The Morality of Treason

Treason is one of the most serious legal offences that there are, in most if not all jurisdictions. Laws against treason are rooted in deep-seated moral revulsion about acts which, in the political realm, are paradigmatic examples of breaches of loyalty. Yet, it is not altogether clear what treason consists in: someone’s traitor is often another’s loyalist. In this paper, my aim is twofold: to offer a plausible conceptual account of treason, and to partly rehabilitate traitors. I focus on informational treason, as the act of passing secret intelligence to foreign actors without authorization. I argue that informational treason is sometimes justified, indeed morally mandatory; even when it is morally wrong, its beneficiaries are sometimes justified, indeed obliged, to make use of the intelligence thereby provided.


I. INTRODUCTION
Kim Philby, Oleg Penkovksy, Oleg Gordievsky, and Aldrich Ames are some of the best known spies of the Cold War. They all worked for their respective country's intelligence services and at the same time passed on its secrets to the other side. Philby and Penkovsky are both dead. At the time of writing this, Gordievsky and Ames are still alive. Gordievsky, who ended his career as deputy head of the KGB station in London, worked for MI6 from the 1970s until he escaped from Moscow with MI6's help in 1985. He currently lives at an undisclosed location in England under the protection of the British security services. Ames, who was a CIA officer with particular expertise in Soviet counter-intelligence, was arrested in 1994 and is currently serving a life sentence in Indiana. 1 All four have been widely described and often condemned as traitors. They are far from alone in this. Not only is the history of international relations writ large littered with such cases: not a year goes by without news that some official, somewhere, has been arrested for passing on official secrets, without authorisation, to a foreign power.
Treason is one of the most serious criminal offences that there are. In the UK, a conviction for high treason carries a life-long prison sentence, while breaching the Official Secrets Act carries a prison sentence of up to 2 years. In the US, a conviction for treasonous espionage carries a minimum of 5 years in jail up to the death penalty. At least, in those two countries, those who are suspected of treason have a right to a fair trial. Had Gordievsky been captured, he would in all likelihood have been tortured and summarily executeda fate that befell Penkovsky. 2 Laws against treason are rooted in deep seated moral revulsion about acts which, in the political realm, are paradigmatic examples of breaches of loyalty. The traitor deceives us, breaches promises -implicit or explicit -he has made to remain loyal, and leads us to doubt in the trustworthiness and dependability of our fellow citizens and public officials. At its most extreme, his treachery goes to the heart of what binds us together as members of the same political community.
And yet, surprisingly, although there is a plethora of philosophical works on political obligation and (more generally) on the obligations of citizenship on the one hand, and a sizeable empirical literature on treason on the other hand, contemporary moral and political philosophy has been somewhat silent on this particular issue. My aim, in this paper, is to provide an account of what the charge of treason amounts to, and to (partially) rehabilitate those against whom it might be levelled. 3 I focus on what we may call informational treason -the act of disclosing official secrets without authorisation to a foreign power. I thus set aside cases in which traitors plot to overthrow their leaders. Likewise, I do not attend to cases in which individuals join foreign forces and fight against their own community, such as (in recent years) Britons who enlisted with ISIS in the Middle East. Informational treason is more common than lethal treason. In some respects, it is also more interesting. Granted, if one believes that killing members of one's own side is morally right under some circumstances, one is likely to believe that passing on intelligence to the other side is morally right a fortiori under similar circumstances. However, many would resist the conditional. I want to show, thus, that even if lethal treason is morally impermissible, informational treason -which need not lead or contribute to acts of killing -is sometimes morally justified. In addition, I restrict my inquiry to informational treason for the sake of foreign actors, and thus do not look at cases in which an agent passes on information to a domestic revolutionary movement, or blows the whistle on her country's wrongful practices for the benefit of the world at large. In some respects, those acts are easier to justify than informational treason to foreign actors. A domestic informational traitor (as it were) might find it easier to show that she is acting for the sake of fellow community members by helping undermine unjust institutions from the inside, than if she did so from the outside. As for the whistleblower, even though she prepares for her act of disclosure in secret, the act itself is fully public: that, in fact, is precisely the point. By contrast, Philby, Penkovsky, Gordievsky and Ames disclosed their country's secrets surreptitiously, often for years. In so doing, not only did they deceive their colleagues, friends, and fellow citizens: by keeping them in the dark, they gave a geopolitical advantage to the beneficiaries of their treason without giving their country the opportunity to counter-act. Whatever one might say of Snowden and Ellsberg, at least their respective administrations found out alongside everyone else (more or less) what they were up to. 4 I proceed as follows. In Section II, I provide an account of the accusation of treason -of what a treasonous act consists in, and why it is presumptively wrongful. In Sections III and IV, I argue that agents are sometimes morally permitted, indeed obliged, treasonously to pass on secret intelligence to foreign enemies as a means to thwart violations of fundamental rights. In Section V, I tackle the delicate issue of treason between allies. In Section VI, I argue that even when treason is morally wrong, its beneficiaries are sometimes justified, indeed obliged, to make use of the intelligence thereby provided. Section VII concludes.
Three final points. First, I take it for granted that all human beings, wherever they reside in the world, have rights to the freedoms and resources they need in order to lead a minimally flourishing life -against all other human beings. Put differently, I take it for granted that national-cum-political borders are morally irrelevant to human beings' fundamental rights. This is merely another way to assert universal human rights and the correlative duties not to harm, but to protect, fellow human beings' vital interests. It is also compatible with the view that there are wrongdoings other than violations of fundamental rights -lesser wrongdoings for sure, but wrongdoings nonetheless. 5 Second, I shall have very little to say about the issue of traitors' motives. These range widely, singly or in combination, from commitment to an ideology, hatred of one's regime, desire to rescue one's compatriots from that regime, financial greed, or the simple 4 There is a small but growing literature on the ethics of whisteblowing. See e.g. Sissela Bok, Secrets: On the ethics of concealment and revelation, 2nd ed. 5 I call those rights 'fundamental rights', rather than human rights, for there is no reason to suppose that the grounds for conferring those rights (sentience, the capacity for moral and rational agency, etc.) are necessarily, in all possible worlds, restricted to human beings. need to survive. As we shall see, the traitor's motives are not relevant to the characterisation of his acts as treasonous. Whether they are relevant to the moral worth of such acts is too complex an issue to be tackled here. Accordingly, my claim that treason is (sometimes) morally justified holds subject to whether motives are relevant and, if they are, whether the traitor acts from the right motives. 6 Third, by 'treason' I do not mean the crime of treason as it is defined in the criminal codes of various jurisdiction. Nor, when I say that a treasonous act is wrong, do I imply that it ought to be turned into a criminal offence. Whether it ought to be so is a separate question which space prevents me from addressing here. To the extent that this paper makes a contribution to understanding the law, it does so by providing an account of the morality of treason tout court-that is to say, of the foundations of a morally justified legal approach to treason. 7 II. UNDERSTANDING TREASON Treason manifests itself in many guises (from political assassination to informing on one's regime, from unauthorised diplomatic negotiations to capitulating to the enemy -to name but a few examples.) Moreover, it is not always clear what and whom the alleged traitor is betraying: one person's traitor is more often than not another person's loyalist. Treason, in other words, is a contested concept. 8 Nevertheless, to describe someone as a traitor implies that he breaches presumptive obligations of loyalty to fellow members of his political community. This, in turn, supposes that he stands in a  (Boulder, Co.: Westview Press, 2001). As K. Brownlee pointed out to me, there is a sense in which treason is whatever the law of the land says it is and punishes accordingly. However, I want to allow for the possibility that the law might get it wrong by deeming as treasonous acts which are not, and by failing to describe as treasonous acts which are. particular relationship to those individuals. As Philby himself once put it, 'To betray, you must first belong. I never belonged.' 9 The question, thus, is what kind of relationship qualifies for treason and what obligations, as grounded in that relationship, the traitor presumptively breaches.

