Concern about “secret diplomacy” in the wake of World War I found its clearest expression in opposition to “secret treaties.” Left-leaning and internationalist groups believed that “publicity” of treatiesFootnote 1 would advance the control of foreign policy by legislatures and citizens, and help construct a more peaceful international order, governed by law rather than power politics. Article 18 of the Covenant of the League of Nations gave expression to these ideals, by purporting to render the binding force of treaties entered into by any member of the League conditional on their registration by the League Secretariat. Registration in turn entailed publication to the world at large.Footnote 2 Article 18 was thus the first instantiation of a norm of treaty publication in international law (with “treaty” here understood in the general international law sense).Footnote 3 Admittedly, this norm was a relatively narrow one. It guaranteed only the publication of treaties once made, and did not address the secrecy of the negotiation phase. Nevertheless, it was a major change to law and diplomatic practice. Aspects of Article 18 were carried over into Article 102 of the UN Charter, which in turn was echoed in the Vienna Convention on the Law of Treaties (VCLT).Footnote 4
This article offers the first detailed history of the norm of treaty publication. It traces the emergence of the norm, and struggles over its interpretation and application, through the practice of the United States, Britain, and France. While these three states are not representative of the international community as a whole (neither in the interwar period nor today), their commitment to greater publicity, and influence over modes of diplomacy, makes them a revealing lens on the norm and its limits. The article draws on public debates in legislatures, international organizations, and legal scholarship, but also discussions within foreign ministries and the League of Nations Secretariat. Analysis from these diverse sources offers a clearer picture of the true legal architecture of publicity and secrecy in treaty practice. It suggests how, and why, margins for secrecy have persisted, even in liberal democracies, and draws out for critical scrutiny the relationship between the norm of treaty publication and legality in the international order.
The article challenges the liberal democratic trajectory some might expect, namely ever-greater publicity over time. It shows that, while statesmen and officials made real efforts to uphold the publication norm, they also fought to preserve some avenue for making commitments that were both legally binding and secret. In international law, the radical potential of Article 18 of the Covenant—the connection it forged between registration and publication, on one hand, and binding force, on the other—was undone in the early years of the League. Article 102 of the Charter is less ambitious than its predecessor. Domestic law requirements, which arguably play as great a role in driving publication as international requirements, have also left avenues for secrecy, although these have been narrowed and systematized.
Within liberal democracies, the article shows that general attitudes to the norm of publication were driven not only by factors like perceived geopolitical vulnerability, but also national political and legal cultures, and even the bureaucratic organization of the treaty apparatus. Specific efforts to carve out exceptions to the norm of publication, or decisions to flout it, were motivated by diverse considerations: perceived needs to preserve some concrete advantage over military or commercial rivals, to stabilize markets or currencies, and to prevent political opposition from national populations, colonies, or allies; but also to conceal markers of inferiority or subordination in interstate relations, and thus shore up a symbolic economy of interstate equality. For bodies charged with upholding the norm, like the League Secretariat, maintaining the force of Article 18 was important for credibility with internationalist constituencies, but there were countervailing incentives to interpret the provision narrowly in order to avoid confronting—and being seen to be defeated by—powerful member states.
Many motivations for secrecy on the part of governments and institutions were self-interested and instrumental. However, officials and reformists also thought about secrecy in principled terms. On occasion, they reconsidered the normative case against secrecy that had been seen as persuasive in 1919. Some came to doubt whether secrecy was always inimical, for example, to peaceful ordering. In some instances, it seemed vital to preserve peaceful relations.
The doubts of contemporaries suggest enduring questions about whether and how the norm of treaty publication favors ends like democracy, peace, and legality.Footnote 5 This article probes, in particular, the relationship between publicity and legality in the international order. In doing so, it gives a sense of the complex empirical terrain in which questions about democracy and peace, too, would have to be pursued.
Actors concerned with publicity articulated different dimensions of the relationship between publicity and legality. Simplified for the sake of analysis, much of the discussion about publicity posited, first, a basic conceptual or normative connection to legality. This could take the form of assertions that legal relations and obligations were inherently public, or that they ought to be public. Second, there was an expectation that the norm of treaty publication would strengthen the role of law in the international order more generally (for example, by improving the dissemination and coherence of the law itself, fostering a commitment on the part of governments to following the law, or levelling out power imbalances between states and vindicating sovereign equality). This combined a general normative sensibility with a number of very condensed assumptions about the effects of publication or secrecy in particular contexts. Finally, there was confidence in law as a means of entrenching the norm of publication. This article shows that each of these stylized dimensions of the relationship between publicity and legality proved more complex—empirically and conceptually—than it had seemed in 1919.
The article intersects with doctrinal scholarship on the law of treaties and emergent research on why states craft interstate commitments in particular ways. However, doctrinal scholarship—cited very sparingly here—often concentrates on problems of classification, like the definition of a treatyFootnote 6 and distinctions between legally binding “treaties” and “gentlemen's” or “non-binding” agreements;Footnote 7 or on questions which flow from classification of texts, like the role of ancillary texts in interpretation.Footnote 8 By contrast, emergent research on why states make agreements in the form they do largely brackets these questions.Footnote 9 This article puts the law at the heart of the inquiry. Unlike doctrinal scholarship, though, it is not aimed at reaching its own classification of the texts considered. Rather, it probes what interlocutors understood themselves to be doing in drafting or handling texts in particular ways.
Attention to these features of treaty practice, often side-lined in doctrinal accounts, gives a much richer picture of the true architecture of publicity and secrecy. It reveals the possibilities created by the interplay of the law of treaties with diplomatic crafts of drafting, and the management of texts within foreign ministry bureaucracies. It suggests a more expansive canvas for states crafting their commitments than can be captured in many studies of state behavior. Governments are not only choosing between treaties and non-binding agreements, for example, but crafting complexes of different texts, making choices about presentation and emphasis which are difficult to capture in quantitative ways. Officials are sometimes uncertain about the status of particular texts, in disagreement with each other, or consciously shaping textual forms to preserve ambiguity about legal status. Finally, close attention to the crafting of texts and the negotiation of the bounds of the category of treaty brings to the fore some paradoxical aspects of secrecy in the international legal order. Secret treaties, by their nature, will never enjoy many avenues for rhetorical or institutional enforcement typically associated with legally binding commitments. They might even thus reflect a faith in legal obligation alone to shape staff behavior. Conversely, however, some secret treaties purport to commit states to actions which may violate international law, calling broader commitments to legality into question.
The article begins with a brief overview of the pre-World War I (WWI) landscape (Part I). It then traces the emergence of Article 18, and the tacit qualification of this provision in the early years of the League's existence (Part II), before addressing interwar confrontations with a comprehensive publication norm in national public law and practice (Part III). While interwar developments are considered in detail, as they shaped approaches taken after World War II (WWII), the treatment of national practice is necessarily selective, focusing on examples which illustrate the range of areas in which secrecy was sought, the repertoire of techniques used, and particular turning points in national approaches. The article then considers the transformation of the treaty publication norm after WWII, a period for which the archival material becomes scarcer (Part IV). It sketches the evolution of techniques of secrecy, and changing legal approaches to its management, from WWII to the present (Part V), before offering a snapshot of the secret treaty today (Part VI). Finally, it summarizes developments over the last century, and what they reveal about the empirical landscape in which to consider the relationship of publicity to democracy, peace, and legality in the future (Part VII).
I. Secret Treaties on the Eve of World War I
Prior to WWI, the secret treaty was a recognized, if criticized, legal instrument. International law imposed no requirement that treaties be published. Such a requirement, if it existed at all, stemmed from national laws or practices concerning legislative knowledge or approval of treaties. There was no consensus on whether a failure to adhere to domestic law respecting treaty-making affected the validity of a treaty under international law.Footnote 10
Of the three states examined here, the United States was the only one with an ostensibly comprehensive requirement for legislative approval prior to ratification, in the form of the requirement for Senate advice and consent for all “treaties” within the meaning of Article II of the Constitution. Once ratified, at least, these texts would be published. Yet, as is well-known, U.S. practice recognized a range of “executive agreements” made without Senate advice and consent.Footnote 11 While most executive agreements were published in one form or another,Footnote 12 some did remain secret, and it was therefore possible to make secret, and legally binding, agreements. However, scholars in the early twentieth century included in lists of “executive agreements” a number of “gentleman's agreements” or “diplomatic agreements” which they considered binding only on the president who made them (and successors, if they did not repudiate the agreements), rather than on the state.Footnote 13 Some of the more notorious instances of secret texts fell within this category, including the 1907–08 “gentleman's agreement,” under which the Japanese Government would limit Japanese migration to the United States.Footnote 14 The “secret treaty” in the sense of a text that was both unambiguously binding on the state under international law, and secret, was therefore marginal to U.S. practice, if it was used at all.
In Britain, the Crown alone had the authority to make and ratify treaties. If a treaty was not of a nature to require any immediate change in the existing law, or further legislative steps for its implementation, there was no requirement that its existence be disclosed to the Parliament or to the public at large. Treaties were customarily published as parliamentary papers, but this publication was not required by law. Accordingly, secret treaties were a feature of British practice, although they were used relatively sparingly.Footnote 15
In France, the president of the Republic, empowered to negotiate and ratify treaties, was required to “give information of their contents” to the legislature “[a]s soon as the interest and safety of the State shall permit.”Footnote 16 This language, taken alone, permitted indefinite secrecy of treaties once made, unless they required promulgation in order to function in domestic law, or entailed legislation for their implementation. However, as in many other Continental European constitutions, certain treaties (“of peace, of commerce, … which involve the finances of the State, … which refer to the condition of persons, [or] to the rights of property of Frenchmen in foreign countries”) were “not final” until voted on by the two Chambres, and no cession, exchange, or annexation of territory could take place except by virtue of a law.Footnote 17 This required certain categories of treaty to be laid before the legislature and thus subjected to formal or de facto publicity. At the same time, many of the most consequential “political” treaties, particularly treaties of alliance, were generally not considered to fall within the categories requiring legislative approval, and could thus be kept secret until “the interest and safety of the State … permit[ted]” their communication.Footnote 18
Although national law and practice sometimes used “treaty” in a narrow sense, and made distinctions between different types of treaties, the orthodox doctrinal position under international law was, by WWI, that there was a single juridical category of treaty:
International compacts which take the form of written contracts are sometimes termed not only agreements or treaties, but acts, declarations, protocols, and the like. But there is no essential difference between them, and their binding force upon the contracting parties is the same, whatever be their name.Footnote 19
This posture sought to impose a formal coherence on what remained a very rich repertoire of forms for the making of promises between states. Variations in nomenclature reflected the fact that older writings on the law of nations had assumed a repertoire of different kinds of agreement, rather than a unitary category of treaty.Footnote 20 Such variations were still used as rough indicia of relative importance and solemnity, and some authors posited distinctions between texts of different kinds which, if they did not indicate a distinction in the nature or degree of legal force, at least marked differences in the circumstances in which the obligations would come to an end. There was a strong sense that the diversity of forms and nomenclature mattered, although it was difficult to say precisely how.Footnote 21
The form of written agreements was not itself regulated by law, but by a diplomatic craft of drafting. Typically, important “political” clauses would be in a principal text, and signed by more senior figures or with greater solemnity, whereas more trivial or detailed clauses could be relegated to an ancillary text, sometimes signed separately by lesser figures (without necessarily implying any lesser degree of legal force). These practices responded to practical and aesthetic considerations, but were also amenable to differentiation between public and secret texts. Secret clauses or “réserves” of the kind satirized earlier by Kant,Footnote 22 and prohibited by some European constitutions, sometimes went as far as contradicting commitments made in the principal (public) text, but there were also less flagrantly contradictory ancillary texts, of uncertain status, which merely gave a different complexion to the principal text. These features of treaty practice—the diversity of forms and terms, and the centrality of a flexible craft of drafting—would prove major obstacles to initiatives for greater publicity.
The use of secret treaties and ancillary texts had already been an object of political contestation, albeit with quite different political inflections, in earlier moments of tension over the relation between the domestic political order and national and imperial foreign policy.Footnote 23 In the late nineteenth century, European jurists, motivated by a need to systematize the material basis of international law, proposed a comprehensive international treaty collection, but conceded that this might include only treaties the publication of which would not be precluded “by reasons of state or by political expediency [convenances politiques].”Footnote 24 In the early twentieth century, gradual revelations of inter-imperial dealings shaped by secret clauses and ancillary texts, associated with mounting tension between France and Germany, prompted more far-reaching attacks on secrecy from across the political spectrum. In France, revelations in 1911 that Germany's non-interference in Morocco had been purchased with the (undisclosed) cession of parts of the French Congo led to the fall of the government and harsh criticism in the Chambres and press. The pro-colonial lobby was most upset about the substance of the bargain, but on the left there were demands for constitutional change to preclude secret treaties. While this failed, the new government undertook to submit foreign policy, insofar as possible, to the Chambres and to public opinion, and to take a broad view of the categories of treaty requiring approval.Footnote 25 Revelations that the British had been apprised of a secret Franco-Spanish treaty gave rise to attacks in Parliament on the perceived underhandedness of this diplomacy.
