoxes in logic. After all, as John Mackie (1973, 238) has it, “A paradox is apparently sound proof of an unacceptable conclusion.” We submit that adopting the disjunctive closure at least tentatively makes sense of the paradoxicality of blackmail. Moreover, and rather crucially, the disjunctive closure, quite unlike what we called the conjunctive closure, seems to be prima facie plausible. The conjunctive closure is immediately problematic, since combining two actions introduces new harms and risks. For example, drunk driving is probably more hazardous than simply adding together the risks of drinking and of driving independently. By contrast, the disjunctive closure is impervious to at least this kind of criticism. For if two actions are independently permissible, then performing either of them (disjunctively) appears to be permissible too. If agent x is permitted both to perform action A and to perform action B, then his performing A or B seems to be permissible a fortiori. Apply this logic to our case of blackmail: The blackmailer by assumption is permitted to (1) gossip about the blackmailee’s secret and (2) ask the blackmailee for money. If so, then there appears to be an even stronger case for (1) or (2) being permissible. It is as if the blackmailer were saying to the blackmailee: “Look, I may gossip about you and I may ask you for money. Now you pick which liberty I should exercise.”^[9]

We believe that the above considerations establish the prima facie plausibility of the disjunctive closure. And if they do, we may venture to logically derive the conclusion that blackmail is permissible, which contradicts the otherwise believed proposition that blackmail is impermissible.^[10] We submit that our reconstruction of the paradox of blackmail (by dint of the disjunctive closure) shows that blackmail presents a genuine paradox rather than a loose puzzle.

What Counts as a Dissolution of the Paradox of Blackmail

In this section, by examining various ways paradoxes are tackled in logic, we will illuminate a normative question: Which treatments of the paradoxes amount to mere dissolutions thereof? These findings will mutatis mutandis apply to various ways in which the paradox of blackmail has been (and can be) tackled. Thus, we hope to end up with a proper standard of dealing with the paradox of blackmail.

Let us start with a brief presentation of the liar paradox, a classical logic paradox, and then proceed to demonstrate how it is dealt with by the Russellian (1908) theory of types. Simply stated, any version of the liar paradox relies on a self-referential utterance (call it U) predicating falsity of itself. In other words, U is “This sentence is false.” Thus, if U is true, then it is false, but if U is false, then it is true—a contradiction.^[11] And thus a paradox. So how does the theory of types deal with the contradiction that U gives rise to? Bertrand Russell’s (1908) idea is to introduce a hierarchy, or types, of expression: Type 0 expressions denote objects; type 1 expressions are first-order predicates predicating certain properties of these objects; type 2 expressions are predicates of first-order predicates; and so forth.

Now, given this framework, the liar paradox simply cannot get off the ground. For when we ask whether U is false, the predicate “is false” that we use in that external question is of a different order from the predicate “is false” as used within U itself. The subscripts will make this clear. Upon analysis, the question would take the following form: Is “This sentence is false₁” false₂? This formulation does not generate a paradox. Hence, even if we were to grant that U is false₂, this would not entail that what U is saying is actually true₁, for there is no uniform predicate “false” at work here. Rather, we only witness “false” shifting between Russellian types. And so, as the Russellian theory would have it, “is false” in the original formulation of the liar paradox is basically used ambiguously. Thus, the liar paradox is allegedly ill-formed and hence no paradox at all.

As we can see, the Russellian theory of types blocks the formulation of the liar paradox: Once orders of “being false” have been postulated, no contradiction arises. Resorting to the theory of types similarly blocks Grelling’s paradox. This paradox kicks off with the realization that adjectives either have or do not have the properties they describe. If they have the property they describe, they are called autological. For example, “short” is autological, as “short” is short itself. If they do not have the property they describe, they are heterological. For example, “long” is heterological, as “long” is not long itself. Now, the troubling question which apparently gives rise to a paradox is this: “Is ‘heterological’ heterological?” For if it is, then it has the property (whatever it is^[12]) that it describes and so is autological, and thus not heterological. And if it is not, then it does not have the property it describes and so is heterological. Hence a contradiction. But again, we can conceive of “being heterological” as belonging to more than one Russellian type. That is, we can conceive of the question “Is ‘heterological’ heterological?” as “Is ‘heterological₁’ heterological₂?” Thus construed, the original question would not generate a contradiction. Hence, the construal in question would block the paradox.