A. Treason, Nationality and Membership
At first sight, we might think that the relationship of co-nationality (by which I mean purely formal citizenship) is both a necessary and (together with the breach of relevant obligations) a sufficient condition for an agent to be aptly described as a traitor. My opening examples seemingly illustrate the claim that it is sufficient: Philby, Penkovsky, Gordievsky and Ames were/are nationals of the country whose secrets they divulged, and that alone (we might think) qualifies them as traitors. Furthermore, to see the force of the claim that nationality is a necessary condition, consider the case of Markus Wolff, the long-serving head of the GDR's foreign intelligence service (or HVA) -itself part of the infamous Stasi. After the reunification of Germany, Wolff was convicted of treason against the Federal German State. The Federal Constitutional Court subsequently over-turned the verdict, on the ground that a national of a state duly acknowledged under international law (albeit defunct by the time of the trial) cannot be accused of treason against another state. 10 Upon closer inspection, however, nationality seems a poor basis for the charge of treason. On the one hand, it labels as traitors individuals who do not seem to act treasonously. Suppose that while travelling through Russia on an American passport, Natasha stumbles upon the abandonned briefcase of a Russian nuclear scientist, and decides to pass on its content to the American Embassy. Unbeknownst to her, however, she is not American: she is in fact a Russian national. Yet, it would seem odd for the Russian authorities to regard her as a traitor. Not only does she have no substantive relationship to the Russian state and its citizens: she is not even aware of her purely formal relationship to them. At the very least, we would need to add a knowledge requirement to the nationality condition, but even so, being a national and aware of it is not enough. For suppose now that Natasha knows that she is Russian, but has lived in the US for over 50 years without ever coming back to Russia, and only has a formal connection, by dint of her citizenship, with the latter. Were she to stumble upon Russian secrets while residing in the US and pass on those secrets to (e.g.) the Pentagon, she would not be aptly regarded as a traitor to Russia either.
On the other hand, nationality does not account for cases in which the charge of treason seems apt, as the example of William Joyce nicely illustrates. Joyce, a.k.a Lord Haw-Haw, was the last person to be executed for treason in the United Kingdom. He was an American citizen of Irish descent, who identified as British and who, thanks to his father's fraudulent application to the British authorities, was in possession of a British passport. Joyce joined the British Union of Fascists in the 1930s, fled to Germany in 1939 and was naturalised as German in 1940. There, he was hired by the German Ministry of Propaganda to work in the English-language section of the radio services, from whence he issued virulenty anti-British broadcasts until his capture by the Allies in 1944. His nationality was a key issue at his trial. At first, the prosecution argued that, as a British subject, he owed a duty of allegiance to the Crown which he had breached by working for Britain's enemy. The trial collapsed as soon as the defence team was able to prove that Joyce was not, in fact, a British subject. The prosecution returned to the breach with a new charge. As the holder of a British passport (albeit one fraudulently obtained), Joyce had enjoyed the de facto protection of the British Crown; as such he was under a legal duty of allegiance to the latter, which he had violated by fleeing to and working for the enemy. The House of Lords ultimately found in favour of the Crown and Joyce was hanged in May 1946. 11 On the prosecution's view, in other words, nationality was not a necessary condition for treason: enjoying the de facto protection of the Crown was sufficient. Whether or not the prosecution was right on point of law, as a matter of morality it was on the right track with respect to nationality: intuitively, it seems that Joyce was a traitor, notwithstanding the fact that he was not British.
Nationality, thus, is both under-and over-inclusive. The question, then, is what kind of relationship an agent must have with a political community and its members in order to be aptly regarded as a traitor to it. In the light of Joyce's case, we might perhaps think that enjoying the protection of the state, irrespective of one's nationality, is an apt basis for the charge of treason. We would be mistaken. For if merely enjoying the protection of the relevant authorities on whose territory one happens to be were enough, a tourist who does enjoy such protection could aptly be regarded as a traitor, which seems counter-intuitive. 12 Moreover, long-term protection does not seem to be enough either. Suppose that James is an African-American citizen living in Mississipi in the 1950s. Although he enjoys consular protection while abroad by dint of holding a US passport, at home he is subject to mandatory segregation laws, is systematically discriminated against on the job and housing markets, has to endure daily manifestations of racism by other American citizens, and so on. In other words: he is treated as a second-class citizen and thus does not enjoy social membership in the United States. Suppose that he stumbles upon a briefcase full of state secrets, and finds a way to pass it on to the Soviet Embassy in Washington. Assuming for the sake of argument that he would act wrongly, it is not clear that he would aptly be regarded as a traitor to his community, in the light of the treatment which the latter's authorities, with the support (be it passive or active) of many of its members, mete out to him. 13 Contrast with Jake, also an American citizen, who in every possible way is treated as a first-class citizen. Were he to 12 For a very interesting discussion of the ways the protection of the British Crown extends to foreigners, see John Finnis, 'Nationality, alienage and constitutional principle', Law Quarterly Review 123 (2007): 417-445. I am grateful to Richard Ekins for the suggestion and for pressing me on the issue. 13 As an anonymous reviewer pointed out, there is a risk that second-class citizens might object that denying them the possibility of committing treason is tantamount to denying them the status of full members, notwithstanding the fact that they benefit from being exempt from the charge of treason. (Similarly, the reviewer notes, someone who is exempt from punishment on the grounds that she is declared insane might object to being stigmatized by the label of insanity though she benefits from it.) I can see the force of the concern. In reply: the mere fact that an agent objects to being characterised in a particular way and treated accordingly is not in itself a decisive reason to rescind that characterisation. Moreover, I wonder whether the second-class citizen's resentment might not be directed, and justifiably so, at those who subject him to grievous injustice yet at the same time refuse to call him a traitor precisely because in their eyes he does not fully belong -rather than at those who also refuse so to call him in condemnatory acknowledgment of that injustice. provide a foreign power with US state secrets, he clearly would be a much stronger candidate for the charge of treason. Compare, finally, with Carl, who is not an American citizen but who is a long-term resident in the US and enjoys all the considerable economic and social benefits of such residence, minus those inherent in formal citizenship (of his volition, as he has decided not to apply for naturalisation.) Pending further consideration, he too, I submit, would aptly be regarded as a traitor under those circumstances.
The examples of Jake and Carl, as contrasted with the example of James, suggest that social membership is a necessary condition for a relationship to count as treason-qualifying. Carl and Jake enjoy it, whereas James does not. The 'social membership thesis' thus reaches the correct conclusion in William Joyce's case. That being said, it needs refining. Return to Oleg Gordievsky. As a citizen of a totalitarian regime, his fundamental moral rights to the freedoms and resources needed for a minimally flourishing life were not enforced in law. As a matter of fact, however, he was enjoying considerably more freedom and had access to a far greater range of material resources than ordinary Soviet citizens, thanks to his institutional position within the KGB -dominated freedom and access, in so far as these could be taken away from him at a moment's notice without due process, but freedom and access nonetheless. Gordievsky's case suggests that we should distinguish between thick and thin social membership. Someone enjoys thick social membership if and only if (a) her fundamental moral rights are enshrined in her community's legal system, such as to give her meaningful recourse should those rights be violated, and (b) those rights are not, as a matter of fact, routinely violated. Someone enjoys thin social membership if and only if her fundamental moral rights are not, in fact, routinely violated, whatever the law of her community might say. Now, it seems apt to describe Gordievsky as a traitor -and, more generally, so to describe members of illiberal and/or undemocratic communities who nevertheless enjoy an array of de facto freedoms and benefits. By contrast, it does not seem apt so to describe someone (like James) whose fundamental rights are routinely violated de facto and de jure. Taken together, those two points suggest that enjoying thin social membership can be an appropriate basis for treason. A fortiori, so is thick social membership. 14 However, although social membership is necessary for a relationship to count as treason-qualifying, it is not sufficient, for two reasons. First, in keeping with my earlier claim that someone who is not aware that she is a national of a particular community cannot be aptly described as a traitor, someone who is not aware of the fact that she is a social member in that community cannot be so described either. Whilst it is hard to see, in practice, how one could fail to know that one is being so treated, so long as it is a theoretical possibility, there is no reason not to apply a knowledge requirement to the social membership thesis.