Spanning different substantive and party-political debates, one can discern a cluster of persistent and related concerns in the early twentieth century. Secret treaties were seen as an autocratic anachronism, out of place in the democratic state insofar as they undermined the proper role of legislatures or, for some thinkers, citizens themselves, in monitoring foreign policy. Secret treaties were seen as inimical to peace, whether because they allowed states to make offensive alliances, or because they fostered an atmosphere of suspicion and apprehension, thereby rendering more likely the outbreak of conflict. This apprehension and resulting possibility of violence in turn favored brute power politics, and was thought to hinder the construction of a law-governed interstate system.
II. The Formulation and Qualification of Article 18, c. 1919–1923
The Early Life of Article 18
World War I threw into sharp relief the power of treaties, public or secret, in the international order. The discovery in 1914 of the extent of British commitments to France was part of the impetus for the formation of the Union of Democratic Control, a group aiming to increase parliamentary control of foreign policy. The Bolsheviks’ 1917 release of wartime secret treaties found in tsarist archives cast secret treaties in a particularly nefarious light, and sharpened preexisting concerns about their effects.Footnote 26 Criticisms of secret diplomacy, while particularly strong on the left,Footnote 27 had some appeal across the political spectrum, especially in Britain and the United States.Footnote 28 These criticisms were both reflected and amplified by Woodrow Wilson's embrace of “open covenants” and “diplomacy [that would] proceed always frankly and in the public view.”Footnote 29
Agendas for reform covered quite a range of positions, from demands that negotiations themselves should be public to demands for publication of final treaty texts, and from a focus on empowering citizens directly in foreign policy, to strengthening legislatures (or committees of experts within them) to act on behalf of citizens. The publication of treaties once made was a relatively minimalist, and seemingly uncontroversial, goal. It had been proposed by various pacifist and internationalist groups during WWI, but the text that would become Article 18 was ultimately introduced into the draft Covenant by American figures.Footnote 30 It attracted no debate in the League of Nations Commission, the body responsible for the drafting of the Covenant. This is perhaps unsurprising, given the sheer number of major issues then under discussion, but it also indicates that Article 18 was consonant with aspirations for the reform of diplomacy that negotiators either accepted, or found difficult to oppose, at least in public.
In its final form, Article 18 read:
Every treaty or international engagement entered into hereafter by any Member of the League shall be forthwith registered with the Secretariat and shall as soon as possible be published by it. No such treaty or international engagement shall be binding until so registered.
A Memorandum of May 1920, drafted by the League Secretariat's Legal Section and the secretary-general, reflected what Article 18 was intended to achieve:
Publicity has for a long time been considered as a source of moral strength in the administration of National Law. It should equally strengthen the laws and engagements which exist between nations. It will promote public control. It will awaken public interest. It will remove causes for distrust and conflict [and] … enable the League … to extend a moral sanction to the contractual obligations of its Members. It will, moreover, contribute to the formation of a clear and indisputable system of International Law.Footnote 31
On paper, Article 18 marked a shift from a pre-war world, in which the secret treaty was criticized but legally permitted, to a post-WWI order in which such treaties would be abolished, at least for League members. The role accorded to the Secretariat in registering treaties sat alongside other provisions which hinted at a larger evolution toward a more collective, institutionalized legal order. Article 19 envisaged that “[t]he Assembly [would] from time to time advise the reconsideration by Members of the League of treaties which have become inapplicable and the consideration of international conditions whose continuance might endanger the peace of the world.”Footnote 32 Article 20 provided that members “agree that this Covenant is accepted as abrogating all obligations or understandings inter se which are inconsistent with the terms thereof, and solemnly undertake that they will not hereafter enter into any engagements inconsistent with the terms thereof.”
The 1920 Secretariat Memorandum offered an expansive, antiformalist definition of the phrase “every treaty or international engagement” in Article 18. This was said to encompass “not only every formal Treaty of whatsoever character and every International Convention, but also any other International Engagement or Act by which nations or their Governments intend to establish legal obligations between themselves and another State, Nation or Government.”Footnote 33
The Memorandum gave full effect to a reading of the Covenant in which registration, and thence publication, was a precondition of legal force. To this end, it proposed that treaties be submitted for registration, at the latest, by the date on which the parties considered them to come into force.Footnote 34
While delegates on the Council approved the 1920 Secretariat Memorandum, closed-door discussions revealed concerns about the full import of Article 18. The delegates agreed between themselves on support for the “grand moral idea” that peoples must not be engaged by governments without their knowledge, and Léon Bourgeois (France) expressed resistance to carving out exceptions to Article 18. However, he joined with Arthur Balfour (British Empire) in arguing that members retained a capacity to conclude treaties that would not be registered—but would nevertheless place the signatories under some moral obligation, which the League might uphold. Dr. van Hamel, the Dutch Director of the Legal Section, was recorded as stating that, under the Covenant, members of the League would not be bound by an unregistered treaty—but that the parties could nevertheless consider themselves bound in some way inter se.Footnote 35 These ominous divergences of opinion about Article 18 were not, however, probed further at this time.
First Tests: French Military Accord and British Financial Arrangements
The first real test of Article 18 was a Franco-Belgian military “accord” of 1920.Footnote 36 At this time, Franco-Belgian military arrangements were controversial within Belgium, where socialist and Flemish deputies tended to oppose arrangements they perceived as likely to render the state dependent on France. It was thus not only military sensitivity, but likely political opposition, which shaped early thinking about the final form of the commitments. After early drafts of the proposed accord had been prepared by military figures, the Belgian foreign ministry contemplated splitting the text into a “political” act, which could be submitted to the Chambres (in accordance with Belgian constitutional clauses requiring legislative approval of certain treaties, and arrangements for foreign troop movements) and submitted to the League Secretariat for registration, and a separate act for the detailed military arrangements, which could be kept secret. The idea of a substantive “political” text was abandoned out of a fear that any treaty involving France alone, rather than France and Britain together, would still attract opposition in Belgium.Footnote 37
The initial “military accord” had not dealt with matters like the casus foederis, and circumstances in which the accord could be brought to an end, but it still had significant political implications. In particular, by requiring Belgium to take action in the event of any German mobilization, it left Belgium potentially implicated in a war between Germany and Eastern European countries; and by envisaging a concerted defense, including along the frontier with Luxembourg, it ran counter to Luxembourg's nominal independence.Footnote 38 A version of this text was signed on July 29, 1920 by Marshal Foch (president of the French Superior Council of War) and General Maglinse (chief of the Belgian General Staff). It was thought necessary for governments to exchange letters approving this accord, which would be published, although this does not appear to have been based on a theory that the military accord otherwise lacked legal force.Footnote 39
The Belgians subsequently proposed amendments to the military accord, and at this point the Belgian ministry of foreign affairs proposed that the public exchange of letters be expanded to cover some aspects of the abandoned “political” accord, and laid before the legislature, without being a treaty requiring approval.Footnote 40 A second version of the military accord thus replaced the initial version (adding, inter alia, a new undertaking by the French to defend the Belgian coast, which was politically controversial insofar as it foreshadowed Franco-Belgian military action without British involvement).Footnote 41 Belgian and French foreign ministers exchanged public letters approving the accord.Footnote 42
When the letters were laid before the Belgian legislature in October 1920, there were demands for the military accord to be published.Footnote 43 The French Ministère des Affaires Etrangères (MAE) expressed concern that, when the letters were submitted for registration, the League Secretariat would immediately demand the military accord.Footnote 44 In Britain, questions were asked in Parliament about whether Britain would press for registration.Footnote 45 Britain was probably the only other power influential enough to press France to submit the accord, and the Foreign Office (FO) was well aware that if other powers withheld politically significant texts as France was apparently doing, “the whole principle of registration will fall to pieces.”Footnote 46 However, Britain had declined to give security guarantees sought by France, and found it difficult politically to do anything which might be perceived as weakening fallback Franco-Belgian arrangements. Moreover, the British Treasury had become concerned, for its part, about registration of various financial dealings.
Internally, French and British officials began wondering if Article 18 might be limited in some way. French officials argued that the military accord should be seen “not as an international treaty, but as an arrangement between general staffs, subject to variations in accordance with military necessity,” and impossible to divulge without compromising national defense.Footnote 47 The British Treasury suggested that “an endeavour should be made to limit the scope” of Article 18 to “International Treaties and Agreements of first class importance.”Footnote 48 The FO dismissed as “useless” any endeavour to limit Article 18 in this sense,Footnote 49 and rejected also an exception for “arrangements of a purely commercial or financial character,” which encompassed exactly the treaties regarding commerce and preferential trading rights that Britain was keen to see states publish.Footnote 50 At the same time, the FO found it difficult to articulate any principled approach:
Possibly the League [Secretariat] might wink at the non-registration of financial agreements of the kind indicated in the Treasury letter; but they cannot logically wink at the non-registration of an instrument [like the Franco-Belgian military accord] which is not so very far removed from an offensive and defensive alliance. There will be a hue and cry about “secret diplomacy” again … . I confess I do not see any formula which will meet with approval, but “agreements of a technical nature” might be stretched to cover anything.Footnote 51
Senior political figures were divided. The foreign secretary, Curzon, thought it was impractical to expect military details to be published.Footnote 52 On the other hand, Robert Cecil (a Conservative politician, and a major supporter of the League, then serving as a delegate for South Africa in the League Assembly) saw the application of Article 18 to the military accord as a crucial point of principle. Drafters of the Covenant, Cecil argued, had hoped that “one of [Article 18's] most valuable results would be to hinder secret military preparations [which] have been in the past a prolific source of suspicion + counter-preparations leading ultimately to war.”Footnote 53
For Balfour, who was to represent Britain at the upcoming League meetings,
[If] we are to interpret [Article 18] as meaning that every nation is at liberty to make what secret arrangement it likes …, provided it does not ask the League to defend that arrangement until it has been published; [then] Article 18 does really nothing to put an end to so-called secret diplomacy. If, on the other hand, Article 18 is to be interpreted as meaning that [withholding a treaty] is an offence against the provisions of the Covenant … then … the refusal of France and Belgium to comply with the Covenant brings us face to face with a very serious international issue.Footnote 54
In the face of this dilemma, Balfour wondered whether, if the Council asked the French and Belgian governments to register the treaty “in its broad outline,” and to give formal assurances that it was purely defensive, the Council might then decide that the military details need not be submitted for registration.Footnote 55
Hurst noted that the Covenant's drafters had probably intended to oblige states to register all treaties—but that they may not have considered “the inconvenience which a strict construction of the Article would entail, or the difficulty of enforcing it.” Having regard to these matters, and the wording of the French version,Footnote 56 he suggested Balfour's approach would be acceptable.Footnote 57 This reflected the British position throughout the ensuing debates: the FO realized the systemic importance of Article 18, yet could generate no really principled limitations to it, and ended up making pragmatic efforts to tailor the provision to national needs.