Now, does the type-theoretic reasoning presented above solve the liar paradox and Grelling’s paradox or merely dissolve them? We submit that the theory of types does the latter. The major problem with the type-theoretic attempt at solving said paradoxes is that it seems to be ad hoc in that the alleged hierarchy of linguistic expressions is called forth precisely to explain away the paradoxicality of certain constructions.^[13] And alas, as an ad hoc move, the theory of types can hardly provide a solution to the liar paradox or Grelling’s paradox, as we are given no independent and positive reason to believe in the theory of types itself. Even worse, as noted by Mackie (1973, 254), “Russell had to introduce the Axiom of Reducibility in order to bring back all the innocent victims of the Theory of Types.” In other words, the theory of types not only constituted an unjustified ad hoc invention to solve the Russellian set-theoretic paradox but also resulted in some legitimate constructions being regarded as illegitimate by its standard. Indeed, there seems to be no principled reason why “being false” or “being true” should not be predicated univocally across the orders of expressions. The theory of types provides us with no independent grounds to believe so. Rather, the theory of types fails to take Russell’s (or the liar) paradox seriously. It tries to show that there is no paradox in the first place or that its formulation depends on a misuse of language. Hence, the theory of types dissolves rather than solves the paradox.

How do the above findings carry across to Mitchell Berman’s (1998) treatment of the paradox of blackmail? This author denies to the blackmailer the independent liberty to spread the gossip around. That is, Berman does not share our premise of P(B′). A fortiori, he would not embrace \(P\left(A^{\prime}\right) \wedge P\left(B^{\prime}\right)\). And so the antecedent of the disjunctive closure would not be satisfied. Hence, finally, he would not be able to logically derive \(P\left(A^{\prime} \vee B^{\prime}\right)\), which we equated to the permissibility of the blackmail proposal. We submit that Berman’s attempt constitutes a dissolution rather than a solution of the paradox of blackmail, as it—rather similarly to the type-theoretic attempt to solve Russell’s paradox—does not allow the paradox of blackmail to even get off the ground. Recall that the paradox of blackmail crucially relies on the blackmailer’s liberty to spread the embarrassing information around. Denying this liberty to the blackmailer indeed blocks the permissibility of the blackmail proposal—the proposal which is contradicted by the otherwise believed proposition that blackmail proposals are impermissible. But what is particularly unattractive about denying the blackmailer the liberty to gossip is that it is, quite like the introduction of the theory of types as an attempt to solve Russell’s paradox, an unmotivated, ad hoc move. After all, granting the blackmailer’s liberty to gossip—in the absence of relevant contracts—is quite rightly a default position. Assuming away relevant contracts, the blackmailer clearly does not owe silence to the blackmailee. Therefore, we are owed an independent, positive reason to believe that it would be impermissible for the blackmailer to spread the gossip around; a flat denial of the blackmailer’s liberty to gossip will not do.^[14] Thus, Berman’s (1998) attempt, we believe, also counts as a dissolution rather than as a solution of the paradox of blackmail.

On the other end of the spectrum, we have Block’s (2013) treatment of the paradox of blackmail.^[15] This author—quite unlike Berman (1998)^[16]—takes the premises of the paradox most seriously. That is, he readily grants that the blackmailer enjoys the critical two liberties (1) to ask the blackmailee for money and (2) to reveal the latter’s secret. Moreover, whatever sort of closure Block actually adopts, let us construe what he submits in the most charitable way. Says this author, “It is impossible for two separately legal acts to be rendered into an illegal one, and, therefore, blackmail should be legalized” (Block 2013, iii).^[17] This is easily compatible with our suggested disjunctive closure—as we saw, the most plausible principle of closure. Block would certainly agree that if it is permissible for the blackmailer to perform action A and permissible for him to perform action B, then presenting the blackmailee with a disjunction—“I will perform either A or B”—must be permissible too. Recall that we can derive the permissibility of blackmail from those three premises:

1. P(A)

2. P(B)

3. \(P(A) \wedge P(B)\) (from 1 and 2 by conjunction introduction)

4. \(\forall A \forall B P(A) \wedge P(B) \rightarrow P(A \vee B)\) (the principle of disjunctive closure)

5. \(P\left(A^{\prime}\right) \wedge P\left(B^{\prime}\right) \rightarrow P\left(A^{\prime} \vee B^{\prime}\right)\) (by universal instantiation from 4)