Second, consider Soviet 'illegals', who were inserted by the Soviets under fake identities (complete with fake passsports, fake jobs, etc.) in Western countries and lived there, sometimes for years. Two of the best known illegals are William Fisher, also known as Rudolf Abel, and Konon Melody, also known as Gordon Lonsdale. Fisher was a British citizen by birth, but fled with his family to the USSR in the 1920s. By a long circuitous route, he joined the KGB and was trained to work and live as an illegal in the US -which he did for 8 years until his arrest in 1957. Molody was a Soviet citizen, and lived undercover in the UK for years while working for the KGB. Both Fisher and Molody enjoyed all the benefits of long-term residence in the US and the UK respectively. If social membership is a sufficient condition for a relationship to count as treason-qualifying, those agents were guilty of treason to their country of residence. And yet, given that they all along were KGB agents, inserted into those countries with the explicit purpose of spying on them, it does not seem quite right to describe them as traitors. As a matter of fact, they were charged with and convicted of espionage -not treason. This strikes me as intuitively plausible. The relevant difference between the illegals on the one hand, and individuals such as Carl or Jake on the other hand, is that the former became social members in order to act against Western communities and their members. Not only (to the best of our knowledge) did they not think of themselves as full members of those communities: it was essential to their work that they not identify with it. Put differently, they were not bona fide members. Jake and Carl, by contrast, are bona fide members. It is that, I suggest, which would mark them as traitors were they to pass on official secrets to third parties without being authorised to do so. 15 Some readers might remain unconvinced. They might agree that illegals do not owe an obligation of loyalty to the regime against whom they have been sent to act by their own regime; but in so far as they have benefitted from their social membership in that foreign community, albeit under false pretence, they owe some obligation of loyalty to that community's members. On this view, bona fide social membership is a necessary condition for being charged with treason against the regime (and so illegals are not traitors in this sense); but it is not a necessary condition for being charged with treason against ordinary citizens (and so illegals are traitors in that sense.) I see the force of the objection, particularly in those cases where although it makes sense to say that the illegals' regime, or state, is in conflict with the regime or state of the community on which they have to spy, it does not really make sense to say that both citizenries are in conflict against each other. (That said, if this is correct, the objection is not particularly strong in the case of Soviet illegals, given the wholesale mobilisation of material and human resources by both sides against the other.) Even so, there does seem to be an important difference between illegals and bona fide members, which might perhaps be captured in the following way. When Philby was unmasked, it was entirely appropriate of his colleagues and fellow citizens to say: 'But he is one of us!' When Fisher was unmasked, it would have been entirely appropriate of those whom he had deceived to say: 'But I thought he was one of us!' Suppose that I am wrong. If so, we are still left with the view that (to repeat) bona fide social membership is a necessary condition for charging wrongdoers with treason against their regime, but not for charging them with treason against ordinary citizens. Either way, it is social membership that matters: not nationality or the fact that one enjoys the state's protection.

B. The Presumptive Wrongfulness of Treason
As I noted at the outset of s.II, to describe someone as a traitor implies that he breaches presumptive obligations of loyalty to fellow members of his political community -which in turn presupposes that he stands in a particular relationship to the latter. I have just given an account of that relationship. I now offer a hybrid, threepronged argument in support of the claim that to commit treason is, presumptively, to commit a wrongdoing -such that the burden of proof lies on the shoulders of those who wish to defend it. (In this respect, it is similar to deception.) I shall use examples of unauthorised disclosure to illustrate those arguments, but the latter apply mutatis mutandis to other forms of treason.
One may wonder what is distinctive about the claim that treason is presumptively wrongful. After all, one might think that the task of justifying the claim that treason is presumptively wrongful is tantamount to justifying the claim that individuals are under a political obligation to obey the law of the state in which they enjoy social membership; or that it is tantamount to justifying the claim that an agent may not be complicitous in wrongful ends. If those points are true, there is nothing special about treason.
I do not think that this is quite right. Even non-residents, such as tourists and travellers passing through, are under an obligation to obey the law; yet, as I noted above, it would not be apt to describe them as traitors were they in breach of the relevant laws. Moreover, everyone is under a duty not to further wrongful ends. The traitor's breach, however, is qualitatively different. Even if there is, generally, an obligation to obey the law and/or an obligation not to further wrongful ends, both of which the traitor might be breaching, the charge of treason gives his presumptive dereliction a distinctive hue which needs explaining. 16 The first argument for the presumptive wrongfulness of treason (as distinct from the presumptive wrongfulness of disobedience to the law) appeals to trust. The traitor, it is said, impairs the political trust which other community members have in him qua fellow member -and therein lies his wrongful breach of loyalty. As a rough cut, two agents display political trust towards one another not merely if they rely on one another to act in particular ways in their political capacity, but also (more strongly) if they both regard the fact of that reliance as a reason so to act and they both know that the other is so motivated. Suppose that Denis is a civil servant in political community Blue and in that capacity is entrusted by his superiors, who in some sense act on behalf of Blue's citizens, with official secrets. When Denis decides to pass on those secrets to political community Red, he breaches their trust in him qua public official. More importantly, however, there is also a sense in which he can be described as breaching his fellow citizens' trust: even though they have never met him and therefore do not trust him, Denis, not to pass on official secrets to Red, they trust whichever civil servants are in charge not to act in this way. Breaching that trust can constitute an act of treason. 17 That being said, appeals to political trust do not exhaust accounts of the presumptive wrongfulness of treason. If they did, treason would by definition be impossible in a political community in which there is general distrust (in that sense) between state officials, between state officials and citizens, and indeed between fellow citizens themselves. It implies, for example, that Soviet citizens could not, by definition, be labelled as traitors to their community, such was the level of distrust within Soviet society. Yet, it does seem entirely apt to characterise Gordievsky as well as those of his fellow Soviet cit- 16 For an illuminating review of arguments against disobeying the law and their applicability -or not -to the crime of treason, see Lee, 'Punishing Disloyalty?'. 17 I draw on accounts of personal trust which can be found in, e.g., Annette Baier, 'Trust and Antitrust', Ethics 96 ( izens who worked for foreign agencies as traitors -as having done something which stands in need of justification. To be clear, I do not dispute that treason can and often does involve a breach of political trust. I merely suggest that we have yet to explain why treason is presumptively wrongful when the relationship in which the traitor stands vis-á-vis fellow community members is not characterised by trust. Now suppose, by analogy, that Carl does not trust Anna to keep his secret. Yet, against his better judgment, he confides in her and asks her not to share what he has just told her with anyone else. In so doing, although he does not trust her, he communicates the fact that he endorses the norm that one generally ought not disclose that which has explicitly been told in confidence. If she gives this undertaking and then goes on to disclose it all to Beth, she is betraying him -even though he knew deep down that she would so act: at the very least, she owes him an explanation, and possibly even an apology. To be sure, she would also owe him an explanation, and possibly an apology, even if she were a stranger to him and had merely come across sensitive information about him which, she ought to assume, he would want her to keep secret from third parties. But her explanation and apology, in this case, unlike in the former case, would not need to take account of their pre-existing relationship since there would be no such relationship to begin with. 18 Similar considerations apply to some cases of treason -cases, that is, in which the parties have mutually understood and articulated expectations that they will not disclose sensitive intelligence to third parties in general, and foreign powers in particular. It is particularly apt in cases in which the agent has taken an oath to the effect that he would serve his country, or occupies a position such that, it is understood by all, he is bound by secrecy. The point applies, moreover, even in cases where the traitor benefits rather than harms the vital interests of his fellow community members. Again, by analogy, even if Anna is benefitting Carl overall by disclosing his secret to Beth, she is still betraying him, and here too there still is a sense in which she owes him an explanation, and possibly even an apology. Likewise, Gordievsky provided the British and American governments with a raft of vital intelligence about the KGB's operations in Britain and, more importantly, about the mindset of Soviet leaders in the early 1980s at a time of heightened tensions between the USSR and the USA. But even if he benefitted his fellow Soviet citizens overall by so doing (which seems plausible), there remains a sense in which he breached a presumptive obligation not to disclose their and their country's secrets in defiance of shared expectations of non-disclosure.
