MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA

PUBLIC INTERNATIONAL LAW — MODULE II: SOURCES

Condensed Exam Notes | Seven Prescribed Readings
0. THE FRAME
Article 38(1), ICJ Statute — the Court shall apply:
• (a) international conventions, general or particular, establishing rules expressly recognised by the contesting states;
• (b) international custom, as evidence of a general practice accepted as law;
• (c) the general principles of law recognised by civilized nations;
• (d) subject to Art. 59, judicial decisions and teachings of publicists, as subsidiary means.
• 38(2) — the Court may decide ex aequo et bono if the parties agree. Never once used.
Three things to notice immediately (they generate the whole Module):
• The word “source” never appears in Article 38. It is an instruction to judges on applicable law, not a legislative list.
• It is copied almost verbatim from the PCIJ Statute (1920). The ICJ added “whose function is to decide in accordance with
international law” — usually read as strengthening the claim that these are general sources.
• The list is ambiguous on hierarchy. Order ≠ rank; the only clear hierarchy is jus cogens.
Reading → syllabus map
Head Main reading Support
C.1 Art. 38 and its list Kennedy Panezi, White
C.2 Custom Akehurst Kelly, Kennedy
C.3 General principles Panezi White
C.4 Equity White Panezi
C.5 Judicial decisions Kennedy, Kelly Jennings
C.6 Jus cogens Kennedy, Panezi —
C.7 UNGA resolutions Jennings, Akehurst Kelly
C.8 Unilateral declarations Akehurst, Kennedy White
D. Treaties / VCLT Aust Jennings

1. KENNEDY — “THE SOURCES OF INTERNATIONAL LAW” (1987)

Thesis: Sources doctrine is not a map of where law comes from. It is a rhetorical technique that lets international law promise
sovereigns their freedom and promise the system its authority at the same time — by never letting either argument finish.

1.1 What sources doctrine actually does

• Every treatise defines each source abstractly, independently of the content of the norms. The boundary is meant to be
“on/off”.
• That only makes sense if the categories are thought to control the content of norms, not merely register it.
• The apparatus is entirely content-blind:
• Entry into treaty: capacity (VCLT Art. 6), executive competence (Art. 46 — violation relevant only if manifest, of
fundamental importance, concerning treaty-making competence), full powers (Arts. 2(1)(c), 7, 47), signature/ratification,
registration (UN Charter Art. 102), duress (Art. 52), reservations (Arts. 19–23).
• Entry into custom: how often, by whom, for how long.
• Exit: abrogation, denunciation, rebus sic stantibus (Art. 62), later custom.
• Hierarchy: jus cogens (Art. 53) at the top, then lex specialis, lex posterior, lex superior.
• Kennedy’s irony: commentators admit no norm has ever actually been found to be jus cogens (Sztucki: “there has been
nothing to codify”) — yet a hierarchy still “seems to need establishing”, or judges would have to choose between norms
on the basis of their content.

1.2 The core tool: HARD and SOFT

• HARD = the norm binds because the state consented.
• SOFT = the norm binds because of something outside consent — justice, good faith, systemic necessity.
• They are exhaustive and mutually exclusive by definition. A source cannot coherently be both.
The endless dialogue (memorise this):
1. State 1: you agreed (hard). State 2: I have changed my mind / never intended to consent.
2. State 1 must say change is impermissible — pacta sunt servanda, it would be chaos, you must keep your word.
3. State 2: none of that has anything to do with my consent. → All hard sources turn out to be soft in disguise.
INTERNATIONAL LAW - MODULE II | PAGE 1
MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
4. Reverse it. State 1 opens soft (the norm is just). State 2: I have a different idea of justice; among equals, justice must be
negotiated or voted. → All soft sources turn out to be hard in disguise.
5. Neither wins; neither trumps. Doctrine survives only by limiting both so they can coexist.

1.3 How doctrine manages the tension

• Allocation — treaty is coded hard, custom and general principles soft. But positivists corrupted their own position by
elevating the soft norm pacta sunt servanda, and by re-reading custom as tacit consent (hence opinio juris and the persistent
objector). Naturalists corrupted theirs by re-reading treaties as community judgments about justice (Hague Protocols,
Kellogg–Briand, UN Charter, human rights treaties).
• Characterisation and hierarchy — Article 38 is read as positivist because of the order of the list and because custom is
“accepted as law” and principles are “recognised”. Which permutation you adopt matters less than the project of making
distinctions, which “protects the image of a doctrinal resolution”.
• Proliferation — when a characterisation wobbles, doctrine breeds more distinctions: a hard rule of creation is paired with
a soft exception, and vice versa.

1.4 The instability shown: UNGA resolutions

• Article 38 omits them (the Assembly has no legislative function), so commentators try to force them into existing boxes.
• Consent argument (Tunkin, Falk, Higgins, Castañeda) → met by: delegates lacked capacity or intent (Sloan, Virally,
Arangio-Ruiz — states table drafts to “embellish their image” because everyone knows they are not binding).
• Justice argument (Bedjaoui, for a New International Economic Order) → met by: a resolution is a passing majority,
tyranny of numbers.
• Kennedy’s knockout: you cannot say resolutions are too instant to be custom and too implied to be consent. The objections
cancel — but the discourse uses them alternately, which keeps the question permanently open.

1.5 Four case studies

(a) Nuclear Tests (Australia v France, 1974) — unilateral declarations
• Facts: France announced it would stop atmospheric testing in the Pacific. The Court held France bound and never reached
the merits. Irony: Australia argued France was not bound (to get a merits ruling); France didn’t appear.
• The Court gave both justifications: (i) hard — intention to be bound gives a declaration the character of a legal
undertaking; (ii) soft — good faith, of which pacta sunt servanda is itself an instance.
• It stabilised them by (i) reading good faith as protecting other states’ reliance, and (ii) reading intent objectively, so France
could not redefine its own consent (“no arbitrary power of reconsideration”; no quid pro quo, acceptance or reply
required). On form it cited Temple of Preah Vihear (1961): no prescribed form if intention is clear. Then it moved the
question to its own role: the Court must form its own view of scope and is not bound by another state’s reading.
• Dissents: de Castro (policy statements, domestic jurisdiction; a promise may be revoked before acceptance); Barwick
(statements of policy, not solemn obligation). Rubin: the holding came from no treaty, no custom, no general principle, no
publicist — hence ultra vires. Franck: intent must be judged by listeners. Suy: look behind formal unilaterality for a
“bilatéralité de fond”.
(b) Rebus sic stantibus — US suspension of the Load Line Convention (1941)
• Roosevelt, on Biddle’s opinion, suspended the 1930 Convention because peacetime commerce — its “basic assumption”
— had vanished.
• Biddle’s two moves: unilateral suspension is mere procedure registering a lapse that already happened; and the lapse
rests on what was “fundamental” to the treaty, i.e. located in the parties’ own intention.
• Briggs (1942) attacked: if you rely on the treaty’s general nature, you don’t need rebus; if you rely on consent,
unilateralism collapses. Germany, Italy, Japan and the USSR could equally justify suspending inconvenient treaties.
Chesney Hill: no right of unilateral termination; recognition must come from the parties or a competent authority. Art. 25
allowed denunciation on 12 months’ notice; Art. 20 allowed modification only by unanimity.
• Kennedy: Briggs edges toward absolute hardness — if only enumerated circumstances count, rebus collapses into pacta
sunt servanda. Both men blend registers by letting each deny itself.
• Anchors: Free Zones (PCIJ 1932, decided on facts); VCLT Art. 62 (essential basis of consent + radical transformation of
obligations).
(c) Anglo-Norwegian Fisheries (1951) — custom
• The Court coded the UK’s ten-mile bay rule in consent terms (not general custom; and in any event “Norway has always
opposed” it) and defended Norway’s straight-baseline system in soft terms (long consistent application, coastal
geography, “vital needs of the population… very ancient and peaceful usage”, “moderate and reasonable”).

INTERNATIONAL LAW - MODULE II | PAGE 2

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• The decisive move: by coding one rule hard and the other soft, the Court made Norway’s rule “the” rule the UK had to
displace. The rest was a “mopping up operation” — then a hard argument against the UK (60 years of silence; formal
protest only in 1933; the UK could not have been ignorant of the 1869 Decree).
(d) North Sea Continental Shelf (1969)
• Denmark/Netherlands wanted equidistance; Germany (concave coast) wanted a “just and equitable share”.
• Stage 1 (both rejected on soft grounds): proportionality is at odds with the most fundamental rule — coastal-state rights
exist ipso facto and ab initio by natural prolongation and “do not depend on being exercised”. Equidistance is not
“logically necessary… an inescapable a priori accompaniment” of shelf doctrine.
• Stage 2 (hard analysis): Germany signed but never ratified the 1958 Convention; Art. 6 was proposed “with considerable
hesitation, somewhat on an experimental basis, at most de lege ferenda”; reservations were permitted; ratifications “hardly
sufficient”. Practice by non-parties would not suffice: acts must “amount to a settled practice” and evidence “a belief that
this practice is rendered obligatory”.
• Dissents: Lachs (practice need not be universal; over-exacting to demand proof each state felt obliged); Tanaka (opinio
juris is unascertainable psychology — infer it from the custom’s existence and felt necessity); Sørensen (uniform conduct
evidences opinio juris unless shown otherwise).
• Kennedy: the majority monopolised the hard register and left the soft to the dissents — so it could kill equidistance but
could not generate a replacement. Hence the anticlimax: delimit by agreement, on equitable principles, with natural
prolongation, non-encroachment, coastal configuration, geology, resources, and “a reasonable degree of proportionality”.

1.6 Conclusions
• Doctrine tilts toward consent; argument tilts toward systemic authority. Together they let the order express and transcend
sovereign power.
• Neither extreme is embraced. Pre-1648 scholarship needed no hierarchy (each category was bounded by its own norms).
Nineteenth-century positivism needed none either (the stronger will prevails).
• Modern writers want neither, because they are hunting the source of norms that bind the dissident.
• Sources discourse is therefore “at once fundamental and trivial”: logically prior to substance, yet the last resort of
argument (“this may not be persuasive, but it is authoritative”). It is “a discourse of evasion” that escapes theory only by
opening “an endlessly proliferating field of legal argumentation”.
Contemporary: The ICJ Climate Advisory Opinion (23 July 2025) is a perfect hard/soft clash. Big emitters argued lex specialis — only the
UNFCCC and Paris bind us (hard). The Court held a customary duty to prevent significant harm to the climate system and to cooperate
binds every state regardless of treaty participation, with breach generating cessation, guarantees of non-repetition and full reparation
(soft). UNGA res. A/80/L.65 (20 May 2026), tabled by Vanuatu, welcomed it 141–8–28 (against: US, Russia, Saudi Arabia, Israel, Iran,
Belarus, Liberia, Yemen) — i.e. converting a soft pronouncement into hard, vote-counting form, which the eight will say is not consent at
all.
Exam handle: Sources doctrine doesn’t tell us where law comes from; it lets the system promise autonomy and authority simultaneously by keeping
both arguments unfinished.

2. AKEHURST — “CUSTOM AS A SOURCE OF INTERNATIONAL LAW” (1974–75)

Opening image: St Augustine on time — “If no one asks of me, I know; if I wish to explain to him who asks, I know not.” Lawyers
use custom daily but cannot define it.
Five questions: (I) what is practice? (II) how much? (III) how consistent? (IV) what is opinio juris? (V) treaties and custom.

PART I — WHAT COUNTS AS STATE PRACTICE?

(a) Acts v. claims
• D’Amato: only physical acts. “A claim is not an act.” Judge Read (dissenting, Fisheries): “The only convincing evidence of
State practice is to be found in seizures.”
• Akehurst demolishes this:
• Law-of-the-sea conferences and diplomatic correspondence argue by citing claims and protests, never asking if enforced;
ten of fourteen judges in Fisheries Jurisdiction (1974) did the same.
• North Sea treated the Truman Proclamation as practice generating custom. At Geneva 1958 nobody argued claims needed
physical backing.
• Asylum: equal weight to “the exercise of diplomatic asylum and… the official views expressed”. Rights of US Nationals in
Morocco: custom found in correspondence and conference records. Pleadings are cited against states later.
• D’Amato is self-contradictory: he counts a treaty as practice — but a treaty is a statement, a promise, not a physical act.
And after Nuclear Tests a unilateral declaration can bind like a treaty; would France’s declaration then evidence a
customary ban on testing?

INTERNATIONAL LAW - MODULE II | PAGE 3

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• Physical acts are not more consistent: states act differently at different times; different departments act differently at once.
• The act/word line is artificial: “recognition is no more than a form of words” (Parry).
• Any statement creates political commitment (states hate being called inconsistent), and if embodied in standing
instructions or legislation, future conduct will likely follow.

(b) Statements in abstracto

• Thirlway: claims count only in a concrete dispute; abstract assertions are merely “confirmatory”, not “constitutive”.
• Akehurst: the line is “singularly thin”. At a conference, Arab states and Israel appear to argue abstractly about straits while
really meaning the Straits of Tiran; conversely a state may take a position in a dispute solely to establish a general
principle (Argentina’s protests over the Eichmann kidnapping). They merge; an observer often can’t tell.
• Practice supports him: replies to the League Committee of Experts were treated as constitutive practice in Mexican Union
Railway (1930), Eschauzier (1931), Mergé (1955), In re Piracy Jure Gentium [1934] AC 586 — sometimes as the sole evidence.
• North Sea rejected the Denmark/Netherlands argument built from ILC work, government replies and conference attitudes
not because such material cannot make custom, but because equidistance had been put as lex ferenda. Fisheries Jurisdiction
cited a 1958 conference resolution and a 1960 amendment as practice.
• Key distinction: lex lata v. lex ferenda. “Ought to be law” is not evidence it is law — it may even show it isn’t. States may
dress up demands as statements of existing law; others may contest. A claim phrased as lex lata and met with
acquiescence makes new custom even if the claimant knew it wasn’t the pre-existing law.