6. \(P\left(A^{\prime} \vee B^{\prime}\right)\) (by detachment from 3 and 5)

But now, Block denies \(\neg P(A \vee B)\) and thus blocks the paradox. Hence, rather trivially, no contradiction arises. In other words, Block is essentially saying that the above reasoning is correct and thus he recommends that \(\neg P(A \vee B)\) should go. But then again, nobody doubts the validity of the above reasoning. Granted, given premises 1–4, we can logically derive \(P(A \vee B)\). But this goes no distance toward explaining why the majority of theorists and prevailing legal practice converge on the impermissibility of blackmail. That is to say, Block fails to address the genuine tension between two independently motivated propositions: (1) \(P(A \vee B)\) and (2) \(\neg P(A \vee B)\). Thus, his position does not solve the paradox. Rather, it turns a blind eye to it in that he simply discards \(\neg P(A \vee B)\), which is, after all, an independently plausible premise. Block takes no pains to explain its plausibility. Hence, this author does not live with the paradox. He merely dismisses one of its prongs.

Given the above, let us look for an analogue of Block’s position on the paradox of blackmail in the realm of various treatments of paradoxes in logic. We posit that Block’s shrugging off \(\neg P(A \vee B)\) closely resembles dropping the requirement of bivalence when faced with the liar paradox. It is as though, when confronted with the contradiction, one were to hold that the underlying assumption “every proposition is either true or false” must be dropped, and to leave it at that. In this manner, the paradox would basically be discarded not by explaining why the tension arises in the first place, but by dropping one of the independently attractive principles that generated it. So, just as Block simply denies that \(\neg P(A \vee B)\), one might take the sting out of the emerging contradiction by denying the principle of bivalence. Once that move has been made, the contradiction would no longer arise, as certain sentences would simply be neither true nor false. Hence, both Block’s handling of the paradox of blackmail and the treatment of the liar paradox under consideration would make the paradoxicality disappear, together with one of its independently motivated prongs. This is why we count Block’s position as a dissolution of the paradox of blackmail.

Finally, it is worth noting that there is another possible way of tackling the paradox of blackmail which would fall on the side of dissolving rather than solving it. This would involve biting the bullet and embracing the resulting contradiction. This move would be on a par with accepting dialetheism. The dialetheist, in contrast to the position considered just above, would happily live with the contradiction. For the dialetheist, the liar paradox poses no logical problem. After all, belief in true contradictions is the gist of dialetheism. But simply embracing the contradictory conclusion stemming from the liar paradox is deflating the paradox. The liar paradox is problematic precisely for the reason that it generates a contradiction. Therefore, merely accepting the arising contradiction goes no distance toward solving the paradox.^[18]

Simply accepting the contradictory verdict of the paradox of blackmail would be equally unsatisfactory. The contradictory propositions (1) \(P(A \vee B)\) and (2) \(\neg P(A \vee B)\) crave an explanation. Hence, merely to embrace (1) and (2) without any systematic argument is clearly to dissolve the paradox.

Having, we believe, established what would count as dissolving the paradox of blackmail, let us now turn to the question of what it takes to solve this paradox.

What It Takes to Solve Rather Than Dissolve the Paradox of Blackmail

As we already hinted above, a necessary condition for a solution of the paradox of blackmail is to take its core premises (i.e., the two liberties enjoyed by the blackmailer) seriously. Denying them would be too high a price to pay for dispelling the paradox. After all, we are independently warranted in believing the blackmailer is indeed permitted to reveal the embarrassing truth about the blackmailee just as he is allowed to ask the latter for money. Furthermore, we are also obliged to take heed of and not to dismiss the contradictory verdicts: (1) \(P(A \vee B)\) and (2) \(\neg P(A \vee B)\). In particular, it seems that, when solving the paradox, we are supposed to explain the resulting (apparent) contradiction. Recall that merely embracing the contradiction does nothing to solve the paradox. Unless we uncover a reason why we should accept the contradictory conclusion, accepting it is merely a way of dissolving the paradox.

To illustrate what would count as a solution of the paradox of blackmail, let us revert to our earlier proposal (Dominiak and Wysocki 2026). Revisiting it makes a lot of sense, since it is clearer in the light of the normative standard for solving paradoxes why our previous attempt indeed passes muster. In a nutshell, our idea was to introduce into the matter another morally relevant property of proposals and exchanges: fraudulence/nonfraudulence of blackmail proposals (and blackmail exchanges). But how specifically did our alleged solution unfold?