The French delegation to the League was instructed to press at Geneva for an interpretation of Article 18 as covering any “convention, treaty or accord constituting … a state-to-state engagement, that is to say, of a diplomatic order,” but not “contracts of a commercial or administrative or technical order.”Footnote 58 Britain did not commit to supporting precisely this position, but the two leading League members agreed prior to the League session that Article 18 could be qualified in some way,Footnote 59 even if British instructions still urged delegates to “avoid the impression of encouraging ‘secret diplomacy.’”Footnote 60
Deliberations Within the League, Jurisprudence and Scholarship
The Secretariat was in a delicate position. On one hand, the need to demonstrate that the League was useful to French security rendered it unthinkable to press for registration of what France refused to publish. On the other hand, the norm of publication was popular and highly symbolic. After the Versailles settlement, which was seen by many internationalists as compromising principles of self-determination, and burdening Germany with a victors' peace, Article 18 was one of the few idealistic provisions carried intact into the League order. The Secretariat was keen to maintain the faith of internationalist supporters by upholding the provision. The Secretariat worked assiduously behind the scenes to hasten submission of treaties,Footnote 61 and Juan Teixidor y Sanchez, the Secretariat official principally responsible for treaty registration, was frustrated to see “[t]he Press … airing opinions of all kinds on the influence that the coming into force of such a [c]onvention, without being registered, or being half-registered … may have on the maintenance of the Covenant.”Footnote 62
Teixidor rationalized that the exchange of letters could be registered and the “technical” (military) text withheld, analogizing the situation to other treaties involving both an agreement between countries and a regulation established by technicians for its execution. However, this analogy did not capture the nature of the texts in issue, and even Teixidor himself noted the need to ensure that the technical regulation did not run counter to the provisions of the League Covenant (something which sat uneasily with registering only the exchange of letters).Footnote 63
Eric Drummond, the League's secretary-general, and a former FO official, was skeptical:
If the Franco-Belgian Convention is made up of two separate instruments, the first, what may be called the Diplomatic instrument, and the 2nd solely the technical means by which the 1st will be effected, then of course there could be no objection to No. 1 being registered, while No. 2 was not. … If, however, the one Convention includes both the Diplomatic and Military sides, I do not think it possible to register one half of it and omit the other.Footnote 64
While approaches within foreign ministries and the Secretariat up to this point had tried either to exempt certain accords from the class of treaties and engagements to which Article 18 applied, or to withhold “technical” aspects of arrangements, Drummond settled instead on something more radical. Echoing earlier Council discussions, he declared that “if countries choose to take the risk of concluding Conventions which will not be recognised as binding by the other Members of the League, I have always felt they were entitled to do so.”Footnote 65 Ultimately, the exchange of letters was registered without complaint.Footnote 66
Dionisio Anzilotti (then an under-secretary-general; and from 1921 a judge on the Permanent Court of International Justice (PCIJ)) articulated the same view, but with greater legal sophistication. He conceded that “the background of Article 18 … is the principle of open diplomacy” and that “the non validity of unregistered treaties was, in all probability, conceived as a sanction intended to enforce the publication rule … .” Footnote 67 However, he argued that “an objective consideration of the provision in itself, with regard to the circumstances from which it arose and to the place it has in the system of rules of which it forms a part” left open a much more minimalist reading of what it required. He argued:
Before the coming into force of the Covenant, the binding force of international treaties only rested upon the moral force of the promise entered into by the contracting parties. … The Covenant … has added to the purely moral sanctions other positive sanctions consisting of the coming into movement of the machinery of the League. It seems therefore quite natural to me that Article 18 of the Covenant be understood in connection with the said sanctions and not otherwise. … if the parties agree to be obliged by a non-registered treaty, the Covenant has nothing to do with it: the treaty remains beyond the sphere of the League. … The Members of the League, therefore, are not unconditionally obliged to register treaties or international engagements entered into by them; they are obliged only if they wish to put the said engagements under the protection of the League.Footnote 68
The early vision of Article 18, reflected in the 1920 Secretariat Memorandum, had understood it as imposing a new condition of publicity on the creation of legal obligation. The Anzilotti/Drummond view reduced Article 18 to a condition merely of access to the particular enforcement mechanisms associated with League membership.Footnote 69 On the latter view, governments retained discretion to move at will between the pre-League order and the League regime. As internal British debates had acknowledged, reading Article 18 in this manner left little role for publicity in ensuring democratic control over foreign policy, or in breaking the cycle of suspicion and belligerence over military arrangements. As Teixidor had noted, it also removed the possibility of ensuring compliance with states' undertaking in Article 20 of the Covenant not to enter into engagements inconsistent with the Covenant (and limited opportunities for the Assembly to advise, under Article 19, on reconsideration of treaties in changing conditions).Footnote 70 Ironically, the Secretariat seems to have embraced a limited reading of Article 18 at least partly because of the pressure of public opinion. The minimalist interpretation avoided both the “patent evasion of the law” inherent in registering only the exchange of letters, but also the appearance of “the weakness and the lack of viability of the League” which would result from a showdown between the Secretariat and the French government.Footnote 71
Nevertheless, these internal Secretariat deliberations did bring to the surface a fundamental difficulty. The view that all treaties should be published, and particularly the Article 18 scheme for registration and publication through the Secretariat, was in some tension with the existing law of treaties. This tension emerged more clearly as the Dutch representative, alarmed at the apparent Franco-Belgian challenge to Article 18, proposed in the Assembly that the Council appoint a committee of jurists to inquire into the provision. The representative, Jonkheer van Karnebeek, outlined three possible interpretations of the status of an unregistered treaty in light of Article 18:Footnote 72
-
(1) States were genuinely not bound by a treaty pending registration.
-
(2) States were bound by an unregistered treaty from the moment of ratification, but could not demand execution of the obligations pending registration.
-
(3) States were bound from the moment of ratification regardless of registration, but precluded from relying on the treaty when appealing to the League unless the treaty had been registered.
None of the three interpretations was entirely convincing. On the first interpretation, a ratified but unregistered treaty had, at best, the status of a pre-Covenant treaty which had been signed but not ratified. This approximated the strong reading of Article 18 reflected in the 1920 Secretariat Memorandum, and shared by most internationalists familiar with the provision. It was supported by van Karnebeek as conforming most closely to the language of the provision,Footnote 73 and embraced also by Giuseppe Motta (a Swiss delegate), who noted that the popular campaign conducted in Switzerland in favor of joining the League had presented Article 18 as “one of the pillars of the new [League] edifice.”Footnote 74 However, this first interpretation involved a major shift from the statist, voluntarist creation of legal obligations which had long been central to the law of treaties to a more centralized system in which treaties would have legal force only after the Secretariat's action in registering them.
The second interpretation would mitigate the severity of this shift by preserving the significance of the parties' intention to be bound, and inserting a further stage prior to full entry into force. But this “suspensive” reading, which was perhaps a better fit with the French than the English text,Footnote 75 appeared to emanate from civil law doctrines applicable to private rights. The interim stage, in which parties were bound but could not demand execution, seemed to some commentators a “purely verbal” construction.Footnote 76
The third interpretation was close to the view articulated by Anzilotti and Drummond (although Anzilotti and Drummond were more reluctant to concede the existence of a legal obligation prior to registration). In the Assembly, it attracted support from the politician and former diplomat Tommaso Tittoni, an Italian delegate.Footnote 77 This interpretation was consistent with existing doctrine on the effect of ratification, and minimized the risk of embarrassing confrontations over non-compliance. However, this interpretation focused on the second sentence of Article 18, while virtually ignoring the opening clause (“Every treaty or international engagement … shall be forthwith registered”). In downplaying the obligation to submit treaties for registration, and leaving governments a choice to make secret treaties provided they did not seek to rely on them before the League, it drained Article 18 of much of its anticipated effect in strengthening democratic control of foreign policy, and in removing suspicions, associated with secrecy, which threatened peace and stability.Footnote 78
From a systemic point of view, perhaps only the first interpretation could have been transformative. The requirement would not have attracted universal compliance—as evidenced by the Franco-Belgian case—but if upheld as a comprehensive norm, it might gradually have caused governments to doubt the reliability of any unpublished treaty, and thus increased incentives to make important undertakings in texts that would be registered and published. However, this kind of systemic argument rarely surfaced in the League debates.
As new scrutiny of Article 18 loomed, the British brought their own challenge into the open. The government wrote to the secretary-general, in correspondence then circulated to all League members, that it had entered into “a large number of financial arrangements, many of them of small general importance … with a view to completing and liquidating the abnormal transactions rendered inevitable by the war,” which the government “[did] not conceive … were in any way analogous to the Treaties contemplated by Article 18 … .” The letter announced the government's view that it would be “unnecessary, and in many cases inexpedient” to publish the detail of such transactions.Footnote 79
By late 1920, the British Treasury had clarified that it did not wish to register: (a) agreements with allies for the settlement of claims relating to goods and services supplied during the war; (b) agreements relating to money advances by or to Britain; and (c) agreements for the sale of surplus material. The Treasury believed that texts in categories (a) and (c) were not treaties or international engagements within the meaning of Article 18, but acknowledged that category (b) presented difficulties. These agreements came in a variety of forms. Some might well be treaties or international engagements. However, the Treasury insisted, if these texts were required to be registered and published,
it might well become impossible for any nation to come to the assistance financially of another nation in an emergency … simultaneous publication [to] the legislative assemblies of the countries concerned … might give rise to serious political difficulty … [or] entail grave risks to the financial credit of the borrowing country.
Moreover, publication of an agreement postponing payment of debts or reducing rates of interest would no doubt produce analogous requests from other debtors “which it might, on political grounds, be impossible to resist.”Footnote 80
In a cursory opinion, the British Law Officers (attorney general and solicitor general) advised the Treasury that it could withhold from registration categories (a) and (c), and any agreements in category (b) reached through other than “diplomatic” channels.Footnote 81 The opinion made no attempt to grapple with either the interpretation of “treaty or international engagement” in Article 18, or the definition of a treaty under international law.
In Geneva, van Karnebeek succeeded in his initiative for a committee of jurists to examine “the scope of Article 18 from a legal point of view” and “prepare … relevant proposals to secure a uniform application of the article in respect of the validity of Treaties.” Footnote 82 This committee (hereinafter Article 18 Committee) consisted mainly of legal advisers to European foreign ministries, and seemed set to provide the in-depth legal analysis which had not been undertaken on adoption of Article 18. Yet the Committee also struggled to find a principled reading of Article 18 which would accommodate states' desire for a residual possibility of secrecy. The Committee's report was fractured, indicating that the report was a product of compromise, or simply that there were no attractive legal answers. The Committee recognized the aspirations for structural change inherent in Article 18, noting that the “fundamental aim” of the provision was to “suppress secret Treaties, which in the political sphere constitute a danger to the maintenance of peace” and to secure “a broad system of publicity for international engagements.” The report offered a relatively strong reading of the existing text of Article 18. It rejected both the Anzilotti/Drummond view that registration was not actually obligatory at all (the third interpretation in the schema above), and the approach taken by van Hamel and Teixidor on the Franco-Belgian military accord, whereby registration of one element of an agreement was considered sufficient compliance with Article 18. It also rejected exceptions like those which France, Belgium, and Britain had purported to find in Article 18, based on the nature, form, importance or duration of international engagements.Footnote 83
Instead the Article 18 Committee devised an interpretation similar to the “suspensive” reading (the second interpretation in the schema above),Footnote 84 but asserted that even this position entailed “impracticable and dangerous” consequences. The report laid out a range of avenues for amending the provision, including: deleting the second sentence of Article 18 altogether; retaining the second sentence but giving registration retroactive effect by changing “unless” to “until”;Footnote 85 inserting an “additional article” reflecting the Anzilotti/Drummond approach (the third interpretation in the schema above);Footnote 86 and adding to the obligation to register agreements the words “subject to such provisions as the Assembly may unanimously decide upon,” thereby enabling the Assembly to exclude categories of treaties, such as “technical” instruments.Footnote 87
The Article 18 Committee's report triggered, in a series of League committees, the same questions which had dogged the Article 18 Committee's own proceedings. A Sub-Committee of the League's First Committee betrayed some unease about any major qualification to Article 18 as written, and modified the Article 18 Committee's suggested Assembly regulations to require that, even for administrative or technical agreements, the parties submit basic details of the agreement.Footnote 88 In the First Committee, the British fought back, insisting that publication of such details “would in itself be sufficient to produce all the disastrous results which might flow from publication in full.”Footnote 89 The lawyer and diplomat Raul Fernandes (Brazil) disagreed; if an agreement opened a credit, there could be no disadvantage in announcing this; indeed “the publicity given to the agreement would constitute an element in the world's tranquillity.”Footnote 90 Hurst suggested the summary registration procedure would provide an uncontrolled loophole. Fernandes retorted that the effect of Article 18 would not be weakened by summary registration but by failure to register altogether.Footnote 91 Balfour returned to the line of attack he had pursued earlier. He argued that the framers of Article 18 had intended “to put a stop … to those arrangements which were fundamentally aggressive in their character,” and contrasted treaties of this kind with “the small, transitory, but very necessary, arrangements” which Britain sought to keep secret. He reiterated that “[c]redit was very sensitive, and very easily shattered by public announcements which might be used or misused by interested parties.”Footnote 92 The British ultimately prevailed, and the First Committee returned to the original recommendations of the Article 18 Committee.Footnote 93
These highly technical debates went largely unnoticed by public campaigners against “secret treaties.” In an effort to attract attention, Konni Zilliacus, a committed internationalist in the League's Information Section, passed the Article 18 Committee report, with an explanation of its implications, to the liberal Manchester Guardian. The editor printed parts of the report, suggesting that it undermined the core of Article 18,Footnote 94 and sent the material to the pro-League British politician Robert Cecil, then representing South Africa at the League.Footnote 95 This may have helped mobilize League of Nations Union and Liberal Party support for the preservation of Article 18, and armed Cecil for debate in the plenary Assembly.Footnote 96
In the Assembly, Cecil protested the narrowing of Article 18, and embraced the point made by others, that it was impossible to distinguish between “political” and “technical” treaties.Footnote 97 The Article 18 Committee itself broke ranks, with the Italian legal adviser announcing a personal view that the current language of Article 18 could be interpreted even without express qualification to require only the registration of treaties or even lesser engagements “which may engage the future fate of nations.”Footnote 98
Debates at the League were frustratingly circular and shallow. No one pointed out that even if the ultimate end of Article 18 had been “put a stop … to those arrangements which were fundamentally aggressive in their character,”Footnote 99 this could not be achieved by insisting that states only needed to register these kinds of sensitive treaties. If anything, this would have made compliance with Article 18 even more unlikely. Only a requirement making registration and thence publication an undisputed threshold for any commitments having legal force had any real prospect of incentivizing governments otherwise tempted to secrecy to make politically significant commitments in texts that would be subject to registration and publication.
Those supporting the Article 18 Committee approach, led by British representatives, ultimately retreated from the attempt to achieve formal amendment of the Covenant, and sought to postpone debate to 1922, subject to an understanding that “in the meantime, Members … are at liberty to interpret their obligations under Article 18 in conformity with the proposed amendment.”Footnote 100 The latter part of the resolution was technically not adopted,Footnote 101 but the general tenor of the debates tended to add to accumulated doubts about the scope of Article 18.