(c) Resolutions of international organisations

• They matter because state representatives vote for them; the organisational setting is irrelevant (a conference outside any
organisation would be the same).
• Authority: Nuremberg used League Assembly resolutions, the 1928 Pan-American Conference and unratified treaties to
find aggressive war criminal even before Kellogg–Briand (followed by Tokyo; Judge Pal dissenting). Anglo-Iranian Oil
Co. v SUPOR (Italy) cited a UNGA resolution as custom. Judge Ammoun (Barcelona Traction): delegates’ positions
“naturally form part of State practice” and “amount to precedents contributing to the formation of custom”; likewise
Judge Tanaka (South West Africa), Higgins, Asamoah, Castañeda.
• But only if declaratory in terms. Clear example: UNGA Res. 96(I) (1946) — “affirms that genocide is a crime under
international law”. Declaratory language is rare: outer-space resolutions are ambiguous; the Charter of Economic Rights
and Duties of States (Res. 3281(XXIX), 1974) speaks of establishing the “new international economic order” — the word
“new” negates any declaratory claim.
• Voting figures and explanations matter. In favour = accepting the declaratory claim (absent contrary statement); against
= rejecting it; abstention ≈ in favour, since a state not participating is bound unless it expressly dissents early. A separate
vote on the declaratory part is far more reliable.
• Non-declaratory resolutions may still influence future development where custom is uncertain or unsatisfactory — never
certainly.
• Against D’Amato’s “consensus” category: treating such statements as a separate source is “unhelpful and distorting”.
When states say something is custom, it is artificial to call that a statement about something else.
• The big pay-off: it is usually assumed the only way to change custom is to break it repeatedly. If abstract declarations
count, states can change it by repeatedly declaring the old rule dead — “a much more desirable way of changing the
law”.

(d) National laws and judgments

• Strupp (extreme consent theory) would count only treaty-making organs — excluding legislation. Trace: Judge Nyholm
(Lotus): “mere municipal laws are insufficient.”
• Rejected by practice: in Lotus Turkey relied heavily on national laws; France said only that they lacked others’ assent —
then cited national laws herself. 1877: the Foreign Office told the Minister at Rio that HMG “would not be justified” in
protesting a Brazilian extraterritorial criminal law because it resembled others’. Cutting incident: Mexico pointed to
similar foreign laws; the US said they were different or too few — then cited national laws itself.
• National courts infer custom from comparative legislation: diplomatic immunity (Lagos v Baggianini), ships’ lights (The
Scotia (1871)), enemy fishing vessels (The Paquete Habana (1900)). Nottebohm relied on national naturalisation laws
requiring a genuine link. In North Sea, Judges Ammoun, Tanaka and Lachs counted national laws and even Parliamentary
bills. The ILC always treats national laws, regulations and judgments as “primary evidence of State practice” (Waldock).
• Weight: frequently applied law weighs more, but mere enactment is practice even if never applied. Caution with
ambiguous statutes (how do courts read them?) and with statutes needing ministerial consent to prosecute aliens (how is
the discretion used?).

INTERNATIONAL LAW - MODULE II | PAGE 4

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• Strupp would also exclude non-foreign-ministry departments (absurd — the interior ministry handles aliens) and national
judgments, forcing writers to study belligerent rights at sea “while closing their eyes to the judgments of national prize
courts”. Even Tunkin includes them.

(e) Omissions
• Lotus: absence of prosecutions proved nothing only because there was no opinio juris — the clear inference is that
omissions plus opinio juris create custom. Nottebohm relied on “the practice of certain States which refrain from exercising
protection”.

(f) Organisations and individuals

• Organs composed of state representatives = state practice; organs not so composed (the UN Secretariat) can also create
custom — in Genocide (Reservations) both majority and dissents cited the Secretary-General’s depositary practice and a
League Council decision.
• International courts’ practice counts, including how they handle evidence and procedure.
• Individuals: no bar in principle, but few if any examples. International commercial custom (bankers’ credits) is not public
international law. States’ reactions to individuals’ acts create custom — that is not the same thing. Even writers have no
effect until states or courts accept them.

PART II — QUANTITY
(a) Repetition
• Most writers demand repetition; Waldock: density “depends on the nature of the case… where the occasions for acting
only arise spasmodically, the density required will obviously be less.” D’Amato, Cheng and Tunkin allow a single act.
• Judicial dicta demanding repetition are all obiter — the judges actually decided on conflicting practice (Genocide,
Asylum), missing opinio juris (Negulesco), no practice at all (Guggenheim), or treaty interpretation.
• North Sea properly read: the few delimitations cited were “inconclusive, and insufficient” because of relevance and
absent opinio juris, not number — and, decisively, Denmark/Netherlands were simultaneously making a new rule and
overthrowing an old one (the Court had already held custom required equitable delimitation).
• Lübeck v Mecklenburg-Schwerin (1928): a single act can — very unusually — create custom.
• Relativity thesis: much practice sits in unpublished archives, so custom is never proved absolutely, only relatively. The
party citing more precedents wins. One or two acts beat none — like citing one writer or judgment against none. Analogy:
title to territory, where courts accept “very little in the way of the actual exercise of sovereign rights, provided that the
other State could not make out a superior claim” (Eastern Greenland).
• Why international society differs: ~150 states, modern publicity and diplomacy. A single act between two states has an
impact two individuals’ act never could. Multilateral treaties, conferences and organisations let many states act at once.
“The number of States taking part in a practice is much more important than the number of separate acts.” One act
involving 50 states beats ten acts involving ten pairs. Hence North Sea’s own concession: “a very widespread and
representative participation in the convention might suffice of itself… even without the passage of any considerable
period of time.”

(b) Time
• Negulesco demanded immemorial usage — quoted often, apparently never approved.
• Right of Passage (1960): 125 years sufficed — without implying less would not.
• Continental shelf: Abu Dhabi (1951) found it not yet custom because practice was inconsistent, not because it was new.
Greece said ten years too short; Israel and nineteen other states disagreed. North Sea: “the passage of only a short period of
time is not necessarily, or of itself, a bar”. Lachs: freedom of movement into outer space established “within a remarkably
short period”. Brierly: sovereignty over airspace arose “at the moment the 1914 war broke out”.
• Baxter, endorsed: “the time factor as a separate element in the proof of custom now seems irrelevant. The new customary
rule will be established as soon as it acquires the necessary degree of acceptance.” Communications spread reactions
faster; organisations give more occasions to state views; and rules on things like spacecraft must develop fast (Wolfke).

(c) How many states?

• Old authority sometimes demanded all (Tinoco; Judge Weiss in Lotus). Others require only general practice — Art. 38(1)(b)
says general, not universal. Akehurst notes Art. 38’s drafting is “notoriously defective” (practice is evidence of custom,
not the reverse), and the wording has not stopped the Court applying regional and bilateral customs.
• North Sea’s* “extensive and virtually uniform” and “specially affected States” must be read in context: it concerned
treaty→custom transformation, was prefaced by the warning that this “is not lightly to be regarded as having been
attained”, and involved a rule conflicting with existing custom.

INTERNATIONAL LAW - MODULE II | PAGE 5

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• Judges demanding fixed minima never agree (unanimity; Tanaka’s “very large majority”; Ammoun’s “half is probably
not enough”, then “Third World consent needed”).
• Akehurst’s rule: the number needed varies inversely with the amount of conflicting practice. Few states suffice if
nothing conflicts. Every dictum demanding many states came in cases of conflicting practice. And with many new states
whose practice is unpublished, a high threshold makes new custom “intolerably difficult”.

(d) Presumptions
• Very strong presumption against changing the law. The better established a rule, the more practice needed to overturn
it. A slightly different new rule is easier than a radically different one. A rule with old and recent practice is firmest; only-
old practice looks outdated; only-recent practice may vanish as fast as it came.
• Also: presumption in favour of analogy, and against exceptions to broad principles. Close analogies beat distant ones;
specific principles beat broad ones — so the very general Lotus presumption of liberty is often overridden.
• Akehurst’s gloss on Lotus: it applies only to acts on a state’s own territory; there is a presumption against acts on
another’s territory. Turkey claimed concurrent, not exclusive, jurisdiction — a state claiming exclusivity limits others and
cannot invoke it. In a territorial dispute both sides deny the other’s right, so neither can use it (Norway said the UK must
not fish; the UK said Norway must not molest trawlers). And Turkey gained nothing from being respondent — the
presumption would be identical had she been applicant.

PART III — CONSISTENCY

• Right of Passage: constant and uniform practice made a rule — but inconsistency is not automatically fatal.
• Asylum: no rule, because inconsistencies were major.
• Working rule: minor inconsistency is survivable — practice must be “virtually uniform” (North Sea), not absolutely
uniform. But inconsistency raises the quantity required.
• Timing matters. The Paquete Habana: inconsistent to 1815, consistent after — so a rule could form after 1815. Conversely,
consistent past practice can only be destroyed by abundant, consistently contrary modern practice; mixed modern practice
cannot destroy it. Applied to Fisheries Jurisdiction: since a limiting rule certainly once existed, inconsistent modern practice
is fatal to claims beyond twelve miles, not supportive of them (Judge Petrén).
• No hierarchy among kinds of practice. Concrete-dispute positions look practical but may be self-serving; abstract
statements may be more considered. The arguments cancel. Foreign ministry v. courts: expertise and impartiality cancel
too — and anyway those are criteria for law-determining agencies, irrelevant to a law-creating process. Bias or ignorance of
state organs no more prevents custom than bias of MPs prevents statutes being a source of English law (and statutes
override judge-made common law).
• No hierarchy among states. Such suggestions come only from academics, never diplomats or courts. Some states
influence custom more because their practice is more frequent or better publicised, not intrinsically weightier. A power
with embassies everywhere shapes immunity rules more than a state with three. Great-power acts get publicity and are
imitated — but it is the imitation that matters, and suspicious small states may imitate other small states.
• The subtle example: when the US and UK claimed the continental shelf in the 1940s, others imitated quickly, probably
reasoning that if the traditional champions of freedom of the seas made such claims, either the claims weren’t really
contrary to it or that freedom was moribund — and they saw no point in being plus royalistes que le roi. Had the USSR
started it, imitation would have been slower.

Persistent objector
• A state can be bound by custom it never consented to (North Sea). Silence = bound.
• But a state that rejects the rule is different: it escapes if it opposes from the early days of the rule’s formation and
maintains opposition consistently. Late opposition is too late; abandoned early opposition loses effect.
• Authority: Anglo-Norwegian Fisheries — “In any event, the… rule would appear to be inapplicable as against Norway,
inasmuch as she has always opposed any attempt to apply it to the Norwegian coast.”
• Akehurst rebuts two readings: D’Amato (it was about special custom) — impossible, the preceding sentences deal with
general custom and special custom was never pleaded; Holloway (it rested on others’ acquiescence) — the Court said
nothing about acquiescence there, though it stressed it seven pages later on straight baselines; that difference proves the
point.
• Support: Judges Azevedo, van Wyk, de Castro, Lachs, Sørensen; The Antelope (1825); Fitzmaurice, Sørensen, Lauterpacht,
Verzijl, Thirlway; Soviet writers go further.
• North Sea’s* contrary dictum (custom “cannot be the subject of any right of unilateral exclusion”) is explained by context
— it answered the argument that signing the 1958 Convention completed the custom, and the reply was that the
Convention allowed reservations to Art. 6. That shows equidistance was not established custom in 1958; dissent ab initio
was not addressed.

INTERNATIONAL LAW - MODULE II | PAGE 6

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• Why the rule is necessary: if one dissent blocked a rule, new rules would never form; if dissenters were bound, custom
would be majority voting — and nobody could agree the size or weighting of the majority, and confident majorities
would stop compromising. “Dissenters need one safety valve; they do not need two.”
• Realpolitik: the UK dissented from the emerging 12-mile exclusive fishing zone after 1930. In theory it kept its rights; in
practice it could not enforce them (Iceland ignored the ICJ judgment in Fisheries Jurisdiction) while foreign trawlers,
excluded elsewhere, crowded to within three miles of the British coast. Britain bore all the costs of dissent and none of the
benefits — and eventually claimed a 12-mile zone itself.

New states
• Traditional theory: bound automatically to all custom existing at independence (Ware v Hylton (1796)). Consent theorists
(Tunkin, Bokor-Szegő) deny it but infer consent from entering relations unless a reservation is made — and such
reservations are never made, so the practical result is identical. Asian and African states’ attitude is “somewhat lacking in
clarity and consistency”: they deny rules that harm them, accept others without question.
• This follows from the objector rule — independence came too late to dissent.
• Legally 100 new states = 1; politically not. Akehurst’s constructive answer: new states can change law from within. New
customs among new states don’t bind old states, but with wide support the remaining opposition disappears (as with
fishing zones). Even before that, their opposition casts doubt on old customs, pushing both sides toward treaties — where
new states will secure some of their wishes — and treaties departing from old custom undermine it further.

Special (regional) custom

• Latin America (Asylum); Continental Europe; or groups linked by history, race, politics or religion — capitalist states,
communist states, the Commonwealth. In Barcelona Traction the Court found no general custom giving Belgium standing;
Akehurst agrees with Judge Gros that it should have asked about a special custom among capitalist states.
• States need share nothing but the custom: Right of Passage applied a bilateral custom between India and Portugal. Hence
“special custom” is the better term.
• Operation: special custom conflicts with general custom by definition. Between the bound states, lex specialis derogat
generali (unless the general rule is jus cogens). Between a bound and an unbound state, the general custom applies — and
if the first state also consistently dissented from the general custom, their relations are governed by the former general
custom that existed before the new one arose.
• Who is bound: general custom binds all except consistent objectors. A special custom among unrelated states binds only
participants. But Thirlway’s view (accepted): a custom among some members of a well-defined group binds all members
of that group except objectors. Borderline cases exist (Spanish-speaking South America v. Central America? Brazil?
Guyana?) but edges don’t defeat rules.
• Why formation rules must be uniform: one often cannot tell whether a custom is special or general. In 150 states the line
is far less clear than among millions of individuals. Where states split roughly evenly, who says which is general? Grotius
already saw this: the law of nations gets its force “from the will of all nations, or of many nations… in one part of the
world there is a law of nations which is not such elsewhere”. Vattel likewise. Alive again today over expropriation and
territorial sea / fishing zones.
• Sub-systems solve otherwise insoluble problems. Intra-group disputes use the group rule. For inter-group disputes
where each group has always dissented from the other’s, neither is bound — and since courts never admit gaps, the only
answer is to go back to a rule both once accepted. Artificial and uncertain, but better than the alternatives. Example:
universality and passive-personality jurisdiction, rejected by English-speaking states but resting on centuries of
continental practice, probably as old as the English territorial principle — “Who can confidently say what (if any) rule
existed before this divergence arose?”