So far, we have two arguments in favour of the claim that treason is presumptively wrongful: the argument from political trust, and the argument from mutually understood expectations (understood, that is, by dint of explicitly articulated commitments and/or institutional norms). The third argument goes like this. Return to the example of Jake, who provides a foreign power with US state secrets. Suppose that those secrets are of such nature that, were he to pass them on to (e.g.) the Chinese authorities, he would expose some individuals to a serious risk of wrongful harm. He is under a presumptive duty not to do so -by dint of obligations which we all have to all other human beings, irrespective of any relationship which we may have with them or of what we have undertaken to do or not to do for them. Now, in one variant of this scenario, he is a British citizen with no particular connection to China. His act consists solely in a breach of the general presumptive duty not to expose individuals to a serious risk of harm. In my original variant of the scenario, however, he is a first-class US citizen (a fact of which he is aware.) If he passes on those secrets to the Chinese, he would be aptly regarded as a traitor. To the extent that his act, qua treason, is presumptively wrongful, it cannot be merely by dint of the fact that he harms the vital interests of individuals who happen to be American. On the contrary, it must be by dint of his special relationship to the US. 19 The question, then, is what it is about that relationship which grounds a presumptive duty not to pass on those secrets. As someone who enjoys social membership in the United States, he benefits from the ties that unite him to this particular political community. Yet, by passing on information to the Chinese authorities surreptitiously, he undermines the institutions which his fellow community members support, thanks to which he can enjoy the 19 On which point I agree with Youngjae Lee. See ibid., p. 319.

THE MORALITY OF TREASON
benefits of membership, and on which they too are dependent. He accepts the benefits of membership while at the same time harms his fellow community members: it is for this reason that his act of disclosure is presumptively wrongful -that is to say, that the burden of proof lies on his shoulder to justify it. 20 It might be objected that this defense of the presumptive wrongfulness of treason does not properly account for the intuition that, at the heart of treason -that which makes it wrong -is a breach of loyalty. For (the objection continues) it seems odd to say that individuals owe loyalty to their state because the latter confers benefits on them; and it seems even stranger to say that they owe loyalty to their fellow community members by dint of the fact that the latter contribute to the state providing benefits in this way. Rather, they owe loyalty to their state because it is their state; they owe loyalty to their compatriots because the latter are their compatriots, period. 21 Loyalty, on this view, is necessarily non-instrumental. Yet (I contend) it need not be so. In institutional settings, in particular, loyalty can simply take the form of commitment and willingness not to undermine institutions which do well by us and, thereby, not to harm our fellow community members. Treason is presumptively wrong in so far as it consists in a breach of loyalty so construed.
As should be clear, the charge of treason -as the commission of a presumptively wrongful act -is context-sensitive to the extent that it is shaped by mutually understood expectations as to what agents are expected to do in fulfillment of their duties to one another. Moreover, it holds irrespective of the agent's motives: whether Jake is 20 Two points. First, the benefit argument implies that someone who undermines the community from which he benefits, without doing so for the sake of, in consultation with, or with the help of, foreign parties, can be characterised as a traitor. If a bona fide member of the United States were to assassinate the US president for profit or ideology, with no reference whatsoever to foreign interests, I would have no difficulty in labelling him/her a traitor. Some readers might disagree. If so, they can add a further condition to the presumptive wrongfulness of treason, to wit, that one must act for and/or with foreign parties.Second, the benefits argument I am sketching out here resembles Ekins et al.'s argument in their Aiding and Abetting the Enemy. But it differs from theirs in the following respects. Unlike me and as noted above (supra ft.17), they tie the receipt of benefits to political trust. In addition, they do not consider cases where a community member does not benefit from his membership (and thus does not, on my view, have special duties to other community members). Finally, they assert that treason is always, and not merely presumptively, wrong. motivated by hatred for his country, greed, or desire to improve the lot of his fellow members, what matters is what he has received from his community and/or what he is expected to do in return. Finally, on the account I have just given, acts of treason are presumptively worse than non-treasonous yet harmful acts, and some treasonous acts are presumptively worse than others. 22 On the first count, other things equal, it is presumptively worse of American Jake to pass on US secrets to the Chinese than it is of British Jake, to the extent that (to repeat) American Jake undermines the institutions thanks to which he benefits from social membership in the United States and on which his fellow Americans depend.
On the second count, suppose that Jake is a state official -say, he is working for the CIA. His wrongdoing is worse, other things equal. My account accomodates this. Public officials are even more strongly expected by those whom they serve not to harm the latter's vital interests (even if, as a matter of fact, they are not trusted to do so); indeed, they will often have made an explicit commitment to do so (even if, as a matter of fact, they honor their commitment more often in the breach than in the observance.) 23 Incidentally, my account also accomodates the following case. Suppose that Jake is a CIA employee. He passes secret intelligence about to the US to China. Suppose further that, unbeknownst to him, US counter-intelligence services are fully aware of what he is doing. They choose not to arrest him but, rather, to manipulate him into feeding to China intelligence which the US administration regime wants China to have. Jake, thus, does precisely what US counter-intelligence services want him to do. However, it is still apt to describe his act as presumptively wrongful, in so far as he thinks that he is acting against the United States in defiance of the fact that he is understood to be under a role-based obligation not to disclose those secrets. 22 The claim that treasonous acts are worse other things equal than non-treasonous acts is compatible with R. E. Goodin's view that our duties to compatriots are derived from our general duties to human beings in general -that is to say, that our having those special duties is simply an efficient way of discharging those general duties. On Goodin's view, wrongfully harming a compatriot is worse than harming a distant stranger, but only contingently on the fact and to the extent that in so doing one fails to discharge one's general obligations. As Goodin notes, those special duties are not so special after all. Nevertheless, to the extent that we have them, they demarcate those to whom we owe them, from those to whom we do not owe them.. 23  One final point about the presumptive wrongfulness of treason. As my three-pronged argument so far suggests, even treason against an unjust regime is presumptively wrongful. The more unjust the regime regime, the weaker the presumption. But the presumption nevertheless stands, such that treason does stand in need of justification even in such cases.

III. PERMISSIBLE TREASON
We now have a handle on what treason is and why it is presumptively wrongful. Someone who wishes to disclose secret information about his political community without authorisation needs to show that he has a justification for violating the trust his fellow members place in him, and/ or for defying shared expectations that he not so act, and/or for undermining the institutions thanks to which he enjoys the benefits of social membership and on which his fellow members also depend.
To say that treason is presumptively wrongful is to imply that the presumption can be overriden by countervailing considerations. In this section and the next, I provide an account of those countervailing considerations.
Let us assume that Blue's leadership is embarking on an unjust foreign policy: it authorises, and its officials commit, or contribute to committing, violations of fundamental rights. At the very least, Blue's members are pro tanto morally permitted not to take part in those wrongdoings in the first instance, and to extricate themselves from the unjust venture at the first opportunity -for example, in the case of officials, by resigning. 24 The question is whether they are morally permitted to go one step further and to act against Blue by passing on official secrets about its policy to Red. As a first and very rough cut, here is a simple analogy. Suppose that Albert becomes 24 I say pro tanto, for as Tony Coady persuasively argues, it does not follow from the fact that one has initiated a wrongful course of action that one may, let alone must, desist: the costs others would accrued as a result must be taken into account. C.A.J. Coady, 'Escaping from the bomb: immoral deterrence and the problem of extrication', in Nuclear deterrence and moral restraint, ed. H. Shue (Cambridge: Cambridge University Press, 1989). The claim that Blue's citizens contribute to committing violations of fundamental rights might seem controversial: yet, in some important sense, they -or at least some of them -do, by paying taxes, supporting the regime and its policies, etc. aware that his business partner Bob is planning unjustifiably to assault Innocent as part of their business venture. It surely is not enough to say that Albert is morally permitted to get out of the business venture. It seems that he does not owe it to Bob not to thwart his unjust plans by warning Innocent of Bob's wrongful plans.