As it turned out, further substantive consideration was deferred indefinitely,Footnote 102 and a “silence” fell over Article 18.Footnote 103 The Secretariat declined to take any public position on whether particular texts were required to be registered. It did keep regular lists, sourced from the press, of treaties apparently made but not yet registered, but this information was not used to shame recalcitrant governments.Footnote 104 After a few initial attempts, Drummond largely ceased writing personally to foreign ministry contacts to press for compliance.Footnote 105
Debate in the wider legal community reflected the same divergence of views as existed within League bodies. A report on Article 18 presented to the Institut de Droit International as part of the Institut's ongoing consideration of the Covenant proposed an amendment to Article 18 broadly similar to the one which had finally emerged from the League debates, but which had never been formally approved (i.e., providing that non-registration would merely suspend the parties' obligations, and excepting large classes of instruments from even this narrow provision).Footnote 106 Comments from Institut members ranged from opposition to any compromise of the universal and absolute quality of Article 18, to assertions that, if there was to be a change to Article 18, the provision might just as well dispense with registration for all treaties whose publication appeared inopportune or dangerous to the signatories.Footnote 107 Interwar efforts at codification of the law of treaties were also divided.Footnote 108 Publicists generally reiterated the history of debates within the League organs, and noted that states had fallen short of rigorous compliance.Footnote 109 Perhaps most significantly, the PCIJ (possibly influenced by Anzilotti) seems not to have raised Article 18 in instances in which the PCIJ adjudicated matters related to unregistered treaties.Footnote 110
Assessing the Renegotiation of Article 18
Debates in the League and the wider legal community left the terms of Article 18 intact, but introduced informal qualifications to its interpretation, and indicated that it would not be strongly enforced. This tacit renegotiation of Article 18 was driven in part by realpolitik concerns: governments' perception of a practical need for secrecy, and the Secretariat's difficulties in enforcing compliance.Footnote 111 However, one can also see in the various League bodies doubts at the level of principle, particularly concerning financial relations: was secrecy always inimical to peaceful relations, or was it, in fact, necessary to maintain them?
These early responses to Article 18 challenged the intuition that publicity and legal force went hand in hand. Much depended on the vision of law at stake. The 1920 Secretariat Memorandum had sought to entrench publicity as a condition of legal force, and this approach, which would have cast doubt on the reliability of secret undertakings, was probably the only path by which Article 18 could have changed behavior. However, this approach was premised on a major shift in international law, in which the League Secretariat held at least a formal role in bringing treaties into operation. This was difficult to reconcile with then-dominant understandings of international law, in which states decided both what treaty obligations to accept, and whether or not they needed to be published. As one jurist put it, Article 18 was “a case in which the generous impulse towards a more perfect international organization runs up against obstacles deriving from a deep-rooted status quo, and in which the principle stated does not correspond to the collective legal consciousness.”Footnote 112
Early struggles over the interpretation of Article 18 showed the law itself to be a complex instrument, rather than a neutral means of transforming the interstate order. Not only did the deep structure of the law of treaties weigh against making registration and publication a condition of binding force; but, as the Franco-Belgian accord showed, the complex vocabulary of treaties, together with crafts of drafting, created avenues for secrecy. Moreover, the transposition of political condemnation of “secret treaties” into the realm of legal obligation arguably blunted the impetus for change. What had appeared a clear objective—all treaties must be public—became entangled in debates that were so complex and technical as to dissuade public engagement.
III. Interwar Experiments: Entrenchment of Publicity, Refinement of Secrecy
Domestic Law and Bureaucratic Practice
Once the force of Article 18 had been called into question, domestic law constituted the major constraint on treaty practice.Footnote 113 In France and the United States, there were no major changes to the constitutional position of the pre-war years. In Britain, on the other hand, the first Labour government sought in 1924 to introduce a constitutional requirement for parliamentary approval, or at least opportunity for discussion, of all treaties prior to ratification.Footnote 114 Cecil Hurst, the FO legal adviser, discouraged any move toward requiring parliamentary approval, and suggested that publicity, and thus some measure of democratic control, could be accomplished by simply laying treaties once signed, instead of after ratification.Footnote 115 The new policy was signaled by a statement by the foreign secretary in the House of Commons, that “every Treaty” would be laid, when signed, for twenty-one sitting days, before ratification.Footnote 116 After a brief suspension when the Conservatives returned to power (1924–29), this “Ponsonby rule” was gradually entrenched as a convention.Footnote 117 However, the focus on treaties requiring ratification may have obscured the non-publication of treaties not subject to ratification. Hurst had believed that such “international agreements, usually of a technical or administrative character … do not as a rule raise political issues, nor do the public at large or the Legislature desire to be bothered with them,”Footnote 118 but it was not always the case that such treaties were devoid of political significance.
In all three states, approaches to registration appear to have been shaped not only by domestic law, but by bureaucratic structures and processes. As the United States was never a member of the League, U.S. publication practice was relatively unaffected by Article 18. At least some “executive agreements” were considered treaties capable of binding the state under international law,Footnote 119 and a new Executive Agreements Series was launched alongside the Treaty Series in 1929, but some executive agreements were not published.Footnote 120 Manley Hudson, who worked in the League's Legal Section during U.S. summers even after leaving full-time service in 1921, brokered an arrangement in 1925 whereby the United States would send treaties appearing in the Treaty Series to the League Secretariat, to be “entered” (but not registered). When another signatory was a League member, the Secretariat would seek submission from them of the text for actual registration, and only later, if no submission was forthcoming from the member, print the text provided by the United States in the League of Nations Treaty Series (LNTS), under a system of numbering distinct from that used in the main series.Footnote 121 This procedure delayed the publication in the LNTS of U.S. treaties. From 1934, again following Hudson's activism,Footnote 122 the United States agreed to submit treaties for registration like any League member, but on the condition that the United States was not thereby acquiescing in the stipulation of Article 18 that no instrument be binding until registration.Footnote 123 Although the 1934 arrangements were not explicitly limited to treaties already published in the Treaty Series or Executive Agreements Series, there is no indication that officials contemplated sending to the Secretariat any texts which they would not otherwise have published.
Within the British Foreign Office, the Treaty Department was responsible for both the integrity of internal records and the transmission of texts as appropriate to Parliament and the League. These tasks were mutually reinforcing, and officials had a certain pride in the accuracy of their work, making them (usually) voices in favor of registration. In the French MAE, on the other hand, there was no ongoing treaty series akin to those in the United States and United Kingdom (treaties requiring promulgation simply appeared in the general Journal Officiel), and this may have been both a symptom and a cause of relatively limited attention to publication. Moreover, the task of registration was given initially to a new League liaison service within the ministry, rather than the services usually responsible for the custody of treaties (Protocol and Archives), and this dispersal of roles may have contributed to the marginalization of registration. In 1928, it was conceded that “until now, and it is a matter which Mr Fromageot [the legal adviser] has raised several times, we have often neglected to complete this formality of registration.”Footnote 124 It was only in 1928 that it was laid down that, once the decree of promulgation had been published in the Journal Officiel, the treaty would be sent for registration. Texts not published in the Journal Officiel, of which there were apparently a considerable quantity (see further below), were not to be sent for registration, other than on specific instructions.Footnote 125
Techniques of Secrecy
The interwar period saw common techniques in Britain and France for the preservation of secrecy where this was considered desirable, including recourse to texts of ambiguous status, and packages of layered texts. However, there were differences in national styles. The French MAE was more inclined to rely on a general exception for “technical” texts, which was highly flexible, albeit legally plausible insofar as it reflected gradations of importance and detail that had long lain beneath the positivist notion of treaty.
In Britain, the Law Officers' 1921 opinion, mentioned above, articulated exceptions to Article 18 on which the Treasury and Board of Trade leaned heavily.Footnote 126 Many agreements settling claims, or for the sale of surplus material, were not registered or published in the Treaty Series, although these may have been regarded as outside the scope of treaties in any event. A large number of borderline texts concerning administration of reparations, not actually referred to in the opinion, were treated as analogous.Footnote 127
The FO tolerated secrecy in other circumstances, accommodating it through texts of indeterminate status, or the old drafting technique of splitting commitments between a (published) principal text and a (secret) ancillary text. The former approach is exemplified by a 1927 “Record of Conversations” between the British ambassador in Rome; the governor of the Italian colony of Eritrea; and the director-general of the Italian Ministry for Foreign Affairs.Footnote 128 This rather convoluted text, referred to later as the “Rome Understanding,” stated that the delegates:
acting under their instructions that the respective policies of the two Governments in Southern Arabia and the Red Sea should be pursued in a spirit of friendly co-operation and developed on parallel though independent lines, have reached the conclusion that their Governments are in agreement on the following points … .Footnote 129
One clause recorded that “His Britannic Majesty's Government regard it as a vital imperial interest that no European Power should establish itself on the Arabian shore of the Red Sea” and that Italy, in view of possessions on the west coast of the Red Sea, had an equivalent interest.Footnote 130
When the signed texts arrived in the FO, the initial instinct was to submit them for registration, but some doubted whether they constituted an “international engagement.”Footnote 131 The secretary of state preferred not to submit them, on the basis that this “might create an undesirable precedent.”Footnote 132 This decision was received in the Treaty Department as a legitimate reading of ambiguous texts in circumstances where the line between treaty and “gentleman's agreement” was hard to draw. Yet when the FO looked again at the texts a decade later, as an expansionist Italian policy in the Yemen seemed more threatening, the earlier judgments seemed questionable. It was feared that any public assertion of British interests would ultimately lead to discovery by members of Parliament (MPs) of something which they “might regard as a secret agreement.”Footnote 133 On one hand, it was argued internally that
there is nothing “secret” about the Understanding in this sense. It was a purely informal agreement, in the nature of diplomatic discussions and correspondence, and at the time there seemed no more reason to publish it than to publish any other part of the correspondence passing between His Majesty's Embassy and the Italian Government.Footnote 134
But this was not quite true. From the outset the texts had struck some officials as having the quality of an “international engagement.” The formality of two languages, the finality of the statement of a mutually shared position, and the significance of the content had seemed to indicate a legally binding commitment. On the other hand, the carefully chosen wording (“parallel though independent policies”) and the “conclusion that [the] Governments are in agreement” tended to suggest a mere statement of fact about the governments' beliefs. Regardless of whether the failure to submit the texts for registration had been technically defensible, officials knew that fine distinctions between engagements and gentlemen's agreements would not be intelligible or convincing to Parliament.Footnote 135
An alternative technique for accommodating secrecy involved splitting obligations across a (published) principal and (secret) ancillary text. Where issues were important, and the government feared that the other party might renege on anything agreed in secret, officials insisted on publication of ancillary texts, reasserting a relationship between publicity and legal force.Footnote 136 On other occasions, the United Kingdom did accept secret ancillary texts, the terms of which had some claim to binding force through their relation to the principal text.
Secret ancillary texts were particularly prevalent in relations with polities which, while nominally independent, were within a British sphere of influence. While the Anglo-Iraqi Treaty of Alliance of 1930, for example, was registered and published, it was nestled among exchanges of notes and side letters “interpreting” the treaty, only some of which were published.Footnote 137 The Anglo-Iraqi Treaty was a prelude to transition from British mandate to a new form of internationally brokered “independence” as informal empire,Footnote 138 and the notes and letters were often crafted to preserve an appearance of independence and symbolic prestige.
One note concerned a special guard, consisting of Iraqi troops under British officers, for the protection of British air bases in Iraq (British air force presence having been a major source of tension in negotiations).Footnote 139 It was understood that the note would not be registered or published, but the Embassy in Baghdad contended that
[i]t was … made perfectly clear at the time that Nuri's note, or secret letter as it is generally called, would be binding on the Iraqi Government as faithfully describing “such conditions as may be agreed to between the High Contracting Parties” [referred to in para. 4] of the Annexure to the Treaty of 1930.Footnote 140
The secret text was thus not legally binding in its own right, but provided further details of a binding and public obligation into which it was arguably merged.