PART IV — OPINIO JURIS

(a) The traditional test and its paradox
• North Sea: acts must “be evidence of a belief that this practice is rendered obligatory by the existence of a rule of law
requiring it”; states “must feel that they are conforming to what amounts to a legal obligation”. Frequency is not enough
— many acts of ceremonial and protocol are performed almost invariably, “motivated only by considerations of courtesy,
convenience or tradition”.
• Paradox: it requires states to believe something is already law before it becomes law. That belief is by definition mistaken;
mistakes are rare; and it leaves no room for deliberately making new rules.

INTERNATIONAL LAW - MODULE II | PAGE 7

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
(b) Attempts to abolish it, and why they fail
• Denial (early Kelsen, Guggenheim): many cases find custom without mentioning it — but no case says it is unnecessary,
and its absence was fatal in Lotus and North Sea. Saying judges have discretion is useless for prediction and for disputes
that never reach court.
• Why it is indispensable: international law has permissive rules too. If states habitually write to each other on white
paper, is that required or merely permitted? Frequency cannot tell you; only opinio juris can. It also separates law from
morality, courtesy and comity. Quadri offers de minimis non curat lex — but that explains white paper, not why
enforcement of foreign judgments (important enough for treaties) isn’t customary, and says nothing about permissive
rules.
• Opinio non juris (Lauterpacht; Sørensen dissenting): treat consistent practice as a duty unless disclaimed. But disclaimers
are usually absent — no state has ever said it is not obliged to use white paper — and a theory forcing that conclusion
“is obviously unsound”.
• Opinio juris as moral/social necessity (Le Fur, Scelle, Kelsen 1945, Thirlway, McDougal). Refuted: the overlap is far from
complete. The immorality of aggressive war was recognised centuries before its illegality. Most developed states admit a
moral duty to give aid — how many admit a legal one? And custom is full of illogical distinctions: a state is liable for its
political sub-divisions’ acts except their contractual debts; a neutral must stop nationals supplying warships but not
tanks. Judicial support: South West Africa (1966) — the Court “can take account of moral principles only in so far as these
are given a sufficient expression in legal form”; Haya de la Torre separated legal obligation from courtesy and good
neighbourliness.
• D’Amato’s “articulation” theory (a rule must be articulated in advance by a state, court, organisation or writer, with
publicity giving notice). Akehurst’s killer example: when Japan adopted the Western calendar in 1872 she said she was
obeying a duty imposed by international law. No state has ever articulated the contrary view that the calendar is
permitted but not required. On D’Amato’s logic, every state using it since has been confirming an obligatory rule —
“manifestly absurd”. One need not show every instance was accompanied by acknowledgment; but the claim that
conduct is required becomes doubtful when instances infinitely outnumber statements of obligation. (Same reason he
rejects Wolfke’s view that opinio juris relaxes as practice grows.)

(c) Akehurst’s reformulation: statements, not beliefs

• Practice makes a duty-rule if accompanied by statements by states that the conduct is obligatory. A minimal departure:
the traditional view looks for evidence of belief; Akehurst looks for statements of belief.
• Genuineness is irrelevant. The Truman Proclamation (1945) claimed international law gave coastal states exclusive shelf
rights. Whether the US believed it is beside the point — it said so, and others concurred.
• This is how custom changes: states assert something is already the rule (honest mistake or knowing falsehood —
irrelevant); if others acquiesce, a new rule is born.
• But the assertion must be that something is law — not that it ought to be, or is required by morality, courtesy or social
need. Moral statements make international morality, not law.

(d) Duties v. liberties — how opinio juris is shown

• Permissive rule: a claim of entitlement can be inferred from the fact that states act that way (Judge Petrén, Nuclear Tests)
— unless negatived (e.g. justifying an act as a reprisal implies it would otherwise be illegal). It must meet acquiescence
from affected states. (If the other state denies the right but permits the act ex gratia, the precedent is weak.)
• Duty-imposing rule: action plus acquiescence is not enough. There must be express statements — a declaration in
abstracto, an acknowledgment of one’s own duty, or an assertion that another state has a duty (usually a protest). Watch
the reaction: denial or inaction weakens the precedent.

(e) Protests and acquiescence

• Motives are irrelevant. A may not protest to protect a trade deal; A may vote for B’s resolution to curry favour. What
counts is what a state says or fails to say in public, not what it secretly believes. (Possible exception: intimidation by
force or bribery of officials — extreme and rare.)
• Knowledge matters: acts others could not know of carry little weight, and silence proves nothing. In Fisheries the Court
stressed the UK must have known of the baseline system.
• Protests are both opinio juris and part of the quantitative element. Acts and protests often cancel out — Judge Read
(Fisheries): seizures met with immediate British protest “must therefore be disregarded”.
• Protests create a rule only if the acts are withdrawn/not repeated, or protests greatly outnumber acts, or protesting states
greatly outnumber acting states.
• Weighing (O’Connell): number, vehemence, subsequent actions, importance of the interests affected, effluxion of time.
Isolated protests cannot stop repeated claims; isolated acts against a well-established rule cannot make a contrary rule, so

INTERNATIONAL LAW - MODULE II | PAGE 8

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
strictly no protest is needed to neutralise them. Protest by the one affected state weighs heavily; protest by a few where
many are affected weighs little.
• Gradation: failure to protest an abstract assertion < failure to protest concrete action; failure to protest an enactment <
failure to protest its application — but enactment still counts, since states do protest against enactments, and Lotus
stressed that objective-territorial jurisdiction laws provoked no protests.
• Against D’Amato (protests are not part of the quantitative element, so only physical acts nullify physical acts):
1. Disastrous for world order — a state opposing satellite overflight could no longer protest but would have to “interfere
with the satellite’s flight” or retaliate, converting a polite disagreement into a major dispute; and since D’Amato implies
acts count only if they succeed, effectively one would have to shoot down satellites. This also unfairly favours great
powers: a small state will protest to Moscow but never act physically against it.
2. Internally inconsistent — he lets protests by nearly all states count as “consensus”, and lets protests defeat special but not
general custom, though the two are often indistinguishable.
3. No judicial support.
• Authorities: Lotus — France and Germany “would [not] have omitted to protest” in the Ortigia/Oncle-Joseph and
Ekbatana/West-Hinder cases had they thought jurisdiction unlawful (Judge Loder’s demand for express consent was
rejected). The Elida (German Prize Court, 1915) said failure to object is not positive concurrence — but was inconsistent,
relying on a German declaration made “without meeting with opposition” in 1882. Fisheries — absence of protest against
straight baselines, later treated as a precedent of general applicability. Judge Lachs (North Sea): the first spacecraft
crossed states’ airspace and orbited above them; “the launching States sought no permission, nor did the other States
protest. This is how the freedom of movement into outer space… came to be established… within a remarkably short
period of time… To postulate that all States, even those which initiate a given practice, believe themselves to be acting
under a legal obligation is to resort to a fiction.”

PART V — TREATIES AND CUSTOM

• D’Amato: yes, treaties make custom. Contrary authority: Lotus (doubted that a stipulation “found in several treaties”
expresses a general principle); the Åland Islands Jurists’ Report (self-determination in “a certain number of treaties” is
not yet a positive rule); The State (Duggan) v Tapley (Ireland); West Rand Central Gold Mining v The King [1905] 2 KB 391,
Lord Alverstone: citing treaty stipulations is “as little convincing as the attempt… to establish a trade custom… by
adducing evidence of particular contracts”; they have little value as “international common law”.
• Akehurst’s position: treaties are practice, but like all practice they need opinio juris. North Sea is decisive: bilateral
delimitation agreements proved nothing because “there is no evidence that they so acted because they felt legally
compelled”.
How opinio juris can be satisfied by a treaty:
1. Treaties providing for action affecting third parties. Since treaties can’t take third parties’ rights, the parties are
implicitly claiming custom permits it; absent protest, a permissive rule may be inferred. Examples: treaties for
prosecuting third-state nationals for crimes abroad; NATO Art. 6 (collective self-defence on attack on “vessels or
aircraft”), suggesting the customary right covers ships and aircraft (note UN Charter Art. 51 calls it an “inherent right”);
the Armed Neutralities of 1780 and 1800 (collective reprisals); The Wimbledon (1923) — nobody ever said Turkish or US
neutrality was violated by their Suez and Panama obligations, so Germany’s neutral duties were not violated by
contraband passage through the Kiel Canal under Art. 380 of Versailles.
2. Statements in the treaty or the travaux. Declaratory preambles are rare, because codification always crosses “the often
imperceptible boundary between codification and progressive development”. A treaty may recite a customary rule and
then add enforcement machinery (Nuremberg’s unratified treaties; the Genocide Convention) or exceptions (Anglo-
American Liquor Treaty 1924) — in which case only the recital is authority. Waiver of rights shows they had some
foundation (waiver of claims is weaker).
3. Treaties routinely cited as declaratory: parts of Hague Convention XIII (1907); the 1958 Geneva Conventions; the
VCDR 1961; the VCCR 1963; and the VCLT. Courts usually assert this without authority because it is common
knowledge — but Akehurst warns “most” ≠ “all” (he objects to the citation of the VCLT’s procedural provisions in
Fisheries Jurisdiction).
4. Knowingly false declaratory statements still work if unchallenged (subject to conflicting practice).
5. Statements that a treaty is not declaratory cut the other way: in Asylum Colombia said the Montevideo Convention 1933
codified Latin American custom; the Court said the argument was “invalidated by the preamble, which states that this
Convention modifies the Havana Convention”. In Re Martinez the Italian Court of Cassation relied on ILC and conference
statements.
Withdrawal clauses, reservations, low ratification:

INTERNATIONAL LAW - MODULE II | PAGE 9

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• Withdrawal/revision clauses probably don’t show non-declaratory character — they may exist to let states reconsider if
custom changes.
• Reservations: North Sea suggests a reservation clause defeats declaratory character. Akehurst doubts it — dissenting
states not bound by the rule may naturally be allowed to reserve; and others may prefer a reserving party to a non-party.
Many actual reservations make declaratory status hard, but mere permission to reserve casts little doubt if few are made.
(Criticised by Baxter, Brown, Goldie, Lang, Thirlway; Judges Koretsky, Morelli, Lachs, Sørensen.)
• Few ratifications need not defeat declaratory status: the usual causes are inertia and lack of parliamentary time —
reasons for delay, not refusal. Value wanes over many years, but some survives, since one reason not to ratify is that the
rules already bind as custom.
Subsequent recognition:
• Any invocation of a treaty by or against a non-party = a statement that custom coincides with it, unless explained by
third-party rules.
• VCLT Art. 38: nothing prevents a treaty rule becoming binding on a third state “as a customary rule of international law,
recognized as such”. The words “recognized as such” were added at Vienna precisely to stop automatic transmutation.
• Reading North Sea correctly: the “widespread and representative participation might suffice of itself” passage was about
the quantitative elements (repetition, time). The very next paragraph adds that practice must “show a general recognition
that a rule of law or legal obligation is involved” — so opinio juris is an additional requirement, not an inference from
practice.
Which treaty rules ripen into custom (predictions, not rules):
• Inverse proportion to novelty. Rules adding precision to existing custom ripen easily — e.g. the Thalweg principle
refining Grotius’s and Vattel’s equal-division principle (New Jersey v Delaware (1933)). Nuremberg found the Hague
Regulations 1907 differed from prior custom but by 1939 were “declaratory of the laws and customs of war” — probably
because they added precision to the principle that a belligerent must not cause suffering unjustified by military necessity.
• Where custom is uncertain, diplomats and municipal judges lacking source-handling experience will simply apply the
accessible text — a multilateral treaty, a UNGA resolution, or an Institut de droit international resolution — and
subsequent application with opinio juris makes it custom. But it may never happen: where the uncertainty reflects a clash
of interests, each bloc keeps its own view. Example: capitalist/communist expropriation disputes settled by compromise
treaties never generated custom, because each side’s non-treaty practice was unchanged — and the treaties actually
increased uncertainty by removing recent precedents.
• Where states are dissatisfied, any well-publicised statement is a catalyst — but a lex ferenda statement cannot itself make
the rule. It is the practice inspired by it, with opinio juris, that creates law. The process may be fast, but it has more than
one stage.
• His 1975 law-of-the-sea prediction: adoption of a new convention would be followed immediately by a rush of claims,
even though most provisions would not be declaratory and entry into force was years away; the mere prospect had
already stimulated claims, and the Informal Single Negotiating Text (1975) made the content guessable. A clear failure
to agree would equally stimulate expanded claims, as the failures of 1930, 1958 and 1960 did. But it is the practice, not the
stimulus, that makes law — and the ICJ in Fisheries Jurisdiction refused to anticipate the outcome.

Akehurst’s own five conclusions

1. Practice = any act or statement by a state from which views about custom can be inferred — physical acts, claims,
abstract declarations (including UNGA resolutions), national laws, national judgments, omissions. Organisations too;
individuals in theory.
2. Participation matters more than frequency or duration. Few states, few occasions, short time can suffice if nothing
conflicts and other things are equal — but presumptions (e.g. Lotus) mean they seldom are.
3.
(a) Major inconsistency is fatal and cannot be explained away by ranking types of practice or ranking states. (b) A
state escapes if it consistently opposed from inception; a new state is bound by rules well established before
independence. (c) Special customs coexist with general ones and are hard to tell apart when numbers are even.
4. Opinio juris is necessary, and consists of statements that conduct is permitted, required or forbidden. Permission may be
inferred from conduct; requirement or prohibition must be stated expressly. The state need not believe its statements —
what matters is that other states do not challenge them.
5. Treaties make custom only with opinio juris. Otherwise, where a treaty is imitated later, it is the subsequent practice, not
the treaty, that creates the custom.
Contemporary: (i) The ILC Draft Conclusions on Identification of CIL (2018) essentially adopt Akehurst: two elements, verbal acts and
legislation and inaction as practice, “general” not universal, the persistent objector preserved, and resolutions as evidence of but not a
substitute for custom. (ii) His “change by declaring, not by breaking” is the strategy of the climate-vulnerable states: ITLOS AO (21 May

INTERNATIONAL LAW - MODULE II | PAGE 10

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
2024) (GHG emissions = marine pollution under UNCLOS) and the ICJ Climate AO (2025) both reasoned from accumulated declaratory
material, not from seizures — impossible on Judge Read’s view. (iii) The eight states voting against UNGA res. A/80/L.65 (2026) are
effectively trying to build an objector record against an emerging customary climate duty — but on Akehurst’s timing rule, objection after
a rule is established comes too late. (iv) His UK-fishing parable (all the costs of dissent, none of the benefits) explains why dissenters
usually rejoin — and also concedes that custom bends to enforcement capacity, which is Kelly’s complaint.
Exam handles: Practice = anything a state says or does revealing a view about law. Participation > repetition > time. Minor inconsistency survivable,
major fatal. Opinio juris = public statements, not private beliefs. Treaties make custom only through opinio juris.