By implication, Blue's members do not owe it to one another not to pass on to Red secret intelligence about Blue's ex hypothesi unjust foreign policy, thanks to which Red will be able to thwart Blue. The point holds of ordinary citizens who happen to have that intelligence in their possession. It also holds of Blue's officials -even if the latter have explicitly taken an oath of loyalty to Blue's leaders. One can no more validly pledge to serve political actors who commit violations of fundamental rights than one can validly pledge to serve a Mafia boss who commits similar wrongdoingseven if one derives considerable benefits from belonging to either the Mafia or one's political community. Whilst the presumption against treason is stronger (for reasons indicated in s.II.B) for officials than for nonofficials, it does not hold no matter what. 25 Crucially, however, it does not follow from the claim that a member of Blue (henceforth, Asset 26 ) does not wrong his fellow members by committing informational treason that he is morally permitted to do so all things considered. Before reaching the conclusion that he is, we must bear in mind the following considerations. First, Asset might have countervailing special obligations to third parties which are incompatible with committing treason. Suppose that if Asset is unmasked as a traitor, his relatives will receive a ten-year sentence in a hard labour camp, which they are unlikely to survive. Granted, Red is under a duty of care to protect both Asset and his dependents, for example by spiriting him out of Blue's territory at the first opportunity -as MI6 did for Gordievsky. Nevertheless, we can imagine cases in which a very high likelihood of discovery and collective punishment together with a very low 25 For the view that, on the contrary, lying by means of promissory speech or in breach of an oath is morally wrong, even if it is done for the sake of saving innocent lives or, more generally, thwarting wrongful ends, see H. Grotius, The rights of war and peace, ed. R. Tuck (Indianapolis: Liberty Fund, 2005 [1625]), Bk.III, ch.I, § §XVI, XVIII and XIX. 26 I use this label, since in the parlance of the intelligence community, the sources which intelligence officers cultivate in target countries are called 'assets'. likelihood of being afforded this kind of protection would combine to render Asset's act of treason impermissible in this sense. 27 The second consideration pertains to what Red will or is likely to do with Asset's intelligence. Suppose that the intelligence is useless. Or suppose that it is useful but that Red will use it to wrongful ends, or will not use it at all. For example, its services fail to realise how important it is. Or they do not trust Asset not to be a double agentin the same way as the Soviet authorities apparently never fully trusted Philby. 28 Or they lack the requisite analytic resources. Suppose further that Asset has very strong reasons to believe that such is the case. On the one hand, by disclosing official secrets to Red, Asset might be aptly described as engaging in an overall harmless, honorable act of rebellion against Blue. On the other hand, by taking this particular step, Asset might be more likely to expose third parties to pointless risks of harms. The difficulty is that Asset may not be in a position to form a grounded judgement as to the effectiveness and risks of his act of treason -precisely and in part because he is ex hypothesi an outsider in relation to Red: whilst treason is in some sense an act of treachery towards one's side, it is also an act of faith in the other side.
Under conditions of epistemic uncertainty, the best that moral agents generally can do -indeed the least that they ought to do -is to ground those decisions in the beliefs which they have formed in the light, not just of the actually available evidence, but of the evidence which they can reasonably be expected to have acquired. If this is correct, Asset is permitted to pass on intelligence to Red's authorities if and only if he has formed the belief on the basis of the relevant evidence that the facts are such as to warrant so acting, and is acting on the basis of such belief. Under those conditions, we can say that Asset is robustly justified in committing treason. By implication, if Asset lacks the relevant evidence, he may not so act -even if (for example) Red would make morally justified use of the intel-27 Red's duty of care to Asset to protect him from being unmasked is itself subject to moral constraints. We may readily agree that Red's intelligence services are morally justified in dangling fake, nonexisting traitors in front of Blue's counter-intelligence services, to divert their attention from Asset. We may also readily agree however that they may not lead Blue's services to another yet no longer useful traitor, if the latter would be executed by Blue as a result. The ways in which the KGB succeeded in throwing the CIA off Aldrich Ames' scent nicely illustrate the point. ligence he would provide. While Gordievsky did have pretty good evidence that the British authorities would make morally justified use of the intelligence he was able to give them (based on Britain's relatively democratic traditions, commitment to the rule of law, etc.), Philby had no such evidence (on the contrary) that the Soviet authorities would do the same. Therein lies an important difference between them. Therein, too, lies an important difference between ordinary citizens and officials: the latter are likely to have better access to the relevant evidence than the former. Even though they are under a stronger presumptive duty not to commit treason than ordinary citizens are when things are equal, they are usually in a better position to overcome those epistemic obstacles and thus to act justifiably.
Let me now address two important concerns one may have about my arguments thus far, drawn respectively from the works of David Estlund and the works of Youngjae Lee. I should say at the outset that, on reflection, Estlund and Lee might both accept my central arguments: my discussion, thus, is not ad hominem. It is however worth rehearsing those concerns, as they naturally arise from their works.
In his influential defense of a duty to obey unjust orders, Estlund argues that 'when authoritative commands arise out of an epistemic procedure of a certain kind, there can be a duty to obey commands to carry out even some unjust policies or punishments.' 29 The procedures Estlund has in mind are democratic procedures, which issue in political justifications acceptable to all reasonable points of viewjustifications which, in turn, ground authoritative commands. Estlund does not dispute that an agent who obeys an unjust order would be guilty of wrongdoing. His point, however, is that fair institutions are always vulnerable to making honest mistakes, and that under the aforementioned conditions, an agent is under a duty to do wrong -and, by implication, is not at liberty to disobey.