First, we did grant the two liberties to the blackmailer. As has been stressed in the course of the present article, the belief that the blackmailer is both permitted to gossip and permitted to ask the blackmailee for money is utterly warranted. We also showed, crucially, how we can make sense of \(\neg P(A \vee B)\).^[19] Finally, and again crucially, we demonstrated how we can ground \(P(A \vee B)\). All in all, we took the whole setup of the paradox for granted. Our attempt at solving the paradox was not an ad hoc move at all. After all, fraudulence is a well-recognized morally relevant property of proposals and exchanges. A fraudulent character of an exchange would render it involuntary. Fraud is also criminalized, and so fraudsters are deserving of punishment. Given this, our invocation of fraudulence is not merely designed to solve the paradox of blackmail. Thus, if fraudulence or the lack thereof can distinguish between impermissible and permissible blackmail proposals and exchanges, respectively, this would count as a solution of the paradox of blackmail.

And indeed, appealing to the distinction between fraudulence and nonfraudulence turned out to be up to the task. That is, there are two distinct psychological profiles of blackmailer. Some bluffing blackmailers intend not to reveal the embarrassing secret when not paid. Bluffers might also simply not intend to disclose the blackmailee’s secret in case of the latter’s noncompliance. If so, then they deceive blackmailees into paying. The latter pay to avoid an outcome (i.e., the revelation of the embarrassing information) that would not otherwise obtain. And that is why they are effectively defrauded. Hence, even if it is independently permissible to gossip and independently permissible to ask the blackmailee for money, the proposal \(A' \vee B'\) is not necessarily permissible. The blackmailer’s intention makes all the difference. The proposal \(A' \vee B'\) might be insincere in that the blackmailer intends not to perform B′ or does not intend to perform B′ (i.e., to reveal the secret) if not paid. If, on the other hand, a nonbluffing blackmailer were to sincerely propose \(A' \vee B'\), this proposal would indeed be permissible. In other words, if the blackmailer were genuinely intent upon B′ when not paid, there would be nothing morally wrong with the whole proposal \(A' \vee B'\).^[20]

One remark is due at this point. It is worth noting that our solution casts doubt upon the disjunctive closure. It now transpires that this closure is a sweeping and unjustified generalization. However, this does not mean that we are manipulating the assumptions of the paradox. Remember, the disjunctive closure was a tentative construction designed to strengthen the sense of paradoxicality of blackmail. Fraudulence is a well-established, morally significant property of exchanges. If fraudulence of certain proposals undermines the universal applicability of the disjunctive closure, then so much the worse for the closure.

Furthermore, it is worthwhile to obviate the possible rejoinder that our proposed solution is not true to the letter of the paradox, as the contradiction in the end disappears, for in its final analysis, any particular blackmail proposal or exchange is either permissible or impermissible. Hence, no contradiction arises, as it is never the case that there is a blackmail proposal or exchange that is both permissible and impermissible. But the demand that we solve the paradox and live with the contradiction at the same time is impossible. Remember, it was precisely the arising of the contradiction that made the paradox problematic and crying out for explanation. This is the problem with all paradoxes: Sound reasoning leads to a contradiction, and the latter must somehow be resolved. Our solution indeed resolves this tension: It shows that the contradiction is merely apparent, for once we consider any particular proposal, it clearly falls either on the side of permissibility or on the side of impermissibility.

Before we conclude, it is time for a concession. We do not purport to have come up with a unique solution to the paradox of blackmail. We believe that any attempt at solving the paradox which alludes to any morally significant property of blackmail proposals (and exchanges) such that some blackmail proposals instantiate that property and others do not would count as equally satisfactory. Our solution employs the distinction between fraudulence and nonfraudulence, clearly a distinction carrying a normative force. Blackmail has often been probed as an instance of coercive proposals, but then again, if all blackmail proposals turned out to be coercive, we could not make sense of the contradiction stemming from the paradox, for the prima facie plausible \(P(A \vee B)\) would then make no sense. Similarly, if, as insisted by Block (2013), no blackmail proposal turned out to be coercive, we would again be unable to explain the paradoxicality of blackmail, for then we could not make a case for the prima facie plausible \(\neg P(A \vee B)\). And yet if coerciveness turned out to be a discriminatory feature in the universe of blackmail proposals, appealing to thus-understood coerciveness would also solve the paradox of blackmail.