Four years after the signature of the Anglo-Iraqi alliance, a new Iraqi government sought to publish some of the 1930 correspondence. As with the Anglo-Italian “record of conversations” discussed above, looking with fresh eyes at the correspondence prompted questions. An official observed that “the dividing line between some of these notes … and an ‘international engagement’ which should be registered … is rather thin.”Footnote 141 Fitzmaurice, the assistant legal adviser, noted that “it would obviously be misleading” to submit the treaty for registration without accompanying notes if the notes “place a special meaning on [a] clause … different from that which an ordinary reader might assume from the terms of the treaty itself.”Footnote 142 It was decided, however, that it was better not to register late, and to simply “brazen out” the earlier failure to register if it came to light.Footnote 143
The use of ancillary texts as an avenue for secrecy was also rife in the trade context, particularly as protectionist measures in the 1930s exacerbated struggles for commercial advantage. The FO thought “the growing [Board of Trade] habit of having their agreements in two parts, the innocent one for publication and the less innocent to be kept secret” was “reminiscent of the secret treaties which we have been trying to get away from since the war,”Footnote 144 and objected to the French practice of withholding agreements “by just labelling them ‘administrative agreements.’”Footnote 145 However, the Board of Trade argued that secrecy
has become so widespread in other countries that we have found it impossible to avoid the infection … most of our trade agreements contain confidential letters … perhaps the chef d'oeuvre is the Polish Agreement which contains more confidential than public matter!Footnote 146
The Board of Trade in turn felt inhibited from complaint about other countries' practices because “their own consciences are not entirely clear.”Footnote 147
In Britain, then, possibilities for secrecy subsisted, but secrecy tended to be accommodated only for matters of apparently secondary importance (however politically salient in reality). It was only as the international situation darkened in the late 1930s—a time when fascist powers were entering into secret treaties foreshadowing warFootnote 148 —that Britain resorted to secret commitments in the realm of high politics.Footnote 149
The French MAE undoubtedly left unpublished and unregistered many of the sorts of reparations, trade and financial agreements which also escaped publication in the United Kingdom,Footnote 150 but went further, keeping secret a series of “military accords” made between France, on one hand, and the “Little Entente” (Czechoslovakia, Romania, and Yugoslavia), on the other. These military arrangements, aimed at reinforcing the continental status quo and fostering military collaboration, were more obviously politically salient.
Drafting followed loosely the precedent of the Franco-Belgian military accord, with one or more public texts (often termed “political”), and military stipulations in an ancillary text (sometimes described as “technical”) not intended for publication or registration. In 1921, for example, France and Poland entered into a “political accord” and a “military accord.” The political accord was signed by the respective foreign ministers; the military accord by ministers of war; and an annex to the military accord by generals and (in the French case) the minister of finance.Footnote 151 The political accord contained very general provisions committing France and Poland to cooperation in matters of foreign policy and economic recovery, and providing that if one party suffered an unprovoked attack, the two governments would cooperate for the defense of their territory and legitimate interests. The military accord, styled as being “in execution of” the political accord, contained more concrete commitments to cooperate should Germany or Russia take particular steps, and provisions requiring Poland to develop its national forces, including by instituting two-year compulsory military service, and maintaining a peacetime strength of a certain number of divisions.Footnote 152 France undertook to contribute military credits to the sum specified in the annex. The general staffs undertook to keep each other informed of preparations and measures necessary to the execution of the accord and the maintenance of lines of communication between the countries.
The political accord was to be published almost in its entirety.Footnote 153 The military accord and its annex were not to be published at all.Footnote 154 Despite the variation in signatories, it seems to have been assumed that the texts all bound the French state as such (rather than ministers or generals personally). Internally, the French analogized this case to the Franco-Belgian episode, in which registration of only the covering letters had been accepted.Footnote 155 The League Secretariat registered the political and commercial accords without posing any questions.Footnote 156 A similar pattern, albeit with less firm and detailed military components, was followed for arrangements between France and other members of the Little Entente.Footnote 157
Internal French documents suggest that diplomats and MAE officials understood the secret military accords and other ancillary texts to have legal force, as far as they went. It was occasionally noted that the Franco-Polish annex on military credits was an agreement of a kind which, under the constitution, was not valid unless approved by the legislature,Footnote 158 but there was no attempt to grapple with this fact. Admittedly, officials conceived of obligations in a holistic way, and rarely tried to disentangle those having force on legal grounds from those whose force was derived from, for example, a moral requirement to adhere to undertakings once made, or considerations of national self-interest. There was not scrupulous compliance with all terms of the secret texts, any more than the public ones. Nevertheless, officials invoked secret texts as a source of obligations in internal correspondence, and discussions with foreign counterparts.Footnote 159 Internal lists of France's treaties, alliances, and “contractual obligations” referred to both public and secret texts, mentioning which had been published and which had been secret but not otherwise suggesting that the secret texts lacked legal force.Footnote 160 In a draft speech for the French foreign minister, which originally claimed that French relations with friendly countries of East and Southeastern Europe “remain founded on public accords which do not threaten anyone and in no way limit the development of peaceful relations between their signatories and the other States of Europe,” the word “public” was carefully struck out by hand.Footnote 161 Other countries, including Britain, knew or suspected that the secret texts existed, and incorporated them into assessments of relations between the various countries.Footnote 162 This might merely reflect the insignificance of distinctions between legal and extra-legal obligation, particularly in the military context, but it also suggests that no sharp distinction was made between the legal force of public and secret texts.
Despite the elaborate drafting, the existence of military commitments did not actually remain secret in any of these cases. The undertakings made in the Franco-Polish military accord required domestic legislative action in both states (in Poland, the institution of two years' compulsory service, and in France, the grant of military credits) that was difficult to reconcile with secrecy.Footnote 163 The French Senate Finance Committee demanded a copy of what it believed was a “financial convention” underpinning the military credits, but was actually the military accord and its annex.Footnote 164 The Polish minister in London observed that “this accord had been drafted most clumsily,” and proposed replacing it with a text which would mention a loan without specifying the military aspect.Footnote 165 Further deliberation was short-circuited when the Frankfurt Gazette published a summary of the (public) political and (secret) military accord. The conclusion of clusters of public and secret instruments with Czechoslovakia and Romania were also followed by press rumors or disclosures. In responding to press and public criticism, statesmen and officials could exploit the ambiguous status of ancillary, secret texts, and shift between a broad lay notion of “treaty” and narrower senses common in diplomatic discourse, but not widely shared by the public.Footnote 166
The Interwar Legacy
Article 18, as the first norm of treaty publication in international law, did have considerable effects. Most League members submitted at least some of their international engagements for registration,Footnote 167 and the rhetorical embrace of publicity inherent in Article 18 gave domestic legislatures and publics a new benchmark against which to criticize secrecy. Nevertheless, Article 18 was not able to achieve the definitive split it promised on its face, between a universe of legally binding obligations, all of which would be registered and published; and a hinterland of secret obligations which, whatever their hold in interest or honor, had no legal significance.
Experience in the 1920s and 1930s suggested, even to proponents of publicity, that the connections between publicity and the ideals it had been intended to serve were more attenuated than anticipated. While perhaps necessary for democratic control of foreign policy, the norm of treaty publication was far from sufficient to secure this end. As even Ponsonby came to acknowledge, opportunities for debate in Parliament were rarely taken up.Footnote 168 The relation between publicity and peace had already been questioned in the 1920s, when secrecy about some interstate financial dealings seemed conducive rather than damaging to peace, and in the 1930s, secrecy of military accords came to seem, at least to the “status quo” powers who had emerged victorious from WWI, essential for defense and the stability of the international order.
The relationship between the norm of publication and the fostering of legality in interstate relations was also highly complex. The basic intuition that legal obligations were, or should be, public had gone hand-in-hand with a new, institutionalized vision of international law. This had been countered by a voluntarist tradition in which states remained the ultimate arbiters of their obligations to each other, and the extent to which these obligations would be rendered public.
The vision that the norm of treaty publication would change the conditions of interstate interaction, and help to construct a more law-governed international order, also came under pressure. The norm probably did contribute to the wider dissemination, and possibly thus the systematization and coherence, of international law, by establishing the League of Nations Treaty Series as a unitary repository of authenticated and translated texts. But whether it entrenched a deeper orientation to legality in governments was difficult to say. The fact that some secret treaties made arrangements for conduct in violation of international law tends to reinforce the notion that secrecy and power politics went together, in opposition to publicity and respect for law. On the other hand, many secret treaties did not contemplate action in violation of law. Indeed, precisely because they were secret, and signatories lacked an institutional and rhetorical apparatus for their enforcement, some secret treaties might actually be seen as a mark of faith in the power of legal obligation alone to guide states' conduct.
There are indications that the norm of publication empowered a broader range of states as legal actors, vindicating formal sovereign equality over informal hierarchy. One striking illustration is furnished by dealings in the 1920s between Britain, France, and Italy regarding respective rights in Ethiopia. Frustration on the British side with Ethiopian reluctance to accept a dam project essential for development in the Sudan drove Britain to enter into a 1926 exchange of notes with Italy, whereby each would support the other in seeking desired concessions from the Ethiopian government. The regent of Ethiopia, Ras Tafari, heard of the agreement, and protested that he “[did] not understand why the two Governments [had] contracted together to consider action inside the jurisdiction of Ethiopia before we heard anything about it.”Footnote 169 The FO and Italian counterparts toyed with various possibilities, including splitting the agreement into secret and public parts,Footnote 170 but the necessity of registration and publication, in light of Ethiopia and parliamentary interest in the agreement, made it clear that dissembling over the text itself would be unwise.Footnote 171 The parties settled for only minor alterations before showing the text to Ras Tafari.Footnote 172
In reinforcing national publication practices, Article 18, together with the existence of the League as a platform, made a tangible difference to Ras Tafari's possibilities of resistance. British MPs asked probing questions about the intent behind the exchange and, while their criticisms were often stymied, as the government repeatedly rejected interpretations of the letters by MPs and the press, they drove Chamberlain to issue an explicit denial of any coercive intent.Footnote 173 In July 1926, Ras Tafari sent a letter to the League, duly circulated to League members and published, contrasting his belief that “all nations were to be on a footing of equality within the League, and that their independence was to be universally respected” with the seeming reality of Anglo-Italian collusion. This did not force the withdrawal of the notes altogether,Footnote 174 but caused Britain and Italy to disclaim formally any intention to “reserve” any part of Ethiopian territory.Footnote 175
Historically, Ethiopian rulers had been bound by treaties which they were sometimes unable to read themselves, and European governments had exploited linguistic ambiguities in correspondence to which Ethiopian rulers had had limited access.Footnote 176 League procedures ensured, at least, that the Ethiopian Government would have access to the actual texts, and could contest their terms on a footing of formal equality. On the other hand, it did not change the underlying dynamics of power relations or avert Italian expansion in the long term.
Finally, while Article 18 reflected a strategy of using law to entrench publicity, the law itself continued to complicate the underlying project. Aside from the deep structural concerns which had limited the interpretation of Article 18, diplomatic crafts of drafting and terminological subtleties central to the law of treaties had enabled officials to preserve margins for secrecy without making them obvious to legislatures and the public. Publics in the democracies surveyed here remained deeply concerned about “secret treaties” throughout the interwar period. The revelation of leaked or forged “secret treaties” produced periodic press scandals. Yet there was a diminishing sense that the public might ever discover the truth of what was occurring. Admittedly, the subtleties surrounding understandings of “treaty,” “agreement,” “understanding,” and the like could trap officials as well as confuse publics. Sometimes instruments which were never intended to have legal force attracted suspicion, and the terminological complexity left governments unable to convey the innocent or trivial nature of the transactions at stake. Overall, though, the complexity of the vocabulary probably weakened faith that the true picture could be known, and thus dissuaded public engagement.
IV. The Drafting and Interpretation of Article 102 of the Charter, c. 1945–1949
Although wartime conditions in 1939–45 interrupted the usual course of both international and domestic treaty publication, the United States and the United Kingdom, at least, assumed that the treaty publication requirement would be reintroduced in some form after WWII. Eleven other, mostly small, states raised the issue in their responses to the Dumbarton Oaks proposals.Footnote 177 However, in 1945, general interest and faith in publicity was weaker than it had been a generation earlier. The norm of treaty publication was perhaps associated with an idealism which was seen to have failed. Moreover, the connections between multilateralism, international organization, and publicity, influential in the design of the League, were weakened in the transition to a new international organization. The Security Council was envisaged as wielding centralized military force, which would give it greater powers to police state conduct without relying on publicity, and would require the United Nations itself to maintain regimes of military and intelligence secrecy.
American officials played a decisive role in the crafting of what would become Article 102 of the Charter.Footnote 178 Although figures like Hudson had been consistently supportive of the League of Nations Treaty Series, and keen that it continue in some form, American officials in San Francisco distanced themselves from the Article 18 precedent. They observed that the word “every” in “[e]very treaty or international engagement” had, “by interpretation and in practice, lost most of its significance,” and that the apparent sanction of invalidity for unregistered treaties had been “weakened,” principally by the PCIJ's failure to apply it.Footnote 179 More fundamentally, they believed that the sanction “probably did not achieve the original aim of suppressing agreements inconsistent with the peace of nations.”Footnote 180 The State Department's legal adviser preferred not to include in the Charter any sanction for non-registration.Footnote 181
The U.S. draft, from which the Committee worked, imposed an obligation to register “treaties and international agreements”: an arguably narrower class of texts than might have been caught by the reference to “international engagement” in Article 18.Footnote 182 The draft listed two alternatives for the second paragraph. One was akin to the second sentence of Article 18, and the other was a provision to the effect that organs of the Organization could “disregard” unregistered treaties. Although some small states, including New Zealand and the Philippines, supported the former, a combination of the U.S. proposals, Hudson's interventions, and British support managed to champion the latter, providing that UN organs could choose to disregard unregistered treaties.Footnote 183 This was similar to the Anzilotti/Drummond view of the interwar years. When others raised the difficulties involved in leaving the status of a treaty to ad hoc determination, particularly if the Council and the International Court of Justice (ICJ) took different approaches,Footnote 184 Hudson intervened and expressed the fear that an absolutist provision like Article 18 would be interpreted in light of PCIJ precedent, and thus involve no sanction at all.Footnote 185
Article 102 eventually provided that
Every treaty and every international agreement entered into by any Member … shall as soon as possible be registered with the Secretariat and published by it.