In the context of informational treason, the argument applies to the officials of democratic states, who by dint of their institutional role are deemed to be under a special obligation not to divulge official secrets without authorisation, and who (if Estlund is right) are not morally permitted so to act even if they have overwhelm-29 David Estlund, 'On Following Orders in an Unjust War', Journal of Political Philosophy 15 (2007): 213-234, pp. 221ff. ingly strong reasons to believe that their silence protects violations of fundamental rights. By contrast, the officials of non-democratic states are in principle at liberty to disobey unjust orders. Thus, Gordiesvky's and Penkovsky's treason does not fall foul of Estlund's argument, precisely because the Soviet regime clearly did not meet the aforementioned requirements. Moreover, Estlund's argument does not apply to former officials, who have been sacked or have resigned, and who therefore are no longer part of a chain of authoritative commands. For example, even if the Russian regime (unlike its predecessor) meets Estlund's procedural requirements, the argument does not apply to Alexander Litvinenko, who worked as a FSB officer for number of years, escaped to and settled in the UK, and allegedly informed on the Russian state for the benefit of MI6 before being murdered in broad daylight in 2006 London by (almost certainly) Russian operatives. Nor does Estlund's argument apply to ordinary citizens, for the same reason. 30 However, there are compelling reasons not to press Estlund's argument into service against the claim that treason can sometimes be permissible. To be sure, fair and democratic procedures through which authoritative commands are issued have a kind of legitimacy which the decision-making procedures of a criminal gang lack -even a gang with all the institutional accoutrements of a state. Nevertheless, even in a democratic country such as the United States, where Congress must authorise the resort to war, such constraints can all too easily be skirted. 31 Moreover, it is worth noting that in France, India and the UK (all fairly robust democracies, all in the top ten military powers) the authorisation of the legislature is not a necessary condition for waging war. If anything, the point is stronger still when foreign policy in general, as distinct from war in particular, is at issue. Decisions to go to war at least are subject to some form of scrutiny -if not by the legislature, at least at the court of public opinion. Decisions to impose economic sanctions, enter into treaty negotiations, authorise arms sales and order incremental troop deployments -far less so. The procedures by which those decisions are made are even less likely to meet Estlund's procedural requirements than decisions to go to war, and it is therefore much harder to object to the treasonous disclosure of official secrets in such cases. Admittedly, the lower down the intelligence cycle officials are, the less robustly justified they are in committing treason, since the less confident they ought to be that they -rather than their superiorsare reaching the correct judgement. It remains true however that in the world as we know it, opponents of informational treason will find little support in Estlund's argument (a point with which, I hasten to reiterate, Estlund himself might well agree.) A second objection to the claim that Asset is (under some circumstances) permitted to commit informational treason is drawn from Youngjae Lee's usurpation argument against treason. In brief: the state cannot fulfill the valuable functions which its members ask of it unless it is physically secure. It cannot enjoy physical security unless it, and it alone, has the power to make use of the required resources in general -and to control the resort to violence in particular. Citizens who pass on state secrets to foreign powers or who disclose them to all and sundry 'usurp' powers which are the state's alone. In Lee's apt turn of phrase, they engage in 'foreign relations vigilantism'. 32 Unlike Estlund's, Lee's argument applies to ordinary citizens as well as public officials, and to citizens and officials of nondemocratic states as well as to those of democratic states. It also echoes one of my own, earlier arguments, to the effect that the presumptive wrongfulness of treason partly lies in the fact that the traitor undermines the institutions from which he benefits and on which his fellow members depend. However, the difficulty with invoking foreign relations vigilantism against the claim that treason can (under some circumstances) be morally permissible is that it seems to rule out ever acting in opposition to one's state in the field of foreign policy. For example, it rules out private enlistment in a foreign war, even if one does enlist on the side for which one's state symbolically stands. It also rules out fighting for the just side, against one's state. Yet, there are good reasons for thinking that, at least in some cases, it is morally permissible so to act. (Consider, for example, the enlistment by (e.g.) French and British citizens with the International 32 Lee, 'Punishing Disloyalty?', pp. 332-341. Lee rejects the view that the wrong of treason lies in the fact that it is an act of disloyalty and it is that which leads him to invoke the notion of vigilantism. As I argued in s.II.B, there is scope for disloyalty in an account of the wrongfulness of treason. Readers who are persuaded by my argument to that effect can also endorse Lee's argument from vigilantism, as the two arguments are compatible.
Brigades in the Spanish Civil War, or the enlistment of French citizens with the Allies and against the Vichy regime.) If this is true, the charge of usurpation alone does not undermine the moral permissibility of informational treason.
To this point, Lee, whose concern is with the moral foundations of the American criminal law of treason, might respond that it does not apply to a minimally just state such as the United States. He might further note that even if treason under those circumstances is morally permissible, it still ought to be a criminal offense. Again, he might well be right. But my focus is on the morality of treason, tout court. Pending further objections to the contrary, thus, the case for informational treason stands.

IV. MANDATORY TREASON
I have argued that individuals are sometimes morally permitted to disclose official secrets about their own political community to another political community, as a means to help the latter thwart the former's violations of fundamental rights -thereby committing treason. As I now argue, they are sometimes under a moral duty to do so -by dint of an obligation of Good Samaritanism.
Suppose that you cannot get resources which you badly need and to which you have a right. At the very least, I am under a duty not to prevent you from getting those resources, and in fact, in some cases, under a duty to give them to you. This applies to material resources -such as the basic necessities of life but also the resources we need to protect ourselves from unwarranted harm (for example, guns). It also applies to what we may call 'personal services' -such as rescuing someone from drowning, and more generally putting in place rescue missions -in other words, time, energy, the deployment of our mental and physical abilities, and so on. Moreover, the duty to protect someone from unwarranted harm takes precedence over duties we may otherwise have to those agents who cause the harm. Were I to obstruct your attempt to procure the relevant resources, or were I to fail to help you, thereby enabling my associates to pursue their wrongful ends, I would be complicitous in those end. And that, I may not do. 33 The following two cases bring these considerations to bear on the issue of informational treason: Passive Traitor Asset is a high-ranking officer in Blue's weapons development programme. He knows that if he absents himself from his desk on a specific day and at a specific time, and leaves his computer logged onto his department's network, Officer Red and her team will manage to download secret information about the programme remotely.
Active Traitor Asset is a high-ranking officer in Blue's weapons development programme. He is in a position to download important information about the programme and to pass it on to Officer Red.
In both cases, Asset is under a duty to Red to help the latter obtain the resources it needs -in this case, secret intelligence. Were Asset to obstruct Officer Red by remaining at his desk, or to refuse to download and pass on the intelligence, he would be assisting Blue in the commission of very serious wrongdoings. I take it that one may not act in such a way as to be an accomplice in grievously wrongful ends.
The duty to commit informational treason encompasses both a duty to pass on the needed intelligence to those who are under a wrongful threat of grievous harm, and a duty to pass on the needed intelligence to those victims's 'rescuers'. In the latter case, the duty is owed to both victims and rescuers when the latter are under a duty to rescue. Generally, agents have rights against one another not to be impeded in the performance of their duties to third parties.
It might be thought that, in this pair of cases, Asset's duty to commit treason is grounded in the fact that he occupies a position which gives him access to the relevant intelligence and thus puts him in a position to help Red. And, relatedly, it might also be thought 33 The literature on duties to provide assistance to those in need, which include a duty to protect them from unwarranted harm, is enormous. that Asset is under a particularly stringent duty so to act if he chose to occupy such a position. 34 It is true that agents who freely take a particular course of action at time t 1 are sometimes under a duty to do x at time t 2 , even if they would merely have been permitted -yet not obliged -to do so had they taken a different course of action. For example, I am not under a duty to go to the beach. But if I go to the beach and notice a child at serious risk of drowning, I am under a (pro tanto) duty to try and rescue her. Had I stayed at home and heard that there was a child in trouble at the beach, I would not have been under a duty to (e.g.) drive to the beach like a lunatic -at serious risk to my own safetyin order to get to the child. Or suppose that I freely promise to do something for you which I am not under a duty to do. In so doing, I impose a duty on myself to help you.
However, agents can also be under a duty to help even if they have not chosen to put themselves in a position where they can help. Suppose that I am frogmarched to the beach at gunpoint. The fact that I did not choose to go to the beach does not imply that I am not under a duty not to help the child. Furthermore, it is not always the case that agents are under a more stringent duty to help if they have willingly put themselves in a position to do so, than if they have not. Suppose that I am frogmarched to the beach at gunpoint whereas you go there of your own accord. We are both in a position to save the child. Other things equal, the circumstances under which we got to the beach do not seem to make any difference to our duty to rescue the child.
If I am right, any agent who finds herself in possession of secret intelligence the disclosure of which would stymie violations of fundamental rights is under a pro tanto duty to pass on that intelligence, thereby committing treason if she stands in a treason-qualifying relationship with the parties whose secret it is.
That said, the claim that treason can in principle be the means by which agents fulfill a general duty of assistance needs qualifying in the light of two considerations which, unfortunately, pull in opposite directions. On the one hand, treason is likely to be extremely costly to traitors. Yet there is a limit to the sacrifices we can reasonably expect putative Good Samaritans to incur for the sake of others.
There thus might come a point where the magnitude of the costs which a traitor would incur together with the likelihood that he would be found out and have to incur those costs exonerate him from a duty to betray. Where to set the threshold for unacceptable costs is impossibly hard to establish -in the same way as it is impossibly hard to reach fine-grained judgements as to how much by way of taxes a well-off individual may be expected to pay for the sake of the needy. I surmise, though, that a high risk of being executed, tortured, or sentenced to a lengthy prison term does constitute too high a cost.