It seems that several accounts of coercion might serve this purpose. One such account is Grant Lamond’s (1996; 2000). To put it briefly, if somewhat simplistically, whether Lamond’s theory would classify a given proposal as coercive or not is mainly a function of whether the proposer promises one of the disjuncts of \(A \vee B\) only because he knows that it is unwelcome to the recipient and he knows that the latter would rather succumb to the proposer than let him execute either A or B, depending on which is a threatened action. To be sure, the majority of blackmail proposals meet this condition. And yet we can easily think of some blackmail proposals which do not satisfy this Lamondian necessary condition of coercion. After all, the blackmailee might happen to be indifferent between the disclosure and his paying for the blackmailer’s silence, so it would then not be the case that the blackmailee would find the disclosure more unwelcome than his paying for silence. With the blackmailee disposed in this way, the blackmailer cannot issue a coercive threat by the Lamondian standard. And so, if the blackmailee decides to pay (simply out of indifference), the transaction would not be coerced. Hence, we would end up with a similar verdict to ours: Some blackmail exchanges are permissible, and some are not.^[21] This, as already established, would address the paradox of blackmail equally satisfactorily.

Another account which would satisfactorily engage with the paradoxicality of blackmail is Scott Anderson’s (2008) nonmoralized coercer-based theory of coercion. Although Anderson (2008) does not explicitly deal with blackmail, it is fair to say that his account of coercion easily predicts that some blackmail proposals would be coercive and some not.^[22] We can reach this conclusion via Anderson’s (2008; 2011) treatment of bluff threats coupled with his preferred nonmoralized concept of coercion. Anderson (2008) posits that the central cases of coercion can be explained by a power differential between the coercer and the coercee. Clearly, the gunman’s proposal “Your money or your life”—a paradigm example of a coercive threat—is predicated on the gunman’s being in a position to inflict serious harm on his victim and on the victim’s vulnerability and inability to retaliate. However, said power differential does not need to actually obtain, according to Anderson (2008, 413), who submits that “coercion by bluff threats is often possible, and this depends not on the coercer’s own powers, but the powers and past actions of those he imitates.” We can infer from Anderson’s statement that he would find some bluffing blackmail proposals coercive and thus impermissible, and some noncoercive and thus permissible.^[23] Granted, our account of blackmail would regard all bluffing blackmail proposals as impermissible (Dominiak and Wysocki 2026), whereas Anderson’s (2008) would find only some of them impermissible. But that is irrelevant. What matters here is that both our account and Anderson’s can successfully engage with the paradoxicality of blackmail.^[24]

By contrast, it appears that Block’s (1986) moralized account of coercion (which at the same time turns a blind eye to the possibility of blackmail proposals being fraudulent) fails to make sense of why blackmail is paradoxical. After all, his moralized theory assesses the morality of the proposal solely as a function of the morality of a threatened act.^[25] Given this, since blackmail by definition involves a threat to do something lawful, Block must deem any blackmail proposal lawful too. And this is precisely the reason why Block (2013) rejects \(\neg P(A \vee B)\). Less technically, that is the reason why he denies the impermissibility of blackmail and calls for its decriminalization.^[26]

Conclusion

In this article, we tried to illuminate which attempts to solve the paradox of blackmail would count as dissolutions of it and which as solutions. Thus, a normative standard, informed by the treatment of paradoxes in logic, emerged. We argued that to solve the paradox of blackmail we must take its core commitments for granted. Thus, we granted two liberties to the blackmailer: (1) the liberty to gossip and (2) the liberty to ask the blackmailee for money. We also took seriously the prima facie plausibility of both the permissibility and impermissibility of blackmail. To strengthen the sense of the paradoxicality of blackmail we resorted to a seemingly justifiable principle of disjunctive closure.

Then we proceeded to show that our previous solution to the paradox of blackmail remains unscathed, even given a more sharply formulated standard for solving it. Finally, we conceded that ours is by no means a unique solution to the paradox of blackmail. Having said that, we outlined a possible solution to the paradox via coercion rather than fraud.

We hope that this article will help open a discussion on standards for resolving paradoxes in the social sciences. Our treatment of the paradox of blackmail may serve as a promising point of departure.