3. KELLY — “THE TWILIGHT OF CUSTOMARY INTERNATIONAL LAW” (2000)

Thesis: CIL should be abolished as a source and replaced by consensual processes — not to weaken international law but “to
encourage the use of more democratic, deliberative processes”. Three limbs: (1) CIL lacks authority; (2) it lacks coherence; (3) it
lacks procedural legitimacy.

3.1 The opening exhibits — “CIL is a matter of taste”

Scholars and states assert opposites with equal conviction: - Torture — Restatement (Third) §702 and Filartiga v Peña-Irala (1980) say
it violates customary human rights law; Weisburd and Watson deny it (Watson: torture is widespread state practice). -
Environmental harm — Restatement §601 and O’Connell assert a prevention duty; Zemanek finds no widespread consistent
practice. - Expropriation — Restatement §712 and Amoco v Iran assert full compensation; Guha Roy and Jiménez de Aréchaga deny
it.
Orthodoxy attacked: CIL is the primary source of universal law, applying to all states regardless of culture since the 19th century;
practice + opinio juris; and treaties bind only because of the prior customary norm pacta sunt servanda.

3.2 What real customary law is (the anthropological baseline)

This is Kelly’s analytical foundation — learn it. - Customary law is empirical: the norms a society in fact believes are legally
required, found by the inductive method. It is not state practice — practice is only material evidence. “It is the community-wide
belief that a norm is legally required that provides customary law with authority and legitimacy.” - It survives in (1) traditional
societies (primary law-making); (2) subcultures — Native American, Roma — preserving themselves against assimilation; (3)
modern societies as a subordinate source (least useful comparison, since there custom binds only once the state recognises it). -
Norms, not habits. Gluckman: among the Barotse, law is the set of rules accepted by all normal members as defining the right and
reasonable ways to behave. Without the belief that conduct ought to be followed, custom is mere habit. - Internalisation gives
authority, often reinforced by religion — among the Dinka of southern Sudan norms are the legacy of mythic ancestors whose
observance lets them share village life. - Sanctions: public shaming (the Nunamiut Eskimo who steals a kayak is renamed
“Kayak”, reportedly more feared than a beating); denial of aid in scarcity; among Roma, fines, corporal punishment and marime
(banishment) — “the equivalent of social death”. - Empirical, not deductive. When Gluckman asked Lozi judges whether they
needed evidence of custom, “they laughed.” In modern tribal courts custom must be proved as fact. - Not consensual. The test is
what members regard as the rule, not the view of a lawyer or judicial class. No room for individual interpretations or exceptions.
- Law from below — not imposed by a legislature or class of specialists. - Definition: “the internalized normative convictions or
beliefs of a social group derived from the material of practice.” A set of norms, not a complete legal system.

3.3 Authority, legitimacy, and the vocabulary

• Authority = the quality a rule has when produced by the body empowered to make binding obligations; in custom, that
body is the community itself, which does not delegate to a legislature or court.
• Legitimacy — Franck’s four indicators: determinacy, symbolic validation, coherence, adherence to a hierarchy of
secondary rules (“right process”). Kelly’s aim: show CIL lacks all four.
• Method note: his authority critique is positivist (custom is empirical, so it cannot be a normative exercise); his
procedural-legitimacy critique is normative. He rejects Tesón’s normative theory as substituting an individual’s moral
position for the community’s beliefs.
• CIL as “nominalization”: it survives only because users mean different things — consent for some, consensus for others;
slow unconscious process for some, instant change for others.
• “Non-empirical CIL” = paper norms unrelated to practice believed obligatory — e.g. the US full-compensation standard,
or its converse, developing states’ permanent sovereignty over natural resources.
• “Empirical CIL” = the few genuinely accepted norms — structural norms like diplomatic immunity, which are permissive
and limit states very little.
• Traditional CIL v. “new CIL” — the latter extending custom to non-binding resolutions and to multilateral treaties the
state hasn’t accepted.
• Revisionists (Trimble; Bradley and Goldsmith) argued CIL should not be federal common law without political-branch
incorporation. Kelly goes deeper: CIL is not a legitimate form of law-making at all.

INTERNATIONAL LAW - MODULE II | PAGE 11

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
3.4 LIMB ONE — no authority
(a) Perception. International society is not organic but “an amalgam of nations with widely different cultural histories, values, and
interests” (180+ states), with no central authority to interpret events. Three problems: 1. Norm-formation conclusions reflect self-
interest or the scholar’s values. “The substantive norms offered as CIL in much of the Western literature are, not coincidentally,
norms associated with individualism and the market economy.” Even Western states don’t accept them as limits on sovereignty:
the US refuses human-rights obligations beyond existing domestic law, and would “strenuously resist” treating capital
punishment as cruel and unusual. 2. Interpretation depends on perception. The 1934 Hull letter to Mexico is cited for “prompt,
adequate and effective” — but Mexico equally forcefully asserted the Latin American national-treatment position. The weight of
BITs or of lump-sum settlements “varies depending upon a writer’s predilection or a nation’s history and perceived self-interest”.
3. What counts as practice is culturally filtered. Sovereign immunity example: many developed states follow the restrictive
theory (US FSIA 1976; European Convention on State Immunity 1972), and the Restatement and Koh call it CIL — but states with a
different conception of the state’s economic role reject it, and even Brownlie notes many restrictive-theory states won’t enforce
against sovereign assets. Convergence is real but driven by the spread of the market economy and World Bank/IMF policy —
“there is little indication that these domestic policy choices are, in any sense, obligatory.” - Two tendencies exposed: scholars find
rules matching their own values; and “many international lawyers feel a mission to increase and promote rules of international law
using CIL as the means in their messianic quest without observing the element of opinio juris.”
(b) Opinio juris cannot be measured.
• Courts rarely inquire directly; on the rare occasions the ICJ demanded direct proof it found the evidence inadequate. If
direct proof were required “there would be few principles of CIL”.
• The fallback — infer opinio juris from general and consistent practice — fails because practice is rarely general or
consistent, and conclusions rest on a few incidents. In a traditional society regularities evidence norms because each
member carries the group’s value system; in a diverse international society they don’t.
• Nicaragua (1986) restated the two-element test (quoting North Sea) — but did not itself analyse practice.
• Trail Smelter: the Tribunal found no relevant state practice and applied US Supreme Court decisions on disputes
between US states, assuming international principles were identical. That is the common-law technique of analogy, not
custom. It “has no precedential value” yet “spawned a vast literature on the customary international law of the
environment” — paper norms inconsistent with practice.
• Practice is unavailable: only the largest states publish digests, and the US stopped publishing its Digest as a budgetary
measure. Japan, China, Africa, Asia and Latin America are “virtually ignored”; even continental sources are used only in
English. Partly language, but also “a peculiar chauvinism that is incompatible with the concept of customary law”.
• Acquiescence and presumed consent are fictions — “the conclusion is presumed rather than proven”. Wolfke presumes
acceptance even where interests are not at stake. Consent without knowledge is impossible: states are unaware custom
is forming, may not know silence binds them, and lack resources to monitor and respond — even the wealthiest do not.
So CIL can form by inaction without the participation or even awareness of executive officials.
• Conclusion: “there is no methodology that has the capacity to determine whether states have, in fact accepted a norm as
law.”
(c) Deduction dressed as induction.
• CIL “remains the domain of experts familiar with and able to selectively manipulate a vast body of legal materials” — not
“the residue of customary norms understood and accepted by members of a society”.
• Louis Sohn, approvingly: this is how international law is made, “not by states, but by ‘silly’ professors writing books, and
by knowing where there is a good book on the subject.”
• The exemplary case — Libya v Malta (1985): the Court found a customary 200-mile shelf entitlement not from practice or
opinio juris but by analogy to the EEZ regime in UNCLOS 1982 — since EEZ seabed rights are defined by reference to the
shelf and there can be no EEZ without a shelf, “the distance criterion must now apply to the continental shelf”.
• Authority objection: the ICJ “possesses only the authority to apply customary law, not to create it”. Customary norms are
social facts, not expert creations.
• Non liquet: faced with no practice, the Court should refuse to decide. Non liquet is inherent in custom, which is a set of
rules, not a complete system; those who deplore it confuse the ICJ with a domestic court of general jurisdiction.
• D’Amato’s “claims approach” defines away the community’s conviction — removing formation from the only body that
can confer authority.
(d) Structural v. substantive norms — crucial.
• Structural/constitutive norms (personality, recognition, territorial sovereignty and jurisdiction) are mostly deducible
from sovereignty itself, confirm power states already exercise, and limit them little. They may be “associative norms” —
minimum conditions of membership (Franck).

INTERNATIONAL LAW - MODULE II | PAGE 12

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• Diplomatic immunity is his model of real empirical CIL: for centuries states granted it against their immediate interests
and acknowledged doing so, before codification in the VCDR 1961. Grotius justified it by extraterritoriality; the ILC used
“functional necessity”; the preamble says the purpose “is not to benefit individuals but to ensure the efficient
performance of the functions of diplomatic missions”.
• Jurisdiction shows what happens when structural rules are used to limit power. The principles — subjective territorial,
objective territorial (effects), nationality, protective — are permissive and near-limitless. The effects test is “a
potentially all-embracing justification”: the US DOJ asserted jurisdiction over foreign corporate activity, legal where it
occurred, on the basis that it might affect the revenue of foreign subsidiaries.
• The Restatement’s reasonableness limit (§403) fails twice: the balancing is “highly subjective” and each state’s courts will
back their own political branches; and it has no basis in practice — “its roots in international law appear somewhat less
than deep”, and no tribunal has so ruled. If jurisdiction is to be limited, it must be by concrete treaty rules.
• “There are virtually no substantive rules of liability that are accepted as empirical law throughout the world.” The CIL
of state responsibility was borrowed from European culture and never accepted by Latin America or the post-colonial
states; after 50+ years the ILC is “no closer to a version acceptable to a significant majority than when they began”.
Content varies by culture: on expropriation, Brownlie says compensation yes but not the “adequate, effective and
prompt” formula, against the Restatement; on environment, Akehurst calls the CIL rudimentary and Brownlie doesn’t
mention it. China’s position — a quarter of humanity — is unknown.
(e) “New CIL” and why it fails.
• Premise: near-unanimous resolutions evidence opinio juris, so norms become CIL quickly on little practice. It is the
preferred method of human-rights and environmental advocates and the basis of the Restatement’s norms.
• Bootstrapping: resolutions are recommendations, yet become both opinio juris and practice — so several non-binding
resolutions alone create obligation. “If this process were accepted, then it would constitute a new legislative form of
lawmaking, not CIL.”
• Deeper objection: “The repetition of non-binding instruments cannot create a conviction that simply is not present in any
of the instruments.” The Universal Declaration, Stockholm and Rio Declarations are aspirational. Eleanor Roosevelt on
the UDHR: “It is not a treaty; it is not an international agreement. It is not and does not purport to be a statement of law or
of legal obligation.” Aspirational instruments adopted while states refuse concrete treaties are evidence that states lack
the conviction.
• Environment: language is so general, and paired with countervailing goals (sovereign right to exploit resources), that
both Stockholm and Rio leave the development/environment balance to each state as a sovereign prerogative. Neither
says harm prevention is a legal obligation. The real questions — cost allocation, liability regimes, intergenerational equity
— cannot be settled by vague resolutions. He concedes aspirational texts can create expectations, galvanise domestic
constituencies and shape willingness to accept treaties.
• Human rights: consent-based conventions are the real regime. For non-ratifiers (China) or heavily reserving ratifiers (the
US), “their level of commitment to these treaties is the best evidence” of what they accept. States scrutinise treaties far
more seriously than resolutions.
• States reject the premise. US Legal Adviser Conrad Harper replied to the Human Rights Committee that its General
Comment “asserts in wholly conclusory fashion that a number of propositions are customary international law which, to
speak plainly, are not” — instancing hate speech unaccompanied by any act, and the juvenile death penalty, where “the
practice of States demonstrates that there is currently no blanket prohibition”; “such a cavalier approach… raises serious
concerns about the methodology of the Committee as well as its authority.”
• Non-Western rejection: the Tunis Declaration (42 African states) and Bangkok Declaration (34 Asian states) suggest
most non-Western states see customary human rights as a goal, not a requirement.
• Cultural chauvinism: apart from genocide and slavery, the UDHR rights reflect discrete cultural values. Rights-
consciousness is spreading, but driven by market capitalism, industrialisation, satellite TV and the Internet — with little
evidence states accept them as international legal obligations limiting domestic choices, or that the attitude penetrates
beyond economic elites. Imposing them while the US resists is “a disturbing cultural chauvinism that does not promote
respect for law”. Note his sharp point: the Restatement’s list replicates only those UDHR rights consistent with US
values, ignoring the social rights implying government responsibility for citizens’ welfare.
• No weighting rule: “There is no objective means of determining the magical point when norms… suddenly are
transformed into binding legal obligations.”
(f) “New CIL” as a rule of recognition.
• Theorists invoke Hart (primary rules + secondary rules of recognition, change, adjudication); Franck and D’Amato say the
system has them. The strategic gain: redefine CIL’s requirements while keeping its formal authority.
• Hart disagreed: international law lacks secondary rules of change and adjudication and a unifying rule of recognition; it is
a primitive system of primary rules binding as custom when accepted. “General acceptance” merely repeats the
INTERNATIONAL LAW - MODULE II | PAGE 13
MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
customary requirement. “CIL… has no separate form or ritual that tells us a norm is binding beyond general
acceptance itself.”
• Undefined processes cannot be rules of recognition: how do you weigh resolutions, conference instruments, treaties?
What identifies Bin Cheng’s “instant custom”? Norms like the right to development, democratic governance, harm
prevention and precaution “have not become standards of behavior internalized by world society”.
• What a real rule of recognition looks like: UN Charter Arts. 27(3), 39–42 — the Security Council acts on nine votes
including the P5 concurring, and states specifically consented on ratification, so resolutions bind regardless of
agreement. WTO Agreement Art. IX(2) — authoritative interpretations on a three-fourths vote. In both, states defined the
process by treaty and required consent. These are “clearly evident rituals”; and rituals not only confer authority but
confine it to formal occasions.
• What states could do instead: amend the Charter; make prescribed-form resolutions bind those voting for them; adopt
law-defining resolutions on a supermajority; or sign a treaty creating a less formal binding process. “If nations are willing
to spend months discussing and negotiating language at the General Assembly, time and energy would not apparently be
barriers.” “New CIL” theories abound “because there is a lack of political will to create binding mechanisms.”
Section conclusion: “CIL discourse is normative discourse masquerading as empirical.” Powerful states use it to justify power
without acceptance; activists use it for progressive change. Kelly is sympathetic to the second but says the tool is inapt for both.