On the other hand, the stronger individuals' degree of causal and moral responsibility for rights-violations, the more grievous their wrongdoing should they fail to help, and the greater the costs they can be expected to shoulder when providing assistance (including, in this context, the costs of being found out.) If I willfully and unjustifiably put your life at risk, and if you can save yourself by killing me, I am under a duty not to kill you in my own defence. By that token, if I can rescue you from my own lethal threat by divesting myself of a good without which I would die (for example, a protective vest), I am under a duty to do so. 35 But now suppose that Asset is sufficiently responsible for the rights-violations which Red is seeking to thwart as to be under a duty to provide intelligence to Red at some risk of being found out by Blue's counter-intelligence services, but not sufficiently so as to be reasonably expected to incur the costs of lifelong imprisonment let alone death if he is unmasked. At the same time, it is quite likely that 35 This claim might seem to be in tension with two separate claims. First, I said a few paragraphs ago that one's having chosen to put oneself in a position where one can help does not (other things equal) make a difference to the stringency of one's duty to help. But if an agent has chosen to occupy a position which entails the commission of rights-violations, and if being in that position also enables him to help victims, then it seems that he is under a stronger obligation to help than if he had not chosen to occupy this particular position. In reply: it is not the fact that the agent has chosen to put himself in a position where he can help which makes a difference to the stringency of his duty; rather, it is the fact that he has chosen to occupy a rights-violating position.Second, and by that token, it might seem that the claim that officials who are responsible for rights-violations are under a duty to incur greater costs, in committing treason, than non-responsible agents, is in tension with the claim (supra s.II.B) that officials are under a stronger duty not to commit treason than ordinary citizens. For now it seems that an agent's official position supports both the claim that he is under a more stringent duty to commit treason -in mitigation of his contribution, qua official, to rights-violations -, and the claim that he is under a more stringent duty not to commit it -on the grounds that he has explicitly undertaken to serve the relevant institutions -than if he were an ordinary citizen. In reply: the latter claim applies other things equal, and is compatible with the view that when things are not equal, for example when the official is partly responsible for rights-violations, that official is under a duty to act treasonably whereas the ordinary citizen is not -or is under a more stringent duty to do so than the latter. he will be exposed to those costs, at Blue's hands, precisely by dint of his high-ranking position. Asset's responsibility for Blue's rights-violating policy renders his duty to commit treason both more and less stringent than if it would be if he were a menial official or an ordinary citizen of Blue's. 36 Here too, we need to know much more about each case to ascertain whether Asset is merely doing his duty or is acting in a superogatory manner. The point remains, however, that it is in principle plausible that, in some cases, at least, Asset is under a duty to commit treason notwithstanding the high costs he is liable to incur if he is found out.
It might be objected that, were we to conclude that Asset is merely doing his duty, we would be both unduly ungenerous to him -in so far as he would be described as merely having done his duty and no more -and unduly critical of those similarly situated agents who, unlike him, remained loyal. This is not so: to say that someone is under a duty to do x is compatible not only with the familiar claim that he can be excused for not doing his duty (for acting under duress) but also with the less familiar claim that he is deserving of praise for doing it (precisely because he is acting under duress.)

V. TREASON, ALLIANCES AND SHARED GOALS
To recapitulate, I have argued that agents are sometimes morally permitted, indeed sometimes morally obliged, to pass on secret intelligence about their political community to the latter's declared enemy. Now suppose that Red and Blue are allies and that Asset, who 'belongs' to Blue, offers to pass on official secrets to Red. Is he committing treason and, if so, can his act of treason ever be justified?
At first sight, it might seem that he is not committing treason. Thus, on Philby's own and admittedly self-serving construal of his work as a double agent during WWII, given that the UK and the USSR were allies, he was working towards their shared goals. On my account of treason, however, an act counts as treasonous if the agent stands in an appropriate relationship with those whose secrets he is passing on. By implication, Philby, who was enjoying social mem-36 This will not always be the case. Asset might be high ranking enough to enjoy high-level protection thanks to which he might be able to get away with a moderately harsh punishment, particularly if his act of treason is or can be dressed up as relatively minor. The general point however is this: Asset's institutional position can make a difference to the stringency of his duty to commit treason -whether or not he freely chose to occupy this position. bership in the UK and who was under a directive not to disclose its military secrets, was committing treason between June 1941 and the end of the war. The fact that the UK and the USSR did share the goal of defeating nazism is irrelevant. In any event, Philby's argument, if it works at all, only applies to those cases in which the intelligence pertains to those shared goals. While Britain and the USSR shared the goal of defeating nazism, their post-war ends were not compatible -and there is evidence that Philby did work to frustrate the former for the sake of the latter. 37 Suppose, then, that Red is pro tanto justified in seeking to obtain intelligence which its ally Blue wishes to keep secret. The justification for treason which I offered in Sections III and IV applies to allies. For a start, the alliance might be one of pure expediency, where both parties know that the other will breach the agreement at the first opportunity. The 1939 pact of non-aggression between Germany and the USSR aptly illustrates the point. Richard Sorge, a German citizen who joined the Communist Party and volunteered to work as an undercover spy for Moscow following WWI, disclosed to Moscow what he knew of Germany's invasion plans in 1941: that he was a traitor, and acted justifiably, seems relatively uncontroversial. 38 Moreover, an alliance might be robust at time t 1 and yet disintegrate at t 5 . What matters is not the fact of the alliance itself but whether Red is epistemically justified in inferring from the facts at its disposal that Blue might be planning to commit serious rights-violations and that it is warranted in trying to ascertain whether its belief is true by acquiring Blue's official secrets. If so, the fact that Asset's act of treason occurs in the context of a geopolitical alliance between Red and Blue does not in itself count against it. This is not to deny that this fact makes a difference. As we saw in s.III, the lower down the intelligence cycle Asset is, the less evidence he has to make the relevant judgement. The existence of an alliance between Red and Blue compounds the problem. For the closer the relationship between Blue and Red, the fewer and the weaker Asset's reasons for believing that Blue might in the future pose a credible and wrongful threat to Red's interests, and the higher, thus, the 37 On Philby's work during WWII, see e.g. Knightley, The Master Spy, ch. 8. 38 Sorge is widely acknowledged as one of the most successful spies of the 20th century. See e.g. evidentiary threshold which he must meet for his act of treason to be justified. However, the fact that the threshold is very high does not preclude that it can be met and that, when it is met, treason is (pro tanto) justified.
So far in this paper, I have assumed that Red and Blue are independent political communities. Now suppose that Blue belongs to a supra-national organisation, and that this supra-national organisation -call it Supra -has a military force, to which member-states contribute both financially and by seconding military staff. Suppose further that Supra is conducting a large scale military exercise (EX-ERCISE), so as to test its preparedness for a multi-front war with a non-member. Asset is a member of Blue's intelligence services and in the course of his work comes across sensitive information pertaining to Blue's side of EXERCISE, which he decides to disclose to Supra; or, he has been seconded to work for Supra and, in the course of his work comes across sensitive information about Supra's side of EX-ERCISE, which he decides to disclose to Blue.
I have not encountered such cases in the contemporary empirical literature. Perhaps there have been some, which have remained protected under the cloak of official secrecy. Even if there has not been a single such case, it is likely that, sooner or later, there will be one. The issue, thus, is not purely academic: it also invites us to reflect anew on the ethics of overlapping and conflicting political and institutional loyalty.
Let us suppose for the sake of argument that the sensitive information which falls into Asset's hands pertains to violations of fundamental rights, and that the disclosure of the information would enable its beneficiary to take the requisite protective steps -subject to countervailing considerations as described in s.III. Let us suppose, thus, that diclosure is pro tanto justified. The question is whether Asset would act treasonously (albeit justifiably) were he so to act.