No party to any such treaty or international agreement which has not been registered … may invoke that treaty or agreement before any organ of the United Nations.
Article 102 was, on its face, open to the same interpretive controversy as Article 18: did it impose an absolute obligation to register, or merely stipulate the consequences that would follow from a failure to do so? The first sentence of Article 102 suggested the former interpretation, but some invoked the limited sanction to suggest the latter.Footnote 186 On the latter interpretation, states could create secret obligations which were legally binding inter se, provided they accepted that they could not invoke these obligations in UN fora. Other aspects of the Charter drafting also weakened the Covenant's model of centralized supervision over treaty-making.Footnote 187
By 1945, there had been no further clarification of the terms “treaty” or “international agreement.”Footnote 188 This left a certain ambiguity about the reach of Article 102, even aside from the limited nature of the sanction. At San Francisco, the drafting committee had discussed an explicit exception for agreements between individual states and the Security Council about military forces (then envisaged as part of an international military force which never eventuated).Footnote 189 However, this exception had been abandoned because it was expected that subsequent “regulations” would confine the reach of Article 102.Footnote 190 The FO and State Department contemplated trying to include in the General Assembly resolution concerning implementation of Article 102 some limitation of the classes of texts requiring registration (in the FO case, by transposing some version of the exclusions identified in the Law Officers' opinion),Footnote 191 but this failed, and the resolution ultimately just deferred to “experience and practice” to set its bounds.Footnote 192 In an early concession to the growing role of international organizations as legal actors, that resolution did make clear that “treaty or international agreement” included agreements between UN member states and specialized agencies. The resolution also set out an alternate procedure (“filing and recording”) for certain texts that were deemed not to come within Article 102, notably agreements between the UN and non-member states, the UN and specialized agencies inter se, and the UN or specialized agencies and other inter-governmental bodies.Footnote 193
Importantly, the resolution required a party submitting a treaty for registration to certify that the text was a true and complete copy thereof and included all reservations. The UN Secretariat thus considered that “[i]t was not permissible simply to register a treaty and to withhold any confidential exchange of letters which might, in fact, modify its substance.”Footnote 194 Although an analogous requirement had existed in the League Secretariat's 1920 Memorandum, it had not been made central to debates over Article 18, and the UN Secretariat's position seemed to signal a stricter approach than had been taken to texts like the Franco-Belgian military accord.
In echoes of the interwar experience, however, early compliance with Article 102 was patchy. Attempts to articulate the strongest possible reading of even the limited terms of Article 102, and to foster compliance, were hampered by institutional constraints within the Secretariat (which was unable to confront permanent members of the Security Council), and the fact that even some strong advocates of registration felt vulnerable to attack over their own past practice.
In the Sixth Committee, the USSR seized on an observation by the Committee Chair that Article 102 did not impose an absolute obligation to register, but merely outlined the consequences of not doing so (a position rejected by a number of other states, including the United States, Britain, and France).Footnote 195 The FO contemplated from time to time whether Article 102 might be more energetically enforced, particularly against the USSR.Footnote 196 In 1949 an official in London noted that:
after several battles in the Sixth Committee, it has been agreed to differ, and Article 102 is gradually losing its force. We do not yet say in public, however, that we are not prepared to register all agreements. … It could be added confidentially that we have not recently been beyond reproach ourselves.Footnote 197
Britain reaffirmed its view internally that Article 102 imposed an absolute obligation to register,Footnote 198 but had by this time experimented with new techniques of secrecy (outlined below), and had been implementing some troubling secret agreements, including one requiring the (sometimes forcible) repatriation of liberated Soviet citizens in the United Kingdom.Footnote 199 Though this agreement predated the Charter and was thus not required to be registered, its implementation proved morally disturbing, and created difficulties under domestic law, meaning there were incentives to avoid provoking mention of it by the USSR.Footnote 200
V. Secret Treaties in the Post-1945 World
Domestic Law and Bureaucratic Practice, c. 1945–1990
In the decades following WWII, all three states examined here continued to permit the making of treaties (in the international law sense) which remained unpublished. In the United States, the experience of wartime governance, and the knowledge that the Senate's rejection of the Treaty of Versailles a generation earlier had weakened the League system, motivated arguments that the president had authority to make treaties under international law by a process other than the two-thirds majority approval of the Senate required for “treaties” within the meaning of Article II of the Constitution.Footnote 201 Scholarship and jurisprudence had already recognized “executive agreements” pursuant to an existing treaty or legislation, and in exercise of the president's constitutional powers.Footnote 202 Post-1945 developments would help to articulate more clearly a further variant: agreements subject to expedited approval or implementation, not by the Senate, but by a majority in both Houses, and subsequent approval by the president.Footnote 203 Such “Congressional–executive” agreements, like Article II “treaties,” would necessarily be subject to the publicity associated with the legislative process. Executive agreements pursuant to existing treaties or legislation, or under the president's constitutional powers, on the other hand, while usually published, could be withheld from publication. A State Department circular of 1955 clarified the Department's view of the power to make executive agreements;Footnote 204 reiterated an undertaking by the secretary of state to consult with Congressional leaders and committees, if circumstances permitted, where any serious question arose as to whether a given agreement should be treated as an Article II “treaty,” or as an executive agreement on one basis or another; and mandated certain notifications to and consultations with the State Department in the course of negotiations.
In the United Kingdom, the Ponsonby rule remained the only “constitutional” limit on the Crown's treaty-making powers. This rule did not apply to a large class of treaties made in a form not requiring ratification. These were sometimes laid for twenty-one days, especially where there was public interest, but not in all cases of important agreements.Footnote 205 It remained the practice to publish treaties in the Treaty Series whether they were laid in Parliament or not, but this was not legally required. It was acknowledged internally that “one often finds that politically most important things are done by instruments which come into force on signature only …”Footnote 206 and governments insisted on discretion regarding the choice of form.Footnote 207 As in the interwar period, focus on the Ponsonby rule may have distracted attention from the question of whether treaties not subject to ratification were published at all.Footnote 208
The initial French post-war constitution preserved the requirement for legislative approval of a specific list of treaties, but removed the requirement for the president to inform the legislature of all treaties “[a]s soon as the interest and safety of the state shall permit.” A 1953 decree gave the MAE exclusive responsibility (as against other government departments) for the ratification (if necessary) and publication of all “conventions, accords, protocoles et règlements internationaux” to which France was a signatory, or by which France was engaged. This decree, however, only required publication in the Journal Officiel of instruments of a nature which would affect private rights and duties.Footnote 209 The 1958 Constitution distinguished between “traités” (i.e., in this context, texts subject to ratification or analogous process) and “accords internationaux” (i.e., in this context, international agreements not subject to ratification or analogous process).Footnote 210 It provided, as under earlier constitutions, that the president “negotiate and ratify” traités, and be kept informed of all negotiations leading to the conclusion of an international agreement.Footnote 211 Again, only a closed list of specific kinds of treaties and international agreements required legislative approval and associated publicity. Any treaty not falling within these categories, and not intended to take effect within domestic law, or affecting private rights and duties, might be kept secret by the executive indefinitely.
Familiar Techniques of Secrecy in a Changing World
In some respects, officials navigating publicity requirements after WWII were moving in a world rather different to that of the 1930s. The United States' ascendance to global influence gave the U.S. government new influence over publication practice. The UK, on the other hand, experienced a rapid decline, and sometimes sought the face-saving secrecy she had formerly bestowed on others. The United Nations played a greater role as a site of contestation over compliance than the League had done, in part because Cold War adversaries embraced the United Nations as a platform. While the Soviet government had not complied with Article 102, American celebration of democracy and openness left NATO governments vulnerable to accusations of hypocrisy for any failure to register. Overall, greater international cooperation, including the proliferation of transnational functional cooperation not channeled through foreign ministries, increased the scale of treaty-making and challenged foreign ministries' institutional control over treaty-making processes (a phenomenon reflected in periodic attempts, discussed below, to reassert foreign ministry consultation and oversight).
That said, there were important continuities in the general areas in which secrecy was deemed desirable, particularly trade (especially before the advent of the General Agreement on Tariffs and Trade) and defense (where the development of atomic weapons, and the formalization of military and intelligence cooperation amid Cold War hostilities contributed to an institutionalization of bureaucratic secrecy well beyond the treaty context).Footnote 212 While the scale of post-WWII decolonization would mark a major break from the interwar order, treaty relations between former colonial powers and post-colonial states, exemplified in the French case explored below, often showed an interweaving of public and secret commitments of a kind seen, for example, in interwar dealings between Britain and Iraq.
Where there was a felt need for secrecy, diplomats and officials fell back on some of the techniques seen earlier: crafting packages of texts, some for publication and some not; and maintaining ambiguity about the legal status of particular texts. Persistent questions about legal status arose in particular in connection with recourse to “inter-agency” or “inter-departmental” agreements, concluded by agencies or departments in their own name with foreign counterparts rather than in the name of the relevant governments.Footnote 213 Though acknowledged earlier (for example, by the British Law Officers' reference to agreements not arranged through “diplomatic” channels”),Footnote 214 these agreements proliferated after WWII, creating possibilities for secrecy, but also challenging foreign ministries' control of international commitments in ways that would ultimately demand correction.
As before, British and American officials were inclined to enforce a legalistic approach to what was required to be published, whereas French officials, when they considered registration, seem to have assumed a capacious exception for “administrative” or “technical” agreements (a posture which probably reflected pre-WWII attitudes, but also precarious control of treaty texts following the disruption of the MAE and its records under Vichy and occupation).Footnote 215 In all three states, fine distinctions made internally between different kinds of text, and the nuanced deployment of treaty terminology, were not well understood by legislatures or the public.
For the United States, the key challenge in the early post-war years was reconciling Article 102 with the demands of a global military presence. The State Department resisted attempts to formulate blanket exceptions for agreements made by agencies other than foreign ministries, or dealing with defense and sensitive matters. Instead, officials held out the possibility of “distinguish[ing] between the substance of an agreement and mere administrative arrangements made … in implementation of it,”Footnote 216 albeit on the basis that the latter must not be “integral parts” of the treaty or agreement.Footnote 217
The United States was willing to take the risk that secret treaties could not be invoked before the United Nations, but was concerned about exposure to criticism by the USSR. While it might be argued that Article 3 of the NATO Treaty and announcements of integrated defense forces gave “[t]he peoples of the world … notice, in broad terms, as to what is being done,” and met “the ‘public relations' test,” this did not satisfy Article 102.Footnote 218
A memorandum reviewed possible techniques, and their limitations. Service-to-service agreements (a variant of the “inter-agency” agreements mentioned above) might arguably not constitute treaties or agreements captured by Article 102. However, governments and the U.S. Department of Defense “prefer the additional stability of a government-level agreement.”Footnote 219 Countries might work through bilateral institutions rather than treaties, creating bodies like the United States-Canada Permanent Joint Board on Defense, which would then make plans and decisions functionally equivalent to what might otherwise be dealt with by treaty. But this depended on the states involved having close working relationships. Finally, parties could reach principal agreements in broad terms, with classified annexes. This final option was seen, at least in the U.S. defense context, as preferable.Footnote 220 The United States proposed this to other key NATO powers, and even contemplated making a public announcement of this policy, a course discouraged by French officials.Footnote 221 In practice, some ancillary texts would drift beyond the position prepared for Congress, namely that “not all purely technical or administrative arrangements, such as details on staffing or equipping or manner of operation of bases established in implementation of the [NATO] Treaty, are published.”Footnote 222
Like the State Department, the FO deemed it acceptable to keep secret a strictly delimited set of ancillary texts: “a confidential letter, or agreed minute which explains in greater detail the effect of the published document,” or “embod[ies] understandings which are not intended to constitute international obligations … but … are rather declarations of policy or of attitude.”Footnote 223 Such texts were used in the defense context, sometimes in ways which strained the limits imposed by the FO. A notorious example is the secret exchange of notes between the U.S. ambassador in London and the secretary of state, accompanying the published 1967 agreement regarding U.S. rights on Diego Garcia.Footnote 224 The secret exchange made arrangements for concealment of associated expenses, including for compensation of existing colonies from which the territory was taken, from both the British and American legislatures. These notes “confirm[ed]” the “financial arrangements which have been reached,” by which was understood the position that the UK would meet these costs initially, and then be compensated for half of them by manipulating liabilities under an unrelated agreement concerning the Polaris missile program.Footnote 225
By 1950 the FO had developed a detailed list of exclusions to the definition of treaty, and thus to the set of texts it considered necessary to submit for registration.Footnote 226 Some of these exceptions, like that for unilateral statements, and for agreements operating under municipal law, were congruent with prevailing understandings of the treaty under international law (even if somewhat stretched in practice). Others, like that for minutes of standing committees composed of national representatives (a device popular with the Board of Trade, but not the FO),Footnote 227 and agreements subsidiary to and not materially altering a principal agreement,Footnote 228 reflected more artificial pathways for secrecy.