3.5 LIMB TWO — disintegration

The paradigm: CIL = general and consistent practice accepted as law; binds all; binds new states; but not persistent objectors.
Every element is controversial.
(a) State practice.
• Practice alone never suffices — it may be a violation or a convenience.
• No agreed list. Traditional: diplomatic correspondence, incident reports, policy statements, national and international
judgments, legal advisers’ opinions, treaty patterns. “New CIL” adds ILC drafts, UNGA resolutions, recitals. The ILC
adds organisations’ practice and common national legislation.
• The choice is normative and distributive. The US State Department’s list stresses government acts, not resolutions —
favouring effective state power and strong centralised governments. “New CIL” stresses normative statements. “The
choice of what practice counts depends on the theory adopted. This choice then determines the existence and content
of customary norms.”
• Expropriation illustration: full-compensation advocates cite BITs; national-treatment supporters cite lump-sum
settlements. Neither evidences conviction — BITs are utilitarian devices to attract capital (and the same states keep
opposing full compensation as the standard); lump sums reflect politics. Empirically the lump sums prove what neither
side admits: there is no consensus. “This duet has played out for nearly 200 years.”
• Acts v. statements. D’Amato and the New Haven school: only physical acts; a normative claim by a participating state
sets a precedent — both “elevate the normative position of a participating state above that of the normative attitude of
community”. Others: statements alone suffice. The stakes are power: acts-only gives law-making to the strong;
declarations-count gives the many a voice. “Choosing what practice counts not only determines the content of norms, it
determines the distribution of lawmaking authority.”
• His own gradings: international judgments are opinions about CIL, not practice; the ICJ’s unpredictability helps explain
“the chronic disuse of international adjudication”. Domestic decisions are of minimal value — domestic courts don’t
analyse other cultures’ practice, adopt their own government’s positions, and overrate the judicial role (praise for
Sabbatino (1964) for judicial restraint; criticism of Chase Manhattan (1981) and Trail Smelter). Resolutions and ILC drafts
are neither behaviour nor statements of acceptance; the few clearly law-declaring ones (Friendly Relations Declaration,
Res. 2625; Definition of Aggression, Res. 3314) are better seen as Charter interpretation.
• Forest and trees: practice is not an end; its purpose is to evidence the community’s beliefs. Henkin’s method — selecting
human-rights provisions from a few constitutions and calling them practice — ignores other constitutions, actual
behaviour, and communities’ convictions.
(b) Consent or consensus — the persistent objector.
• The paradigm uses two inconsistent narratives: binding new and silent states = consensus; letting objectors opt out =
consent.
• History. Classical consent theory (Grotius, Wolff, Vattel): CIL is “voluntary law” resting on tacit consent, applying only
to accepting nations, exitable at will. Vattel (1758) based positive law on actual practice; in the 19th-century US his treatise
was the primary authority, a university textbook, the source most cited by judges and counsel. But his authority vanished
by 1900: few states ever participated or expressed a will; tacit consent is “an obvious fiction”; and states could not in
practice exit. It was also incompatible with new non-Western states (Turkey, Japan).

INTERNATIONAL LAW - MODULE II | PAGE 14

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• Hence the universal/consensus theory on “common consent” — with the practical value of binding the successor states of
the Ottoman Empire to European international law.
• The universal theory had no persistent objector rule until recently. Latin American states were held bound to the law of
state responsibility “despite continuous objections by a large number of nations” — an era “better characterized as one of
domination over the expressed will of many nations”.
• It contradicts customary law itself, which is communally determined and independent of individual will. And if it is to
be part of CIL, it must itself be generally accepted — no evidence non-Western states accept it. Supporters cite only Anglo-
Norwegian Fisheries and Asylum — and Asylum was special/regional custom (D’Amato), which is contractual by nature.
Both are only dicta; other ICJ opinions say all states are bound. Sharpest thrust: many supporters “quixotically” say the
rule doesn’t apply to human rights or fundamental norms — “This subjects some norms to consent, while others,
associated with Western individualism, are not so subject.”
• Incoherent on any theory: if consent-based, new states may dissent; if sovereignty-based, they need not accept limits; if
consensus-based, the idiosyncratic state is as bound as the new state.
• The political history (his best passage). During the Cold War the Anglo-American literature treated CIL as universal —
“with the evident virtue that customary norms were binding on a recalcitrant Eastern Bloc”. The 1965 Restatement had no
persistent objector rule and applied restrictive immunity to the USSR despite objection; Oppenheim’s had none until the
9th edition; Tunkin was “a voice in the wilderness”. Then in the 1960s–70s new states became a majority and the West
began losing control of the territorial sea and investment regimes. “The notion of the persistent objector can be seen as
the Western counter-reformation to this revolution.” The Restatement (Third) Reporters, after consulting a State
Department hostile to the ICJ post-Nicaragua, adopted it for the first time. Ted Stein: since the Reporters held CIL to be
federal common law, they faced “the rather horrifying prospect of hostile majorities at the U.N. making federal law” —
the objector rule “provided the necessary escape device to preserve U.S. sovereignty.”
• The paradox: custom is a consensus concept everywhere else, and was applied as such to new states, silent states and
non-consenting Latin American states — “Yet powerful nations, utilizing the persistent objector concept, will not accept
normative principles with which they disagree. A universal theory has only been acceptable to the more powerful states
when they control the normative process.” On China he is blunt: “although China’s human rights record is abhorrent, a
state has as much right to dissent from these norms as does the U.S.”
(c) Disintegration. The indeterminacy is not discretion in application but that “there is no common understanding of how to
determine customary norms” at all. The elements are “empty vessels in which to pour one’s own normative theory”. Every theory
— D’Amato’s acts-only, McDougal/Reisman’s policy jurisprudence, Meron/Sohn’s new CIL, Henkin’s domestic constitutionalism
— has the same dilemma: making norms from your own view of a better order destroys the authority needed for compliance.
“CIL has degenerated into pure subjectivity.”

3.6 LIMB THREE — procedural legitimacy

(a) Democratic legitimacy.
• Process values (participatory governance, procedural rationality, per Summers) matter in themselves — and especially
internationally, because rules that don’t match a culture’s values may still be accepted if the procedure is seen as fair.
• “The majority of nations and peoples of the world rarely participate in the creation of customary rules that limit their
policy choices and sovereignty.” This “democracy deficit” infects traditional and new CIL alike.
• The methodologies destroy participation. Under the “claims” approach, protest is not an act and cannot stop formation,
so the normative element reduces to the justifications of states able to exercise power. Example: a few incidents of asserted
humanitarian intervention are said to create a customary exception to the Charter prohibition on force, widespread
protests deemed irrelevant.
• “New CIL” doesn’t cure it: states consent to abstract recommendations, not binding norms, and don’t decide which
become obligations.
• Silence proves nothing. “Silence in the absence of an obligation to speak does not imply consent.” A system where
silence speaks requires perfect knowledge, unlimited resources and awareness of legal consequences. Analogy: in contract
law silence is not acceptance (Restatement (Second) of Contracts §69), except where the offeree takes offered benefits or
invited reliance — “the seller cannot force the buyer to respond or be bound by silence.”
(b) The new-states dilemma.
• Theorists justify binding them by analogy to newborn citizens: D’Amato’s “inherited entitlements”; Franck’s associative
rights and obligations of membership.
• The analogy dies once you add the persistent objector: individuals in domestic societies cannot opt out of laws they
dislike. And if the system is consensual, new states were denied the opportunity to object that others enjoyed.

INTERNATIONAL LAW - MODULE II | PAGE 15

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• The historical effect: post-colonial states are bound to pre-existing custom against their interests, while older states can
block formation or exempt themselves. New states particularly objected to the state-responsibility limits on expropriation,
perceiving that much of the property was acquired by force in the colonial period.
• But they accepted the structural norms — statehood, boundaries, immunities, allocation of competence — because these
are genuinely reciprocal. State responsibility “had a differential impact on capital exporting and capital importing
countries. Different interests created different perceptions and positions that were incompatible, not reciprocal.”
(c) International tribunals.
• Much “CIL” is judge-made; scholars weight dicta, concurrences and even dissents heavily.
• Implied delegation fails: ICJ jurisdiction is voluntary; decisions bind only the parties and are not precedent (Art. 59); and
the Statute unusually defines the sources. “Nations delegated to the Court only the authority to apply custom as a fact,
not to create law under the guise of CIL.” And states show it: the Court is rarely used, and non-consenting respondents
often fail to appear or comply. “States register their opposition to the uncertainty of judicial lawmaking by not utilizing
the Court.”
• Revisability fails: the Assembly only recommends; the Council is limited to peace and security and is veto-bound; only a
multilateral treaty can correct the Court, at high cost and with no consensus available.
• “No harm” fails: creative lawmaking “may diminish respect for the Court’s judgments, even when its authority is clear,
thereby undermining the entire system”.
• Special capacity fails: the two-chamber election over- and under-represents states; and “the few international judges…
cannot fairly represent the rich diversity of international society”.
• The common-law analogy is inapt: common law reasons from precedent within one culture; custom asks what a society
in fact regards as required, “even if it is irrational and short-sighted”.
(d) Efficacy and the case for treaties.
• Compliance follows perceived legitimacy; Tyler’s empirical work shows procedural justice provides a “cushion of
support” for unfavourable outcomes.
• Four procedural failures:
1. Interests unrepresented — under acts-only, only the powerful count.
2. No open discussion — the process is a clash between two states around a contentious event, encouraging “political
statements to bolster positions at the expense of reasoned dialogue”. State responsibility “was historically imposed on
Latin American nations often by force of arms”.
3. No negotiation or trade-offs. The Law of the Sea is his proof. The Anglo-American literature long asserted a three-mile
customary limit; developing states extended claims to exclude advanced fleets; developed states wanted fisheries access
and naval mobility. By 1958, only 21 of 86 attending states supported the asserted customary rule — yet the US and
Japan would not budge. The treaty process produced compromises for nearly everyone: 12-mile territorial sea, a new 200-
mile EEZ protecting developing states’ fisheries, and a special regime for international straits protecting military access.
Policy problems are polycentric; negotiation can reach acceptable, if imperfect, solutions.
4. Injustice reduces compliance pull. Expropriation is “a striking example”. Disputes settle by lump sums far below the
asserted standard; only recently has an investment regime emerged through BITs, which are two-party and “inherently
fragile”; a comprehensive Uruguay Round agreement failed because developing states refused full compensation and
developed states refused obligations on investors. “Seventy years after the Chorzów Factory case, there continues to be
no consensus on an international standard for compensation for expropriation.”
• What he concedes exists: agreement on structural norms, and “a limited number of foundational norms of human
dignity… such as genocide, ethnic cleansing, and racial segregation”. Beyond that, sovereignty-limiting norms need
actual agreement.

3.7 Conclusion and the objections answered

• Against Franck: CIL lacks determinacy, symbolic validation, coherence and any secondary rule of recognition.
• The deepest incoherence: states want to bind others universally without being bound themselves. “The trump card of
actual consent is reserved for norms one dislikes.”
• Objection: necessity. Answer — necessity can’t justify norms without acceptance; and if states really had the conviction,
they could sign a treaty. Modern communications make negotiation routine, and the genuinely accepted (structural)
norms have largely been codified. Illustration: developing states adamantly opposed any IP agreement in the Uruguay
Round; after years of compromises and trade-offs, TRIPS was accepted by all members.
• Objection: progressive norms. Answer — this is “an argument for the incorporation of preferred values into international
law, without the inconvenience of democratic processes.” Western traditionalists (humanitarian intervention, full
compensation) and human-rights/environmental activists use remarkably similar techniques. “Customary norms can be
progressive or reactionary… depending upon who is spinning the tale. If any possible norm can be justified, then no

INTERNATIONAL LAW - MODULE II | PAGE 16

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
legitimate norms are possible.” Genocide shows the alternative: Nuremberg, then a near-universal treaty — “However, it
was not always so.”
• The historical debt: the difficulty in building an investment regime rests with “the historical failure of the Western
nations to respect the rights of property and life when effectively expropriating the land… of the indigenous peoples of
Africa, Asia, and the Americas” — Europe’s “largest land grab in history” in Africa (Pakenham, The Scramble for Africa)
ran in the same period as the US pressed a minimum standard on Latin America.
• The ozone model: the Vienna Convention (1985) and Montreal Protocol (1987); India and China declined without
financial support, so the London Agreements (1990) created a multilateral fund for transition costs. Progressive norms
can be negotiated.
• Four systemic reasons to abolish CIL: it breeds cynicism and disenfranchisement; it does not encourage compliance; it
creates exceptionalism (powerful states conjuring exceptions to the prohibition on force; weaker states and advocates
conjuring norms from resolutions); and better means exist — trade-offs, incentives, World Bank/IMF leverage, changing
public attitudes. “Patience, as well as vision, is a virtue.”
• Final position: practice remains “the tentative steps in the ongoing process”. “The definitive articulation of a norm
requires a ritual” — an authoritative court decision or treaty ratification — specifying the norm’s existence, content,
effective date and bound states. Scholarship “should be more modest”.
Contemporary: (i) Vindications. The objector-as-Western-escape-hatch thesis is visible in the “no” votes on A/80/L.65 (2026); the
ozone/Kigali trajectory played out as he predicted; and the ILC’s Articles on State Responsibility remain uncodified 25 years on. On
investment, the fragility of BITs is now open crisis — EU and member-state withdrawals from the Energy Charter Treaty (2023–25) and
UNCITRAL Working Group III ISDS reform are exactly the consent-based renegotiation he urged. (ii) Difficulties. His “if it’s real, put it
in a treaty” claim has fared badly: the plastics negotiations collapsed at INC-5.2 (Geneva, Aug 2025) and the WHO Pandemic Agreement
(May 2025) took years and left key annexes unfinished. Where treaty-making stalls, custom returns by necessity — Jennings’s answer to
Kelly. (iii) The sharpest test — the ICJ Climate AO (2025). On his method the Court did what he condemns: it derived a universal
customary duty largely from treaties, declarations, prior judgments and principles rather than an inductive survey. But the request came
from an Assembly resolution tabled by Vanuatu with cross-regional co-sponsorship and a near-record volume of submissions — the wide
deliberative participation whose absence he said delegitimised CIL. Progressive scholarship reads it as the Global South using the
customary form against the powerful for once; a Kelly-faithful reply is that a norm’s direction cannot cure a defective process.
Exam handles: CIL claims custom’s authority while abandoning custom’s defining feature (the community’s internalised, observable belief). Its method
is deduction dressed as induction. The objector/new-states pair is incoherent — consensus for the weak, consent for the strong. The remedy is ritual:
treaties and defined supermajority processes.