In s.II, I argued that the traitor is someone who enjoys bona fide social membership in a political community -and who thus derives considerable benefits from that community -and who acts in defiance of explicit undertakings or mutually understood expectations that she will not violate the trust placed in her by fellow community members and/or harm their interests qua fellow members. To ascertain whether Asset would act treasonously, we need to distin-guish between two cases. In the first case, Supra is not a kind of interstate organisation in which individuals enjoy social membership. NATO is a textbook case of such an organisation. To be sure, in so far as Blue is subject to NATO's directives and operational principles in relation to EXERCISE, Asset's primary (presumptive) professional obligation is understood to be to NATO and not to Blue. Moroever, it is also plausible to say that there is a sense in which Asset benefits from Blue's membership in NATO. However, whatever benefits he so derives are not granted to him directly by NATO. For this reason, even if he discloses sensitive information to Blue about EXERCISE without being authorised by NATO and in awareness that such authorisation would not be forthcoming if asked for, it is not apt (on the social membership thesis) to describe him as a traitor to NATO. By implication, and precisely because he is a social member of Blue, if he discloses information about the latter, in relation to EXERCISE, to NATO, without being authorized to do so and in awareness of the fact that such authorisation would not be granted, he is acting treasonously.
In the second case, Supra is a kind of inter-state organisation in which individuals enjoy social membership. The European Union is a textbook example. Its individual members enjoy two kinds of citizenship, and thus two kinds of social memberships: citizenship in one of the member state on the one hand, and EU citizenship on the other hand. While the latter is derivative of the former (in that one cannot be a EU citizen unless one is a German, Romanian, French, Irish, etc., citizen), it confers, via EU law (and its incorporation into domestic legislation) a range of rights to its holders. 39 Suppose, then, that Blue belongs to the EU. Asset is a citizen -or social memberof Blue and (thereby) a citizen -or social member, of the EU. As it happens, the EU has a number of battlegroups, which are fully functional and deployable although they have not yet been deployed. So, suppose that Asset is seconded by Blue to work with the European Defense Agency (EDA), and in that capacity is tasked by the 39 For an accessible and recent account of what EU citizenship is (and in particular of the challenges it faces following Brexit), see Jo Shaw, 'EU citizenship: Still a Fundamental Status?', in Debating European Citizenship, ed. Rainer Bauböck (Cham, Switzerland: Springer International Publishing, 2019), 1-17. A further point on the text: to be crystal clear, when I say that to enjoy citizenship is to enjoy social membership, I am not implying, contra my argument in s.II.A, that citizenship is a necessary condition of social membership. As I noted there in connection with long-term residents, one can be a social member without being a citizen. latter to work on EXERCISE. The question, you recall, is not whether he would act unjustifiably by passing on secret information about Blue to the EU and vice-versa (I assume that he would be justified): rather, the question is whether he would ask treasonously, given that he benefits from membership in both. On my contextsensitive account of treason, whether he does depends on parties' mutual understandings of what their individuals members are expected to do when faced with dilemmas of that sort. It is not implausible to suppose that, in so far as Blue is subject to the EDA's directives and operations principles in relation to EXERCISE, Asset's primary (presumptive) obligation is understood to be to the EDA and not to Blue. If he discloses sensitive information to Blue about EXERCISE without being authorised by the EDA and in awareness that such authorisation would not be forthcoming if asked for, he acts treasonously. By implication, and precisely because his presumptive obligation is understood (or, at any rate, should be so understood by his own goverment) to be to the EDA as far as joint exercises are concerned, if he discloses sensitive information about Blue (in relation to EXERCISE) to the EDA without being authorised to do so by Blue, and in awareness of the fact that Blue would not provide such authorisation if asked, he does not act treasonously.

VI. BENEFITTING FROM WRONGFUL TREASON
Suppose that, in a given case (or even simpliciter, for those readers not convinced by my arguments so far) Asset's act of informational treason is not morally justified. This raises the interesting question of whether Red is nevertheless morally entitled to use that intelligence. The question is not unique to treason: for example, it also arises with torture. In the latter case, many (I suspect) would hold that political actors ought not to accept intelligence from third parties which has been collected in this way. The thought is not just or even mainly that the intelligence is unlikely to be reliable; rather, the point is that accepting such intelligence makes one complicitous in the commission of a grievous wrongdoing. 40 It is doubtful that we may never under any circumstances wittingly and freely make use of the proceeds of another person's wrongdoing -even when that wrongdoing is interrogational torture. To ascertain whether Red may justifiably make use of Asset's wrongful treason, we must consider the grounds upon which Asset acts wrongly. Suppose that Blue's policy vis-á-vis Red is unjust and that Blue's regime and citizens do not have a right that intelligence about this policy remain secret. Red, in other words, has a justification for seeking to acquire that intelligence. Suppose however that Asset is morally prohibited from betraying Blue, on the grounds that he is thereby unjustifiably exposing third parties to a very high risk of incurring a serious harm. Whether Red may justifiably make use of Asset's wrongdoing depends on at least two considerations: (a) whether Red would thereby compound the risks accruing to those third parties, and (b) whether Red is itself under an obligation to those third parties not to expose them either to the same risks as Asset or to the higher risks arising from its decision to act on the intelligence. Suppose that Red's decision would make no difference to the likelihood that Asset would be found out, or that Red is not under the same protective obligation to those third parties as Asset himself is. Under those circumstances, Red is (pro tanto) justified in making use of Asset's wrongful act of treason. Contrastingly, suppose that Red's decision would increase the likelihood that Asset would be found out (by raising suspicions in Blue's intelligence services), and that it is under a protective obligation to those third parties not to expose them to a much higher risk of a greater harm. Under those circumstances, Red may not justifiably use Asset's intelligence. 40 I am doing a bit of reconstructing here. Although the claim that one ought not to benefit from interrogational torture is familiar in public discourse, it is not explored at length in the philosophical literature on torture -which is mostly focused on scrutinising justifications for the resort to torture in the first instance. For a notable exception, see Matthew H. Kramer, Torture and Moral Integrity: A Philosophical Inquiry (Oxford: Oxford University Press, 2014), pp. 309-316 -though Kramer merely argues, in line with article 15 of the 1984 UN Convention against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment, that the use of intelligence obtained under torture should be regarded as inadmissible in judicial proceedings. As Kramer notes, this is compatible with the view that officials are morally permitted to use such intelligence as part of their deliberations over which course of action to take. Suppose now that the reason why Asset acts wrongly is that Blue has a right that the intelligence be kept secret. Specifically, suppose that Red would use the intelligence to pursue unjust ends. For example, Red would be able to facilitate a coup against a democratically elected foreign leader whose overall just programme threatens its own geopolitical interests. Or suppose that Red would be able to pursue just ends by unjust means. For example, although Red is justified in imposing targetted economic sanctions against Blue's regime as a means to get the latter to end its wrongful military occupation of another country, it is not justified in imposing a comprehensive trade embargo; yet, Asset's intelligence suggests that such an embargo would be successful, which prompts Red to go ahead. Under those circumstances, and to the extent that Asset can reasonably be expected to know all of these facts, he is not morally entitled to betray Blue. Nor is Red morally justified in accepting and benefitting from his offer. Of course, as a matter of fact, it is overwhelmingly likely that regimes in Red's position would accept the offer. Nevertheless, it still would be wrong to do so.

VII. CONCLUSION
In this paper, I have argued that an agent can aptly be described as a traitor only if he/she enjoys social membership in the community whose secrets he is disclosing to foreign powers, and offered three arguments in support of the intuition that treason is presumptively wrongful. As we saw, however, passing on political secrets to the officials of a political community other than one's own (be it one's foe or friends) can be justified, indeed mandatory, as a means to thwart violations of fundamental rights.
There is more to be said about treason, of course. In particular, one may wonder whether violations of fundamental rights are the only grounds for justified treason. Moreover, treason often goes hand in hand with the betrayal of trust in personal relationships. Philby committed treason against the British state and British citizens, as well as multiples acts of betrayal against his friends and colleagues. The same can be said of Gordievsky. A full normative assessment of their acts requires not merely scrutiny of their decision to pass on secret intelligence to a foreign power in defiance of their orders or of expectations to the contrary, but also scrutiny of their ongoing deception of those with whom they had personal relationships. Whether treason is justified as a means to thwart wrongdoings other than violations of fundamental rights, and whether personal betrayal tips the balance against treason, are a matter for another paper.
OPEN ACCESS This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creativecommons.org/licenses/by/4.0/.