France seems to have been less preoccupied by the Article 102 requirement. Where secrecy was felt necessary, the government took advantage of the possibilities provided by domestic law. This is particularly evident in relations with former African colonies, including provisions, inter alia, for French intervention to maintain internal order. The felt need for secrecy in these instances may have emerged from a French desire to conceal details of military arrangements, or to avoid being pressed by other states to ratchet up support, but the main impetus for secrecy seems to have come from the new regimes' concern to preserve at least a façade of independence (or, conceivably, anxieties about possible obstacles to UN membership arising from a perceived lack of independence).Footnote 229
Post-independence treaties with Cameroon offer a particularly clear example of the drafting strategies used to maintain secrecy. The suite of texts included a Treaty of Cooperation, annexing various published accords and conventions. At the same time, the same plenipotentiaries signed a secret defense accord, providing, inter alia, that the Republic of Cameroon could request France to provide, on conditions defined by further special accords, the assistance of French armed forces for the maintenance of internal order,Footnote 230 together with various annexes (on aid and mutual facilities and strategic primary resources).Footnote 231 Plenipotentiaries signed one protocol of ratification for all the public accords and a separate secret protocol for the defense accord.Footnote 232 This was possible under both the French Constitution and the new Cameroonian constitutional framework, which referred to the making of “agreements [accords] and treaties” but provided only for “treaties” to be submitted to the National Assembly for approval prior to ratification.Footnote 233 Texts forming part of the same cluster of transactions were thus sorted, by form and terminology, into public and secret.Footnote 234
Most of the post-1945 secret treaties and ancillary texts described thus far reflect a gradual systematization of a secrecy that could be cultivated within and through law, or at least in the complex relations between the law of treaties, and the diplomatic crafts of drafting. Yet there were also instances of radically irregular “secret treaties,” like the “Protocol of Sèvres” (1956),Footnote 235 settled outside the usual machinery of foreign policymaking and envisaging conduct in breach of international law.Footnote 236 The Protocol was in part a response to Nasser's nationalization of the Suez Canal, and an effort to check Nasser himself, who was thought to be a conduit of Soviet influence in the Arab world. It provided for an Israeli attack on Nasser's Egypt, following which Britain and France, ostensibly onlookers to the operation, would issue appeals to both sides to withdraw troops from the Canal zone, and press Egypt to accept temporary occupation by Anglo-French forces.Footnote 237 The Protocol was, in substance, a secret offensive alliance, one of the very phenomena that proponents of publicity had aspired to eliminate.
On one view, the secrecy of the Protocol was tangential, as the nefariousness of its contents placed it outside the register of law altogether. There is an inherent contradiction in appealing to law to hold other parties to an undertaking to behave unlawfully. Translated into doctrinal terms, one might argue that there could not have been, in the circumstances, an intention to create legal relations at all.Footnote 238 That said, there are indications that at least some of the signatories considered the Protocol to have legal force. The text, drafted by French and Israeli officials, had characteristics reminiscent of a treaty. Drafters had intended for it to have a preamble, but ran out of time to craft one. The text was signed by representatives of the parties, and referred to the fact that its provisions would enter into force [“entreront en vigueur”] after the agreement of the three governments (the delay being occasioned by the fact that the British signatory was a deputy under-secretary of state, who could only sign ad referendum).Footnote 239 Israeli negotiators, who had requested preparation of the document, were keen for the political and symbolic prestige associated with signing a formal document of some kind (even in secret) as an equal with major powers;Footnote 240 and the French signatory, in his memoirs, did not exclude the notion that the Protocol was intended to have legal force.Footnote 241 Subsequent letters confirmed the French and Israeli governments' agreement to both “the result of the conversations of Sèvres and the terms of the final protocol to which they gave rise,”Footnote 242 giving the Protocol a status independent of the conversations. The British prime minister, by contrast, was dismayed that the Protocol had been brought into existence at all, and tried to suppress all material trace of it.Footnote 243
We might see this behavior as reflecting the ubiquity of legal forms. Even where parties considered themselves to be making, at best, a “non-binding agreement,” they drifted by habit and training into forms characteristic of a treaty. Alternatively, it might speak to the endurance among some diplomats and statesmen of a radically voluntarist model of international law, in which states' agreements inter se are thought to have some legal force despite being at odds with other undertakings, not least the UN Charter. To the extent this is the case, it highlights starkly the terrain left open once a failure to register (and thence have published) a treaty was deemed not fatal, in itself, to its legally binding status.
New Trends in the Management of Secrecy, c. 1960s Onward
The law of treaties underwent major scrutiny in the 1950s–60s, culminating in 1969 with the VCLT. However, the codification process had relatively limited impact on publication. Crucially, the VCLT preserved the ambiguities over the outer bounds of the category of treaty which had been so central to the accommodation of secrecy.Footnote 244 In stipulating the limited extent to which a failure to respect provisions of internal law, including requirements that treaties be laid before legislatures or published beforehand, might be invoked to invalidate consent, it left open the possibility that secret treaties might bind states even in circumstances where they had been made in violation of domestic law.Footnote 245
From the late 1960s, however, there were a number of changes to the management of secrecy in the three states surveyed here, driven primarily by democratic concerns and functional imperatives linked to the growing scale of treaty-making. This part sketches three such changes: (1) requirements for provision to the U.S. Congress of (and thus publicity for) executive agreements, some of which had escaped Congressional attention to date; coupled with public acknowledgement of the existence of a category of “classified agreements,” to be supplied only to House and Senate committees under injunction of secrecy; (2) efforts to grapple with the legal status of inter-agency agreements; and (3) formalization and elaboration of the genre of the “non-binding agreement.” Avenues (1) and (2) further systematized a margin for secret, yet legally binding, texts, continuing the dual history of entrenchment of a norm of publication alongside preservation of possibilities of secrecy. Avenue (3) pushed secret arrangements out of the realm of legal obligations altogether, thus fulfilling the spirit of Article 18, but not in a way which definitively resolved the boundary between “treaties” and “non-binding agreements.”
As the “vast sub-structure of intergovernmental paper” grew after WWII, the difficulty of distinguishing treaties from mere diplomatic correspondence became more acute.Footnote 246 Congressional hearings in the late 1960s over contested issues of foreign policy brought to light that major allocations of funds and materiel, as well as representations about how the United States would conduct itself in particular contingencies, rested not only on Article II “treaties,” but rather on executive agreements of various kinds, including oral statements. While State Department officials were well-versed in diplomatic wording that conveyed distinctions between executive agreements and lesser texts, Congressional committees did not necessarily grasp these nuances, and nor did foreign governments.Footnote 247
Congressional concern about unchecked executive power over foreign policy spurred the Case-Zablocki Act of 1972, which required “any international agreement, other than [an Article II] treaty” to be reported to Congress within sixty days of entry into force.Footnote 248 Any agreement “the immediate public disclosure of which would, in the opinion of the President, be prejudicial to … national security” was to be transmitted only to House and Senate Committees on Foreign Relations under injunction of secrecy.Footnote 249 The State Department interpreted the Act as further qualified by an inherent executive privilege which might permit the withholding of an otherwise reportable agreement.Footnote 250 However, the Act did acknowledge the existence of classified executive agreements—secret treaties, in the international law sense—and provide limited Congressional oversight. In the United Kingdom and France, there was no equivalent explicit acknowledgement and, to the extent that secret texts were available to legislatures, this occurred, in all likelihood, only on an ad hoc basis.
In the United States, efforts to implement the Case-Zablocki Act led to a reckoning with “inter-agency agreements.” Scholarly commentary had expressed doubts about the extent to which such agreements could bind the state, at least independently of some principal text made in the name of heads of state or government.Footnote 251 In the United States, agencies did not always supply the texts of agreements to the State Department in a timely way, or at all.Footnote 252 Recourse to such agreements thus undermined both Congressional oversight under the Case-Zablocki Act, and the ability of the State Department to exercise centralized control over U.S. commitments. In 1976, the State Department clarified that it regarded “agency level agreements” as capable of binding the United States under international law, and as falling within the Case-Zablocki Act as long as they met certain criteria.Footnote 253 The Act was modified in 1977 and 1978 to require all agencies to transmit agreements to the State Department within twenty days, with oral agreements to be reduced to writing and reported; and to consult with the State Department consultation, at least over certain categories of agreement, before signature.Footnote 254 The State Department position on what was captured by the Act, and the obligations of other agencies, were incorporated in regulations in 1981, albeit subject to a stipulation that failure to comply with the regulation did not affect the validity of an agreement in domestic or international law.Footnote 255
In France, too, the perceived need for the foreign ministry to reassert control over increasingly dispersed transnational cooperation seems to have driven the adoption of circulars in the 1980s requiring that all persons signing international engagements have instruments of full powers, and that, once signed, texts be sent to the MAE.Footnote 256 In the United Kingdom, although there had been discussion, for example, of the capacity of bodies like the Bank of England to enter into agreements with counterparts, the preference appears to have been to treat many “inter-agency” agreements as non-binding, and thus separate to some extent from the treaties for which the Foreign and Commonwealth Office was administratively responsible.Footnote 257
If new developments on legislative oversight in the United States, and bureaucratic disciplines on inter-agency agreements, tended to formalize possibilities for secret yet legally binding commitments, recourse to “non-binding agreements” offered an alternate means of securing secrecy (along with other benefits, like simplicity of negotiation and amendment). Such “non-binding agreements” were, in one sense, merely a continuation of nineteenth-century “gentlemen's agreements.” However, the latter tended to be expressed in brief and sometimes personal terms, and rest on the honor of individual diplomats, whereas post-WWII arrangements could be lengthy and detailed, and purport to engage agencies, governments and the state rather than individuals.Footnote 258 While non-binding agreements could be used for high-profile “political” commitments like the Helsinki Final Act,Footnote 259 in which case they might be widely published, and even reproduced in collections of legal materials, they were of particular utility where states wished to avoid the publication requirements associated with treaties.Footnote 260
By the 1970s–80s, British practice and scholarship had developed a fine-grained account of the textual and other indicators differentiating a non-binding agreement (which came gradually to be labeled “memorandum of understanding” or “MOU”) from a legally binding treaty. Recourse to non-binding agreements appears to have been embraced in Britain and the Commonwealth even for arrangements which other states preferred to make in treaties.Footnote 261 However, the distinction between a treaty and a non-binding agreement has not always been clear, even among signatories. The British contention, in a major arbitration, that an intergovernmental MOU was not legally bindingFootnote 262 seems to have surprised the U.S. government, and brought to the surface the fact that at least some of the MOUs which the United States regarded as treaties (in the international law sense) were not considered as such by the United Kingdom and Commonwealth states.Footnote 263 There may still be instances in which the U.S. and UK governments take different positions on the status of particular texts.Footnote 264 Whether this is an oversight or a tacit Atlantic “agreement to disagree” about legal status in certain problematic cases, it confirms again the porous and negotiable outer bounds of the “treaty” category.
VI. The Secret Treaty Today
Today, Article 102 is considered to have customary status, and the Secretariat interprets “[e]very treaty and … international agreement” as any “written agreement governed by international law entered into by parties with the requisite treaty-making capacity.”Footnote 265 However, practice suggests that states consider Article 102 to be something less than a comprehensive obligation to register all such texts.Footnote 266 A stipulation that a given text is not to be registered may be an indication that the text is not a treaty, but this is not determinative; and the non-submission of a text for registration cannot be presumed to preclude it being a treaty. The ICJ has rendered decisions based on interpretations of unregistered treaties.Footnote 267 Even as many aspects of international law and interstate interaction are subject to greater transparency, a central mode of generating legal rights and obligations in the international system remains free of any comprehensive publication requirement.