4. PANEZI — “SOURCES OF LAW IN TRANSITION: RE-VISITING GENERAL

PRINCIPLES” (2007)
Thesis: General principles have gained a decisive role in international economic law, and this strengthens their position as a
source generally. The two features usually treated as defects — their flexibility and the notion of civilisation in their definition —
are exactly what makes them useful.
Two temperaments they attract: a “blanket of safety” (a last resort when nothing concrete regulates an issue) versus “an open
door through which unpredictable legislation might enter”. This tracks attitudes to the state: those who see them as a tool for
innovation are less attached to sovereignty; the suspicious want the state at the centre of law production.

4.1 History
• Little used because traditional international law focused on treaty and custom, and because states distrusted a source “not
so clearly based on state consent” but looking like something judges introduce when nothing concrete exists (Friedmann).
• Already recognised at the Hague Conferences of 1899 and 1907: developed legal systems trying to coordinate needed a
common denominator as a means of communication when physical distances were hard to bridge (Mosler). Then written
into the PCIJ Statute.
• The travaux show the drafters did not think they were innovating. They wanted the Court to draw on as many legal
orders as possible; they worried about the incompleteness of treaty and custom; they wanted to avoid non liquet without
letting judges make their own law. They regarded general principles as law, and all their examples came from domestic
legal systems.
• PCIJ v. ICJ texts: identical wording of the sources; the ICJ adds “whose function is to decide in accordance with
international law…”, and ex aequo et bono becomes a separate paragraph 2.
• Textual analysis (VCLT Art. 31 method) — four elements: (a) general; (b) principle — not a conventional rule, not a
general practice accepted as law; (c) already recognised, not in the making; (d) recognised by “civilized nations”.
“General” may mean not regional, or more abstract and substantial, or simply not special. Civilisation and principal legal
systems reappear in Article 9 of both Statutes (composition of the bench).

INTERNATIONAL LAW - MODULE II | PAGE 17

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• Democratic reading: the inclusion of general principles, like Art. 9’s representativeness, was “one of the demonstrative
facts of the democratic legitimacy attempted by the founders of the League of Nations” — though the veto made the
League dysfunctional and the war ended the experiment.
• Not a dead letter today, despite the proliferation of legislation. She invokes the ICJ in Aegean Sea Continental Shelf
(Greece v Turkey), para. 80: judges must interpret rules as they are now, not at the time of drafting. So their function is
not fixed by the drafters’ motive (fear of non liquet, largely solved by proliferation); they can find a new position today.

4.2 Definition
• They are law. Drafters’ intent; and structurally, (a), (b) and (c) are all classified as “international law”, with (c) in its own
paragraph, expressly distinct from custom.
• Praetorian point: a court need not explain where it found a principle. “It can only invoke it, identify it as a general
principle, that being enough to provide for legitimacy.”
• Mosler — they are autonomous, but their autonomy is linked to national orders: they are produced by general
recognition in domestic fora yet are not applied internationally in their “domestic” capacity. “They have two hats, the
domestic and the international, and they change, depending on where they are used.” They form “an autonomous
group of norms… inspired by the ius gentium”. The judge has a creative role: he must shape the principle without
departing from its core — “the norm which he applies is a norm of international law, taken from principles observed in
domestic legal orders and adapted by him to the particular needs of international relations.”
• Hanessian (arbitration): they either assist in applying national law or are themselves the applicable substantive law —
autonomy is clearest in the second case.
• Autonomy matters because it means a legal problem can be solved solely by a general principle. But autonomy is context-
relative: her example is the MFN clause in the WTO, recognised as a general principle underlying the whole WTO–GATT
system yet not a codification of, or relevant to, custom.
• The scholarly definitions (Bassiouni’s catalogue), each with her critique:
• Lauterpacht — not principles of moral justice, not equity in the ethical sense, not speculative law; rather “those principles
of law, private and public, which contemplation of the legal experience of civilized nations leads one to regard as obvious
maxims of jurisprudence of a general and fundamental character”. Critique: an amalgam of equally vague areas; and he
adds that in practice they are the general principles of private law.
• Cheng — “cardinal principles of the legal system in the light of which international law is to be interpreted and applied”.
Critique: almost circular; the one deduction is that they are fundamental to a legal system.
• Schlesinger — “a core of legal ideas which are common to all civilized systems”. Critique: not circular, but enters the
world of ideas — which may be right, since it is custom that must be found in practice, so general principles “can be only
ideas, notions not escorted by any practical evidence”.
• Verzijl — principles “so fundamental to every well-ordered society that no reasonable form of co-existence is possible
without their being generally recognized as valid”. Critique: this overestimates them (echoes of Rousseau’s social
contract). Coexistence rests on all three sources. Her own reflection: coexistence “is a mixture of real and normative
factors. We coexist because the earth is of finite dimensions and we inevitably find ourselves in a pragmatic proximity” —
states could not distance themselves because they shared borders; the normative layer came later, as they turned patterns
of interaction into rules making proximity bearable. “General principles are not the glue that holds us together.”
• Friedmann — the “maximum measure of agreement on the principles relevant to the case at hand”. Critique: also circular
but adds agreement. They resemble self-evident facts: “an extreme skeptic might challenge them, but most agree to
them.” They have a fact-like function — taken as granted so we can get on with what we disagree about.
• Gutteridge — derive from private law, recognised in substance by all major systems, must not violate those systems’
fundamental concepts. Critique: creates two domestic categories (principles v. fundamental concepts) without explaining
the difference.
• Jalet (her favourite) — “that unformulated reservoir of basic legal concepts universal in application, which exist
independently of institutions of any particular country and form the irreducible essence of all legal systems”. Critique:
closest to both the drafters’ thinking (a reservoir ensuring lacunae never prevent decision) and a modern interpretation.
• Her seven features: (a) part of international law; (b) more or less autonomous; (c) they promote creativity by judges; (d)
fundamental to a legal system; (e) they may be “ideas”, so have a philosophical aspect — but “normative notions” is the
better term, since it shows they are law; (f) only one of three sources; (g) a reservoir.
• Her definition (learn verbatim): “General Principles are an autonomous, created by general consensus, systemically fundamental
part of International Law, that consists of different normative notions, in which judges refer to, through a creative process, in order to
promote the consistency of International Law.”

4.3 Hierarchy
• Clearly above rules not mentioned in Art. 38 and above Art. 38(1)(d) judicial decisions and publicists.

INTERNATIONAL LAW - MODULE II | PAGE 18

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• Their relation to treaty and custom is open. The Statute gives no guidance; the existence of jus cogens shows there is
some hierarchy among rules. Mosler says the Statute “indicates the order the international judge has to follow”; Panezi
replies this cannot be inferred from the mere order of listing.
• Mosler’s position: they can be used in applying treaty and custom without becoming auxiliary; they are examined after
the other two, but their legal quality is the same, and they may “even be stronger… if a peremptory norm belonging to
the general principles group supersedes contractual provisions or customary rules.”
• Restatement (Third) §102(4): general principles “may be invoked as supplementary rules”; comment (l) makes them
independent only where treaty and custom are silent or leave lacunae; the Reporters call them an “independent, though
secondary source”.
• Panezi’s own resolution: is conflict even possible? If general principles come from general agreement, and custom needs
practice plus opinio juris, “it is rather difficult for states to create customary international law which opposes the General
Principles they accept”. Equally hard to conclude a treaty contradicting the principles underlying your own legal order —
“One does not refute their own principles.”
• General principles surface when the other sources are insufficient (lacunae) or have become so complex and numerous
that coordination is needed to decide which rule applies first. In both cases their role is as essential as custom’s or
treaty’s.
• Conclusion: the three sources are equal, autonomous and independent; general principles can be a primary source. If
two sources together answer a question, citing both — in any order — only adds legitimacy.

4.4 Categories
• Mosler’s three:
1. From generally recognised national law — liability/responsibility/reparation; unjust enrichment; property, expropriation,
indemnity; denial of justice; right of passage; prescription; error; implied agreement; preclusion by conduct;
presumptions; general principles of administrative law; procedural principles (following one’s own precedent; res
judicata; res judicata affecting distribution of powers in other UN bodies; requirements of judicial process — equality of
the parties; admission of indirect evidence).
2. Originating in international relations — (a) from the specific features of the international community: protection of
diplomats and consuls; sovereign immunity; primacy of international law and the limits between domestic and national
jurisdiction; territorial jurisdiction; (b) generally accepted: freedom of the high seas; diplomatic protection and exhaustion
of local remedies; laws of war; delimitation of sea boundaries, equidistant lines; (c) principles connected with treaties; (d)
elementary considerations of humanity.
3. Applicable to all kinds of legal relations — good faith, pacta sunt servanda, unilateral declarations; equity; estoppel;
principles of judicial procedure; respect for basic human rights.
• Panezi highlights the third category: the “systems” Mosler contemplates include municipal, inter-state, institutional and
other autonomous systems — so a principle can derive from any established regime or from the activity of other actors.
“This inclusion by Mosler is quite worthy of note.”
• Friedmann’s three, which she finds “more clear-cut”: (i) principles of approach and interpretation; (ii) minimum
standards of procedural fairness; (iii) substantive principles widely recognised in the leading systems. So general
principles have an interpretive, a procedural and a substantive aspect — which adds to their autonomy.

4.5 ICJ jurisprudence (Mosler’s survey)

• South-West Africa (1950) — Judge McNair: international law borrows domestic principles “not by means of importing
private law institutions ‘lock, stock and barrel’ ready-made and fully equipped with a set of rules.” (The single most
quotable line on 38(1)(c).)
• Chorzów Factory — reparation as a general principle; expanded in the advisory opinion on Application for Review of
Judgment No. 158 of the UN Administrative Tribunal.
• Barcelona Traction — property, expropriation, indemnification; denial of justice.
• Right of Passage over Indian Territory — right of passage. Temple of Preah Vihear — plea of error.
• North Sea Continental Shelf and Interpretation of the WHO–Egypt Agreement — principles of contentious proceedings.
• Tehran Hostages — principles of international relations. Corfu Channel — freedom of maritime communication.
• She then asks whether tribunals still see themselves as conditioned on state consent. Many do not: WTO panels (reports
binding once adopted by the DSB) and UN special tribunals — “Their purpose is not to please states.”

4.6 The two sources of suspicion

(a) Judicial discretion. Many see general principles as a carte blanche (Jalet). Ford brings in Justice Aharon Barak: with general
principles, discretion is “not just a possibility but an inevitable situation”; Barak advises judges to step outside themselves and
examine the social reality rather than impose their own values. Ford then asks whether general principles invite extensive

INTERNATIONAL LAW - MODULE II | PAGE 19

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
comparative law analysis to test universality across representative systems. - Panezi’s critique: this “system-representative”
method is state-centred and “old-fashioned since it is based on… an arbitral geographic classification.” Sovereignty is stepping
aside as organisations regulate more — the IMF imposes standards of legislative conduct through conditionality; WTO panels
and the Appellate Body issue binding reports once adopted. - Her reorientation: a judge should ask not which geographical system
the issue belongs to but what area of activity it belongs to. On an international-investment problem unsolvable by treaty or custom,
the judge should not survey every country’s law for a common denominator; he should make the decision systemically
compatible with the general principles governing international financial transactions — citing Metalclad v Mexico (2000), Pope
& Talbot v Canada (2001), S.D. Myers v Canada (2000, 2002). - Ford is right about representativeness, but there must also be a
shift from a state-sovereign system to one where non-state actors participate. This shift both locates general principles more
easily and enhances their effectiveness, since their source is more geographically diverse — near-global membership of
organisations, plus the voice of civil society. - Legitimacy cuts both ways: participation of interest groups and civil society, even
consultatively, is more democratic; but accountability questions arise about those very actors, raising a new scepticism about how
solid the principles are.
(b) “Civilized nations”.
• Bassiouni dismisses the issue: all UN members today are civilised. Panezi’s demolition: what about non-members?
Switzerland only joined in 2002 — it did not become civilised then. Yugoslavia’s membership was suspended in the
1990s, so on Bassiouni’s standard it was civilised, then stopped, then resumed. “This is logically flawed.”
• His second argument — that “civilized” signals a blend of positivism and naturalism, or of civil and common law — she
accepts “in one sense”: the “general” element requires deducting contested notions to make them acceptable, so it is a
compromise; and since the common/civil division is fairly accurate, general principles plausibly are the means of finding
notions common to both. His point that principles are not value-free and so are a tool for prioritising is “a good guide to
their functional significance”: where rules proliferate, rules more compatible with the linking principles outrank less
harmonious ones.
• Fidler’s “standard of civilization” — the provision survives with a fresh interpretation.
• The old standard was “the structuring principle for the expansion of… the Western civilization” into the non-Western
world: basic rights for foreigners; an organised, predictable bureaucracy; a “rationalized” legal system; diplomatic
capacity; the obligation to respect international law; adherence to developed societies’ norms. (She flags Rajagopal,
International Law From Below, for the opposite view.)

4.7 General principles of international economic law

• The old standard and the market: it provided for liberalised commerce, empowered private actors and limited
government’s role in trade, and forced non-Western countries into economic intercourse with the West while
demanding obedience to international law. “This system mainly worked through imperialism and capitulations.” An IR
alternative is solidarism — the society of states rests on political, economic and cultural like-mindedness within states.
Under the old standard a state needed only to be “Westphalian”, not fully Western.
• The new standard emerged after the Cold War, in “a globalized system of multiple actors (states, NGOs, international
organizations, multinational corporations, credit rating agencies)”. “Civilized” was redefined for “the phase of liberal
globalized civilization”, and necessarily includes market-oriented reforms: free trade, market economy, democratic
governance, good governance, the rule of law and human rights. In state–market terms: trade and IP liberalisation;
investment liberalisation through BITs; structural adjustment by the IFIs; and international anti-corruption legislation.
“General principles of International Economic Law are in effect structural legal principles that guarantee the regular
operation of the market economy.”
• Threshold test — patterns are not principles. A market regularity becomes a general principle only when law assigns it
a normative role. Her example: credit rating agencies exist independently of law and their ratings are not legal rules —
another institution may lend to a badly rated party. But if the IMF cites one of those systems in deciding to grant or
deny credit, the market pattern acquires legal significance — and the most reasonable label is a general principle.
Three growth areas:
• International trade law (WTO). 151 members of 191 states (~80%). Panels treat prior reports as persuasive authority;
there is no stare decisis. But rules crystallise: panels reuse “pre-existing definitions, interpretations, logical methods and
schemes” for new problems, and a “rule of thumb” emerges. “If in a legal system an expectation develops on the
repetition of previous interpretations, then they receive a normative effect that goes beyond the two or more parties
before the Court and becomes a principle.” It cannot be custom — even the “customary practices” in WTO Agreement
Art. XVI:1 lack opinio juris. (Political detail: the judicial body had to be called “panels”, nothing near “Court”, or the US
Congress would not have ratified.) Once the DSB adopted reports, states developed “the GATT acquis” creating
“legitimate expectations” (Weiler). Given DSU Art. 3.2’s goal of “security and predictability”, rules that are neither
custom nor treaty “very well could be General Principles of international trade law.”

INTERNATIONAL LAW - MODULE II | PAGE 20

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• International investment law (BITs). Over a thousand signed since 1959, 800+ in twenty years. Even if not custom
(Guzman), identical provisions “develop at least to become General Principles”. Gudgeon: the growing network
“constitutes an important new element of State practice”. Panezi’s formulation: “when a provision appears in every BIT,
then there is a heightened normativity in this rule, because we can infer a consensus on its existence.”
• Fair and equitable treatment (at least equal access to administrative and adjudicative procedures regardless of
nationality), usually with full protection and security.
• Expropriation — Schwebel on the US 2004 Model BIT: Art. 6 is supposed to reflect custom, but the content is highly
controversial; wording across BITs is “closely parallel, if not identical”.
• The Hull formula (“prompt, adequate and effective”) appears in about two thirds of treaties. Not custom, since not all
states agree — but part of it creates “an umbrella for investors, an expectation of similar treatment” under the other third,
and states cannot abstain because (i) investors expect it before investing, and (ii) peripheral states seek uniformity with
larger states to enter the global market. “Even if Hull is not customary law, it might very well be part of a general
principle of international investment law.”
• Lowenfeld’s line to memorise: an agreement to arbitrate under particular rules, or a prohibition of performance
requirements, are not general principles and depend on the treaty text. But “the understanding that international law is
applicable to the relation between host states and foreign investors, that expropriation must be for a public purpose and
must be accompanied by just compensation, and that disputes… should be subjected to impartial adjudication or
arbitration are general principles, and do not depend on the wording or indeed on the existence of any given treaty.”
• Lowenfeld goes further (BITs have ripened into custom). Panezi is cooler: custom “is rather more concrete and it is likely
for states to contest this approach”, whereas general principles are more flexible and need no opinio juris with practice
— so if we want to recognise normativity in BITs, general principles are the safer common ground.
• Weiler on NAFTA Art. 1105: these rules show “the existence of a unified body of international economic law
principles”, permitting us to speak of international economic law “as a coherent set of norms, rather than a loosely
associated set of treaty rules”.
• Lex mercatoria. Generally accepted international commercial law principles; a field of alternative norms adapting to the
market. National law is inadequate for a dynamic, multinational field, and national legislation leaves windows open: US
UCC §1-303 defines usage of trade as “any practice or method of dealing having such regularity of observance… as to
justify an expectation that it will be observed”. The Iran–US Claims Tribunal may apply “such choice of law rules and
principles of commercial and international law as the Tribunal determines to be applicable”, taking account of “relevant
usages of the trade, contract provisions, and changed circumstances” — and it actually invoked unjust enrichment, force
majeure, account stated and clausula rebus sic stantibus. ICC Rules Art. 13(5) requires arbitrators to take account of
contract provisions and trade usages, which Crook says “could be read to permit ICC arbitrators to decide wholly on the
basis of generally accepted international commercial law principles (the celebrated lex mercatoria).”

4.8 Conclusion
• Globalisation makes crises spill over: “There is no local economic crisis” (she instances the East Asian currency crisis) —
the result of coordination through globalised regimes and near-universal IFI membership.
• Spengler (1922): “world-economy is the actualized economy of values that are completely detached in thought from the
land, and made fluid.” If true then, the modern economy “has long ago ceased being fluid; it is so complex that it consists
of values only comparable to untamable vapors.”
• So the consolidation of general principles of international economic law gives the necessary common starting point,
“consequently transforming the traditional notion of General Principles of international law”. Invoking Luhmann on
the economic system setting the pace of “world-society building”, she ends: “Whether this is for better or for worse, is
something that only time can tell.”
Contemporary: (i) Her bet was overtaken from another direction — the ILC’s Draft Conclusions on General Principles of Law
(concluded 2025) confirm the two categories (transposed from national systems; formed within the international legal system), require
comparative analysis plus a compatibility check with the international legal system, and drop “civilized nations” as a colonial
anachronism. So her Fidler-based defence now reads as a description of an ideological function, not a recommendation. (ii) Her WTO
thesis has been complicated by the paralysis of the Appellate Body since 2019 and the makeshift MPIA — the very “GATT acquis” she
treated as generating principles was frozen because a powerful member objected to panels behaving like a court (Kelly’s warning in
miniature). (iii) In investment law, UNCITRAL Working Group III, the ECT withdrawals, and continuing FET/legitimate-expectations
awards keep “heightened normativity” alive — though Global South scholarship reads that same normativity as Fidler’s “new standard of
civilization” entrenched by arbitral rather than legislative means. (iv) Her point that the IMF’s use of private ratings converts market
patterns into legal principles is now mainstream in sovereign-debt scholarship and underlies calls (UNCTAD; the 2025 Sevilla Financing
for Development conference) for a multilateral debt workout mechanism.
Exam handle: General principles are an autonomous, judge-mediated reservoir of normative notions, equal in rank to treaty and custom, operating
where the others are silent or over-complex; their modern growth area is economic law; their historic problem is the ideology inside “civilized nations”.

INTERNATIONAL LAW - MODULE II | PAGE 21

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
5. WHITE — “EQUITY: A GENERAL PRINCIPLE OF LAW RECOGNISED BY CIVILISED
NATIONS?” (2004)
Question: is equity a general principle under Article 38(1)(c)?

5.1 Two framing cautions

• Equity in international law is not co-terminous with municipal equity — common, Roman, civil or Germanic. Generally
it “tends to suggest justice attained through what is fair”.
• The judges are differently situated. Every society has its own conception of justice, and these do not necessarily coincide
(Rosenne). Domestic judges are shaped by and shape their society. “That cannot be true for judges of the International
Court” — by Article 9 of the Statute they represent the main forms of civilisation and the principal legal systems. Values
may be shared, but “the paths to the desired curial outcome will be various.”

5.2 Genealogy
• Earliest traces: Hittite treaties of the 14th–13th centuries BC, which pre-empted dishonesty in performing strict treaty
terms by specifying acts of bad faith incompatible with the parties’ oaths. (Psalm 98 praises the Lord because “You have
established equity, justice and right”.)
• Across cultures: Greeks — clemency; Romans — aequitas/equality; ancient Chinese law — compassion; Hindu philosophy
— righteousness; some Islamic schools — istihsān, to avoid undue hardship from applying the law.
• Equity as received into international law is most closely related to Western traditions, because the rules developed in
Europe after the Peace of Westphalia (1648) and with 19th-century statecraft.
• Aristotle is the root (Nicomachean Ethics V.10): the universality and completeness of law, plus the need for systemic
correction of the shortcomings caused by that very generality. This generates the enduring question: is equity of the
law or outside it? — and it entails discretion both as to application and extent.
• Roman law — the jus honorarium. Magistrates (praetors), advised by judges, issued edicts aiding, supplementing or
correcting the civil law:
• Aiding — the interdict letting a civil-law heir obtain possession of the deceased’s goods.
• Supplementing — the widow of a man dying intestate without blood relatives could claim his property though not his heir.
• Correcting — where the formal requirements of a valid will failed, the nominated heir might still be recognised.
• Stein: by the early 3rd century AD the jurists realised there was little more to add — “They knew when to call a halt”,
and codification began.
• England: the Chancellors “never called it a day”, merging with the common law via the Judicature Acts and continuing.
Kitto’s account: the Chancellors could proceed by common law or “according to the rule of equity” where the common
law fell short — and for the latter “they had no guidance or point of reference save in their own opinions as to the
standards of conscientiousness”, “for ‘equity’ was as vague as aequum et bonum”. Swelling business plus a taste for
consistency drove them to precedent, so principles emerged. Maitland, quoted: but for the Star Chamber and Chancery
the old national law “would have utterly broken down, and the ‘ungodly jumble’ would have made way for Roman
jurisprudence and for despotism. Were we to say that equity saved the common law, and that the Star Chamber saved
the constitution, even in this paradox there would be some truth.”
• The law of nations: Grotius adopted the twofold Aristotelian idea (equity as understanding what is right and as
corrective moderation of the general law). Both he and Pufendorf saw the tension in judicial discretion: Grotius said
equity must be applied “with an abundance of circumspection”; Pufendorf set a standard of prudence — “sentiments
echoed in modern international law jurisprudence”.
• Civil law systems: codification worked against a separate equitable stream. Instead general clauses ensuring equitable
interpretation against formalism were built into the codes — making those systems “rather closer to the Aristotelian idea
of equity”.

5.3 The 1920 drafting — how equity got in through the back door
• Art. 38 descends from Art. 35 of the PCIJ Statute, incorporated without relevant change.
• There is no express reference to equity. And ex aequo et bono under Art. 38(2) is not equity: a decision ex aequo et bono
“would be against the law”, implies law creation by the Court, and no party has ever agreed to it.
• Equity was proposed for express inclusion and rejected.
• The Committee included Lord Phillimore (UK) and Elihu Root (US — Secretary of State, 1907 Hague delegate, first
President of the Carnegie Endowment, PCA member, Secretary for War, Senator for New York, Nobel Peace Prize 1912),
plus the Japanese Ambassador to Belgium.
• “Civilized nations” had meaning then, “which need not be spelt out here. It bears no real meaning in international law
today” — with 189 UN members, the relevant principles are “those generally recognised across a range of different legal
systems”.
INTERNATIONAL LAW - MODULE II | PAGE 22
MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• “General principles” covers both principles of international law and principles common to international law and
municipal systems (Hudson).
• The conflict: the common lawyers (Phillimore and Root) wanted sources restricted to conventions and custom — to
induce states to accept jurisdiction, and to avoid unconstrained judicial power. The continental and Latin American
representatives wanted to avoid non liquet, “anathema to those from code systems”. “In over-simplified terms, it was a
struggle between legal positivism and natural law theorists” — though Phillimore suggested the real difference was
about the role of the judge in common law and civilian traditions.
• The compromise: Baron Descamps (Belgium) confined 38(1)(c) to principles common to all states — the “fundamental
laws of justice and injustice” — to limit judges’ liberty and prevent subjective reasoning. He included “objective
justice”, which he would have called equity but for the risk of misunderstanding. Ricci-Busatti (Italy) added that
judges applying general principles would not create new rules but apply existing ones. Phillimore said all the principles
of the common law were part of international law.
• Why equity was rejected as an independent source: the different meanings it bears in different systems (van Dijk); and
its “sometimes disastrous inclusion in the terms establishing numerous important arbitrations as an undefined standard”
— the Venezuelan Preferential Claims (1904) and the Jay Treaty Arbitrations (1794). Some wanted express reference (de
Lapradelle, Ricci-Busatti); others thought it too vague and fraught (Hagerup; de Lapradelle resiling). Phillimore would
accept “maxims of equity” but opposed equity generally — too much liberty for the judge unless the technical English
meaning were adopted — while accepting good faith and res judicata as general principles.
• The settled understanding: equity itself is not an independent source; but particular equitable principles as recognised
in the various systems can operate as general principles. There was “a failure to ‘colour in’ the words” — no content
was given. The result accommodated both fears. What they meant to exclude was “pure” equity operating against the
law.

5.4 The three registers

• Infra legem — where a law bears more than one interpretation, equity finds the one best serving the law’s purpose.
• Praeter legem — where the law is silent, equity brings the case within the law so that the states’ intention can be
implemented.
• Contra legem — where the text goes against its real intention or purpose. A tribunal needs explicit powers to do this (her
example of such an express grant: the special agreement between Canada and the Cayuga Indians (1926)).
• Reopening the question: Judge Manley Hudson in Diversion of Waters from the River Meuse (1937): “under Article 38
of the Statute, if not independently of that Article, the Court has some freedom to consider principles of equity as part of
the international law which it must apply.” Vigorously debated — some see it as impermissibly empowering judges to
depart from recognised principles (Lapidoth; McWhinney; E. Lauterpacht).
• Confirming 1920: the Norwegian Ship Owners’ Claims (1922) arbitrators, told to apply “law and equity”, said the words
cannot bear their Anglo-Saxon sense and mean “general principles of justice as distinguished from any particular
system of jurisprudence or the municipal law of any state”. Two commentators call this “the spirit of the law”.
• How the Court behaves: judges rarely say they are applying equity as a general principle; it is taken for granted that
rules like estoppel and “he who comes to equity must come with clean hands” are part of international law — consistent
with drawing on doctrine common to “civilised nations” without importing system-specific technicality.
• The constraint (Rosenne): equity does not automatically correct a decision where strict law produces an unsatisfactory
result; a principle of equity operates only if generally recognised across legal systems.
• Frontier Dispute (Burkina Faso/Mali) (1986): the Court would not use equity to modify an established frontier inherited
from the colonial powers. Equity “is a direct emanation of the idea of justice”, but not “an arbitrary concept of
‘fairness’” interposable at will. Where the boundary did not precisely delimit a water pool, the line “should divide the
pool… in two, in an equitable manner. Although ‘equity does not necessarily imply equality’ where there are no special
circumstances the latter is generally the best expression of the former.”