Dominant interpretations of Article 102, together with the limited attention to enforcement, mean that domestic law requirements and bureaucratic pathways probably remain the strongest determinants of publication. Domestic law requirements have strengthened since the early twentieth century, but in ways that preserve possibilities for secrecy—albeit refined and limited over time. Of the three states considered here, the United States is most forthright in asserting the existence of a category of secret treaties, in the form of “classified agreements.” Article II treaties, and most executive agreements, are published or at least made publicly available,Footnote 268 and thence submitted for registration. “Classified agreements,” or classified ancillary texts to published agreements, are not published or submitted for registration, but are provided to the foreign relations committees of the Senate and House (unless deemed subject to executive privilege).Footnote 269
France occupies a middle ground, insofar as there is a requirement to publish all treaties and agreements, but one that acknowledges some instances in which this will not be possible. Certain categories of treaty are required by the 1958 Constitution to be submitted to the legislature. An amended version of the 1953 decree still requires publication in the Journal Officiel only of treaties of a nature which would affect private rights and duties,Footnote 270 though a circular of May 30, 1997 provides that, other than in exceptional cases, all accords must be published in the Journal Officiel.Footnote 271
In the United Kingdom, a statutory crystallization of the Ponsonby rule, introduced in 2010, now requires that treaties subject to ratification or some equivalent procedure be laid before Parliament for twenty-one sitting days, and published in a manner that the relevant minister deems appropriate, prior to ratification.Footnote 272 The publication of treaties not requiring ratification or some analogous process is governed by a convention that all such treaties be published and laid once they have entered into force.Footnote 273 This issue received only glancing attention in the recent legislative reform process.Footnote 274 The government makes occasional references to confidential material in treaties or ancillary texts, but often instead frames unpublished commitments as non-binding MOUs. In light of ongoing uncertainty about how to distinguish between binding treaties and non-binding agreements, and what factors will be given most weight,Footnote 275 use of MOUs entails close attention to both textual indiciaFootnote 276 and negotiating strategies.Footnote 277
The three democracies examined here are outliers. They are unusually committed to publicity of treaties relative to states as a whole. Although numbers of treaties submitted to the UN Secretariat for registration do not indicate the proportion of treaties made by each submitter which are registered, these numbers do still give some rough sense of engagement with the Article 102 obligation. The United States is orders of magnitude ahead of other states in terms of the number of treaties submitted. The United Kingdom and France (the latter since the mid-1960s at least) are also, in relative terms, steady submitters. Other major powers have submitted relatively few treaties. The number of treaties submitted by the Russian Federation, for example, is actually lower than under the USSR, and the Peoples' Republic of China, admitted to membership in 1971, submitted no treaties until 1985, when it began a modest rate of submission.Footnote 278 Moreover, in many states, domestic law allows governments leeway to make treaties without the legislative approval or notification which is a major impetus for publication in the states surveyed here.Footnote 279
For obvious reasons, it is difficult to describe secret treaties currently in force. Some of the secret treaties examined here have been replaced over time. Many early base rights agreements have now been renegotiated, for example, as have many of France's relations with former colonies.Footnote 280 Where secret treaties, or secret ancillary texts, remain, they may take a more systematic and detailed form than equivalent treaties in the nineteenth and early twentieth century, commensurate with the increasing functional specialization of transnational relations. Taking the United States as an example, it is clear that some classified agreements currently in force are iterations of a known genre (such as status of forces agreements, agreements concerning base rights, or older “security of information” agreements).Footnote 281 Others, however, might be rather unique arrangements. There is said to be a government-to-government agreement between the United States and Pakistan, providing for approval of at least some drone targets by Pakistani military and intelligence officials.Footnote 282 There seem also to have been written arrangements in relation to at least some of the “black sites” controlled by the CIA and used for the detention, interrogation and torture of “High Value Detainees” after September 11, 2001.Footnote 283 The circumstances surrounding the drafting of these texts remain unknown at this stage. It seems difficult to conceive that the “black site” arrangements, in particular, could have been intended to be legally binding, yet they were apparently intended to produce some legal effect, however attenuated, by managing the relation between legally dubious activity and the public law framework in which this activity might be assessed.
This snapshot of the secret treaty today must also acknowledge the changing role of treaties in the larger landscape of interstate interaction. Recourse to MOUs, and the proliferation of international organizations, may be moving whole domains of activity out of the realm of treaty governance altogether. While international organizations' own treaty-making has been subject to the norm of treaty publication,Footnote 284 such organizations also offer an alternative means of coordinating action. Matters which might once have been agreed in a treaty may now take the form of an internal decision, resolution, or policy of an organization or network—and escape the norm of publication applicable to treaties altogether.Footnote 285
VII. The Survival of the Secret Treaty: Past and Future
The history traced here reveals contrasting trends. On one hand, the development of an international norm of treaty publication strengthened domestic publication processes. It brought into being a centralized (and now digitized, and freely accessible) repository of treaties, featuring authenticated texts and translations, thereby enabling tangible instantiations of treaty law as a genuinely international corpus. It created a dynamic in which one state's publication was a matter of legitimate concern to, and discussion in, other states, and made it impossible for democracies which wished to uphold the values of the League and United Nations to assert baldly that they were making secret treaties. That said, the relationship between publicity and legality sketched first in Article 18 of the Covenant has never been wholly realized, either in international law or the domestic law of leading democracies.
Even if democracies make relatively few secret treaties, the overall workings of treaty-making offer a contrast with the hopes of 1919. Reformists of the early twentieth century had envisaged registration and resulting publicity as marking off a realm of legally binding treaties from other possible commitments, grounded in honor or interest, but without legal force. This vision has been seriously compromised. The secret treaty remains a viable legal instrument in many states, hindered only by an uneven patchwork of domestic law. In the states surveyed here, foreign ministries have accepted that published treaties may have secret ancillary texts, the relation of which to the principal text is policed with varying degrees of strictness. Secret commitments may be channeled outside the register of law altogether, into MOUs. Such MOUs may constitute a law-like corpus of obligations, even while not formally legally binding. They exist in a sort of parallel system that may not be superior, on rule of law or practical grounds, to secret but legally binding treaties. Moreover, secret texts—whether framed as ancillary to the published treaty or as MOUs not having legal force themselves—may well shape the interpretation of published texts, creating stratifications of knowledge about putatively public transactions.
The gulf between the state of affairs today and the hopes of reformers in 1919 is partly the result of short-term interests of governments and institutions pressing against the norm of publication and the ideals it was intended to serve. But the history also suggests significant complications in the reformists' vision. If it might broadly be conceded that publicity of treaties assisted in democratic control of foreign policy, it became clear that publicity alone was insufficient. Meaningful review of treaties required a more sophisticated institutional machinery for legislative oversight (something reflected in changing structures and practices in the legislatures of many states since WWII). The notion that a norm of publication would foster peace, too, was challenged, with secrecy about interstate financial dealings seen as potentially fostering peace and stability. Some of the secrecy intended to mask humiliating inequalities of negotiating power between states might also be understood as upholding peaceful avenues for interstate interaction. Even with military arrangements, views about the genesis of conflict shifted, such that secrecy came to seem not a catalyst for suspicion and belligerence but a necessary element of the defense of the existing order. The complexity of any relation between publicity and peaceful relations, or at least the avoidance of aggression, is borne out by post-1945 game-theoretical and strategic approaches to the role of information in the shaping of state behavior.
As for the imagined relationship between publicity and legality, the basic intuition animating Article 18––that legal obligations are, or should be, intrinsically public—was countered by the persistence, alongside increasing institutionalization of international law, of a voluntarist understanding in which states themselves retained discretion about the forms of their legal obligations inter se, and the disclosure of these obligations.
The relationship between the norm of publication, and the general construction of a law-governed international order, proved similarly conflicted. On one hand, the forging of a centralized corpus of treaty law, and the improvement of national publication practice, seems likely to have disseminated and help systematize international law. Moreover, the fact that secret treaties were sometimes used to coordinate unlawful, or legally dubious, actions which could not have been agreed publicly does support the intuition that a norm of publication would tend to bolster the commitment to legality in interstate relations (and, conversely, that violation of the norm of publication would tend to undermine legality). Yet many secret treaties do not anticipate a violation of law. Secret treaties of this kind may reflect, in a perverse way, faith in the power of legal obligations to guide behavior, even in the absence of the mechanisms for institutional vindication of rights, and reputational pressure, which usually accompany (public) treaty obligations. Episodes like the 1926 dealings with Ethiopia do suggest that the norm of publication sometimes fostered a more inclusive, open interstate interaction, vindicating sovereign equality, instead of perpetuating hierarchical, self-selecting clubs. However, the history given here suggests that a norm of treaty publication may not displace power politics even from the limited domain of treaty-making, but rather transpose it onto new procedural terrain, favoring parties best able to manipulate the subtleties of drafting.
Law itself proved resistant to reordering in favor of publicity and double-edged as an instrument of transformation toward greater publicity. Legal advisers were often among those who upheld strong readings of publicity requirements, and tried to entrench them in bureaucratic practice. Yet the law was not a neutral instrument. Not only was the deep structure of the law of treaties resistant to a universal norm of publication, the law of treaties both accommodated, and obscured from public view, the preservation of secrecy in more quotidian ways. The uncertainties about the outer bounds of the category of “treaty,” coupled with the diplomatic craft of drafting layered and fragmented arrangements, offered avenues for secrecy. These avenues only multiplied as the scale and extent of treaty-making grew, testing the institutional capacities of foreign ministries and legislatures. The technical complexity of the law meant that the persistence of secrecy was often opaque to the very legislatures and publics thought to be empowered by publicity. More generally, the transmutation of political opposition to “secret treaties” into a specific legal discourse steered contestation over the limits of publicity into highly technical legal vocabularies, and sapped much of the political impetus and public engagement which had initially animated Article 18.
The foregoing may seem a dismaying conclusion. Not only has legal reform failed to eliminate secrecy, but law has in some ways helped sustain secrecy, and legal vocabularies have deflected public scrutiny and engagement. However, the historical account provided here does not suggest that the ideals initially underpinning Article 18 ought to be abandoned, or that they are unattainable. Rather, this history invites new research into the incidence of secrecy in treaty-making, its effects and implications—but with a more sophisticated understanding than was available in 1919 of the architecture of publicity and secrecy, and the economies of knowledge to which it gives rise.
This historical account suggests the importance of domestic law and bureaucratic processes of treaty-making in shaping responses to the norm of publication. Secrecy may be permitted by domestic law, or negotiated in the interstices of publication requirements, through some of the techniques illuminated here. This article has focused on leading liberal democracies. The global incidence of secret treaties, and current motivations for secrecy, especially in the relations of autocratic or illiberal states, remain to be explored. Secret treaties may be more prevalent in illiberal states than among democracies (although, paradoxically, difficulty in making credible commitments may require such states to use public and enforceable agreements on points of importance). There remain also interesting questions about patterns of secrecy in dealings between more and less open states. In many of the interactions surveyed here, the liberal democracy has been the more powerful or influential interlocutor, but where roles are reversed, or two states relatively evenly matched, will the publication practices of democratic states prevail, despite preferences for secrecy among less open interlocutors?
Analysis of the effects of secrecy must take account of the complexity of secrecy as a phenomenon. Texts are not always sorted into binary categories of “secret” and “published,” but arranged into larger transactional complexes. There is sometimes ambiguity, or disagreement between the parties, about the legal status of particular texts. Secrecy is not an “all or nothing” quality. It sometimes entails strict controls on flows of information, and sometimes a much looser arrangement, which does little to prevent information leaking out, but relieves governments of the need to detail, or admit, a state of affairs which is widely understood. Particular commitments might be entirely unknown to the public, known in their general lines but not their specificities, or well-known only to certain segmented audiences (for example, foreign intelligence services).
Once understood in this light, secrecy might have diverse consequences. In circumstances where even published treaties are proliferating at a rate that challenges the capacity of sophisticated bureaucracies,Footnote 286 secrecy may affect the likelihood of a treaty ever being fulfilled. In other words, some connection between registration and publication, on one hand, and binding force, on the other, may operate in practical terms. Conversely, provided a secret text is known to relevant personnel, it might shape official actions despite not being processed by foreign ministries or included in the corpus of published commitments. Secrecy may have implications for the way in which international commitments can be invoked in domestic law, and thus the state's ability to implement undertakings, or to justify its conduct. Secrecy may also play an enduring role in the interpretation and evolution of commitments over time. Where secret commitments are present, the interpretive process is necessarily opaque (even more so than it might be in the case where travaux préparatoires and other material pertinent to interpretation are not widely distributed). This may favor the position of the more powerful party, although dynamics in particular instances will vary enormously.Footnote 287 All of this suggests that the effects of secrecy in particular cases depend greatly on the institutions and regimes involved, the stratifications of knowledge among different sets of interlocutors, and the dynamics of the substance of the treaty.
Analysis of secrecy's instrumental effects does not exhaust the intellectual challenge it poses. Secrecy is not only a subject of study in its own right, but a revealing window onto some larger questions. To return to the theme of the relationship between publicity and legality, attention to secrecy calls into question what law is. In stripping legal obligation of its institutional and rhetorical accoutrements, it offers glimpses into—often contradictory and inchoate—understandings of the nature of legal obligation. It also sharpens normative questions about what law should be. Effects of secrecy on, for example, the coherence of the law, or of the state as a juridical actor, may be limited by strong centralized oversight of commitments that maintains coherence (or at least non-contradiction) between obligations despite the secrecy of some texts. But even in such cases, the possibility of secrecy would still call into question, for example, the very notion of international law as a knowable corpus of rights and obligations. In this respect, attention to the aberrant cases of secrecy has the potential to probe exactly what it is about lawful ordering that we find normatively compelling.
Target article
The Survival of the Secret Treaty: Publicity, Secrecy, and Legality in the International Order
Related commentaries (3)
Introduction to the Symposium on Megan Donaldson, “The Survival of the Secret Treaty: Publicity, Secrecy, and Legality in the International Order”
Secret Treaties in International Law and the Faith of States in Decentralized Enforcement
The Substance of Secret Agreements and the Role of Government Lawyers