5.5 Three equitable doctrines in the case law

(a) Estoppel / acquiescence
• Serbian Loans (PCIJ 1929) — Serbia was to repay French bondholders in gold francs; the bondholders accepted paper
francs for a time; Serbia pleaded estoppel. No estoppel: there was “no clear and unequivocal representation by the
bondholders upon which the debtor State was entitled to rely and has relied”, and “no change in position on the part of
the debtor State”.
• Fisheries (UK v Norway) (1951) — the British stayed out of Norwegian coastal waters for almost 300 years until 1906;
trouble began in 1911 with the seizure of a British trawler. The Court upheld Norway’s baselines, treating it as significant
that Norway had applied its method consistently over a very long period, that this was well known to the UK, and that
the UK had abstained from complaint.

INTERNATIONAL LAW - MODULE II | PAGE 23

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• Temple of Preah Vihear (1962) — in 1908 Thailand accepted as accurate a map placing the temple in Cambodian territory,
drawn under a 1904 Siam–French Indochina agreement. Its later conduct showed acceptance — e.g. a member of the
Thai Royal Family visited the temple while the French flag flew. Cambodia had relied, and for 50 years Thailand had
taken the benefit of a stable frontier; it was precluded from denying the boundary.
• Nuclear Tests (1974) — estoppel “may have played a role”. The Court relied on the French statements made to the world
at large, and on good faith: “Just as the very rule of pacta sunt servanda… is based on good faith, so also is the binding
character of an international obligation assumed by unilateral declaration.” Comparative point: detriment, essential on
Waltons Stores (Interstate) v Maher (1988) 164 CLR 387 principles, was not determining — though detriment could be
found in Australia and New Zealand not pressing the litigation.
• Gabčíkovo-Nagymaros (1997) — an express reference to estoppel by Vice-President Weeramantry: Hungary had allowed
Czechoslovakia (Slovakia inheriting the regime) to believe it would perform the 1977 Danube project, and so was
precluded from saying the treaty no longer bound it.
(b) Clean hands
• Diversion of Water from the River Meuse (1937) — “a rare example of judges applying equitable principles expressly as
‘general principles’”. The Netherlands complained that Belgium had built a lock in breach of an agreement that both take
water only at a certain point — but the Netherlands had built and operated a similar “unlawful” lock itself. Judge
Hudson: “where two parties have assumed an identical or a reciprocal obligation, one party which is engaged in a
continuing non-performance of that obligation should not be permitted to take advantage of a similar non-performance…
by the other party” — the maxims being “Equality is equity” and “He who seeks equity must do equity”, with “a very
similar principle… received into Roman Law”. Relief denied.
• Tehran Hostages (1980) — while the Court deliberated, the US launched a military rescue operation inside Iran and
imposed sanctions. The majority found for the US. Judge Morozov dissented: the US “has now deprived itself of any
right to refer to the treaty of 1955”. The majority said the legality of the rescue and sanctions was not before it.
• Nicaragua (1986) — Judge Schwebel would have dismissed the claim because of Nicaragua’s own support for rebels in El
Salvador (predating US support for the Contras) plus its misrepresentation of those facts to the Court. He cited River
Meuse and Tehran Hostages to declare clean hands a general principle of law.
• Arrest Warrant (2002) — Judge ad hoc Van den Wyngaert: the Congo did not come with clean hands, having failed to
investigate and prosecute Mr Yerodia itself.
(c) “Equity will not suffer a wrong to be without a remedy”
• Barcelona Traction (1970) — Belgium claimed for a Canadian company with over 80% Belgian shareholders; Canada,
the state of nationality, chose not to act. The Court accepted that “for reasons of equity, a State should be able, in certain
cases, to take up the protection of its nationals, shareholders in a company which has been the victim of a violation of
international law” — but denied Belgium standing, weighing the practical difficulties and the risk of “an atmosphere of
confusion and insecurity in international economic relations”. It is read as evidence that equitable considerations cannot
be brought in opposition to the law, since here they would have “opened the door to legal anarchy”.

5.6 The maritime delimitation exception — equity at its most liberal

• North Sea Continental Shelf (1969) — two notions reflecting opinio juris: delimitation must be by agreement, and
agreement must accord with equitable principles. So in delimitation, equity as a guiding principle had been accepted by
states as a rule of custom. Acceptance also rests on a broader basis: the decisions of a court of justice must be just, and
in that sense equitable. But equidistance “soon ceased to be anything more than one method amongst others”.
• Continental Shelf (Tunisia/Libya) (1982) — “the legal concept of equity is a general principle directly applicable as law”;
and: “The result of the application of equitable principles must be equitable… It is, however, the result which is
predominant; the principles are subordinate to the goal.”
• Consequence: each maritime boundary is unique, hence not susceptible to general rules — so decisions “do not
demonstrate any systematic definition of the equitable criteria” (Gulf of Maine (1984)).
• The critique she endorses (Charney): “Law is valuable only if it guides the behaviour of its subjects in the real world… If
the law is so flexible that any result is possible it fails to fulfil that essential function.”
• Judge Nelson (later President of ITLOS): the “infinite variety” of maritime situations prevented UNCLOS producing
definitive delimitation rules, investing tribunals with wide discretion and “creating a situation that is closely akin to a
grant of ex aequo et bono jurisdiction.”

5.7 Conclusion
• “There is deep unease amongst international scholars both from the common law tradition and the civil law tradition
at the unconfined discretion which would repose in judges were they permitted to have recourse to equity as an
unstructured concept.”

INTERNATIONAL LAW - MODULE II | PAGE 24

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• Nonetheless, in specific cases “equity has wide acceptance and is part of the general stock of legal norms of the
international order.”
• Back to River Meuse: Judge Anzilotti — the Italian jurist who had been rapporteur of the 1920 Advisory Committee —
though dissenting, agreed with Hudson, calling “one who seeks equity must do equity” “so just, so equitable, so
universally recognised, that it must be applied in international relations. [It is one of the] general principles of law
recognised by civilized nations.”
• Why the Court avoids abstraction: the 1920 debates showed profound differences about equity’s content; the number of
independent states has since more than doubled and all wish to share in the world’s economic wealth. The Court
achieves surprising unanimity in equity cases by “connecting elements of equity with very concrete circumstances”. “To
this extent equity is a general principle of law recognised by civilised nations.”
Contemporary: (i) Her worry that delimitation approaches an unlegislated ex aequo et bono has been partly answered by the three-stage
method (provisional equidistance → adjustment for relevant circumstances → disproportionality check) settled in Black Sea (Romania v
Ukraine) (2009) and Bangladesh/Myanmar (ITLOS 2012) and applied in Somalia v Kenya (2021): equity has been proceduralised, not
abandoned. (ii) Equity’s centre of gravity has moved from delimitation to distribution — the ICJ Climate AO (2025) puts common but
differentiated responsibilities and respective capabilities inside the customary due-diligence standard; the Loss and Damage Fund
(operationalised from COP28) is equity praeter legem institutionalised, though a fund without an agreed contribution key is precisely
Charney’s “any result is possible”. (iii) Clean hands remains where she left it: never codified by the ILC in state responsibility or
diplomatic protection, surfacing mainly in separate/dissenting opinions and investor–state illegality awards. (iv) Her line that “civilized
nations” “bears no real meaning today” is now institutionally endorsed — the ILC’s general-principles project dropped the phrase.
Exam handles: Equity entered by the back door — rejected as an independent source in 1920 because of its variable meaning, admitted through 38(1)(c)
as particular principles. Three registers: infra, praeter, contra legem (the last needs express power; Art. 38(2) has never been used). Estoppel: Serbian
Loans, Fisheries, Preah Vihear, Nuclear Tests, Gabčíkovo. Clean hands: River Meuse, then only dissents. Remedial equity limited by systemic
consequences: Barcelona Traction. Delimitation is the liberal outlier.

6. JENNINGS — “RECENT DEVELOPMENTS IN THE ILC: ITS RELATION TO THE

SOURCES OF INTERNATIONAL LAW” (1964)
(A talk given at Edinburgh, 14 December 1963. Jennings was then Whewell Professor at Cambridge; later President of the ICJ.)
Thesis: Stop looking for a statute-substitute. Article 13(1)(a) machinery — ILC → Sixth Committee → General Assembly →
diplomatic conference — is already genuinely law-making, and treaties are sources of obligation, not of law.

6.1 Framing
• He set out to comment on the ILC’s recent reports but decided to “work on a larger canvas”: how the Commission’s work
fits traditional notions of sources.
• Against the pessimists: critics called progress “slow and painful”; Jennings “cannot easily understand” this. Seen against
a century of the codification movement, “the acceleration… since the inception of the International Law Commission
can only be described as vivid” — witness the place of the 1958 Geneva Convention on the Continental Shelf in live
law of great economic and political importance. “This whole procedure that has developed under Article 13(a) of the
Charter now seriously rivals the International Court of Justice in its importance for international law.”

6.2 Codification and progressive development are inseparable

• The ILC Statute distinguishes them. But:
• Brierly (1931): “The legislative element in the attempt to codify any part of international law is not merely incidental and
subordinate; it outweighs the codifying element to such an extent that it becomes misleading to describe the process as
one of codification at all.”
• Lauterpacht: the area needing only “precise formulation and systematisation” “is insignificant”.
• The Commission itself (1956, law of the sea): the statutory distinction “can hardly be maintained”; provisions based on a
recognised principle were framed so as to fall into progressive development; it “had to abandon the attempt” to classify
articles, “as several do not wholly belong to either”.
• Rosenne: the most emphatic such statements came precisely on topics originally listed as suitable for codification.
• So progressive development predominates — “But it is here that the real problem arises. The term ‘progressive
development’ begs the question. Can these procedures be regarded as being in any degree law-making?”

6.3 The many stages — where is law made?

• After the Final Report goes to the Assembly (usually after a Sixth Committee debate), the Assembly may adopt it, or
“note” it while refusing adoption (as with arbitral procedures), or call a diplomatic conference. If a treaty emerges there
are further stages: adoption by the required majority, signature, ratification (before or after entry into force). At what
stage is there law-making?

INTERNATIONAL LAW - MODULE II | PAGE 25

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• The orthodox answer: law is made only where a law-making treaty gets its ratifications — and then only for the parties.
Short of that, a code has the authority of a restatement, i.e. some evidence of existing custom; a progressive-development
draft is no more than a proposal.
• Why that fails. Judge Jessup on the 1958 Geneva Conference asked the question “at the back of all our minds”: the
Conference’s action did not make international law — “or did it?” And “one knows as a matter of fact that… the law
relating to the continental shelf was not the same after 1958 as it was before 1958”, and not merely because 22
ratifications might one day arrive.

6.4 The municipal analogy and the “statute-substitute”

• “It is not always realised how much the treatment of the very sources of international law rests upon assumptions
which emanate from municipal law experience” — above all the assumption that what international law needs is
legislation, producing a hunt for statute-substitutes in treaties.
• “The use of the soubriquet ‘international legislation’ for multilateral treaties is an example of wish-fulfilment that has
been allowed to become almost scientifically respectable.” So is “the doubtful and awkward attempt to divide treaties
into law-making treaties and contract treaties.”
• His question: has this search “obscured from us not only that international law may develop law-making processes of a
quite different kind but even that this development is happening under our very eyes?”

6.5 Fitzmaurice’s axiom

• Fitzmaurice (Symbolae Verzijl, 1958): a treaty “is not really ever a formal source of law at all. It is not properly speaking a
source of law but a source of obligation. The law says that the obligation must be carried out, but the obligation is not,
in itself, law.”
• Jennings: “one of those axioms that may remain undiscovered for generations; but once it has been stated is so obviously
true as to compel acceptance.”
• The consequence, in Fitzmaurice’s words: the position would be no different even if every state in the world became
party to a soi-disant law-making treaty. “Admittedly there would be general consent, but it would be this consent that
would generate the law, and the treaty would merely be evidence of the consent… Treaties have in fact, in themselves,
only a pseudo-legislative character.”
• This explains a real drafting fight: Fitzmaurice’s (and earlier Lauterpacht’s) report said a code on the law of treaties
should not itself take treaty form but have “an independent basis”. When the Commission later reversed that decision it
observed the difficulty was not peculiar to treaties — codifying the law of treaties by treaty merely forces the point on
one’s attention.
• Worked illustration: ratifying the Geneva Convention on the High Seas does not create law; it creates “a particular,
essentially contractual, obligation”. That is why the ICJ applies it as one of those “international conventions, whether
general or particular, establishing rules expressly recognised by the contesting States” — which explains why Art.
38(1)(a) lumps particular and general conventions together, and means “we may need therefore to revise the commonly
accepted assumption that Article 38 is a statement of the formal sources” (the word “source” appearing nowhere in it).
• So the treaty is not a formal source in Alf Ross’s sense — “the channels through which validity pours into the world of
reality and lends to the law its ‘binding force’”. This is not to denigrate ratification, “merely to categorise accurately
instead of fancifully what governments actually understand by a treaty obligation.”

6.6 The gain

• “Once we get away from the search for the quasi-statute and so learn to reject the treaty as a formal source of law we
are by the same stroke free to recognise the treaty as a very cogent material source of that law founded in the practice
of States and in the general opinio juris generally called customary international law.”
• He qualifies the label: much of it “is clearly… something entirely different in important respects from custom”. His
image: “Maybe it is harmless to think of the law of the continental shelf as a sort of hot-house forced custom even if it
is rather quaint.” We must escape “the strait-jacket of traditional thinking about sources”.
• What becomes cogent as a material source: a draft treaty adopted by two-thirds of a diplomatic conference representing
most of the family of nations; a more or less unanimous UNGA vote adopting an ILC Report (or virtually rejecting it, as
with arbitration procedures); even the Commission’s own adoption of a Final Report.

6.7 Political infiltration — why this is real law-making

Documented by Rosenne: - ILC candidature depends on straight governmental nomination. - More and more members are or
have been governmental legal advisers. Though there on scientific merit, “they do represent also a confrontation on the scientific
plane of the varied interests of different States; a glance almost anywhere in the verbatim proceedings will demonstrate this.” -
The Sixth Committee — “a purely political body” — has become more and more prominent. - Conclusion: “Here then, out of

INTERNATIONAL LAW - MODULE II | PAGE 26

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
Article 13(a), is a genuine law-making or if you like, law-shaping machinery of actual as well as political importance, which has
already reached a stage of considerable sophistication.”

6.8 It can change law, not merely add to it

• The Norwegian Fisheries example. Britain argued that the mere failure of the 1930 Hague Codification Conference
cannot make new law. “This looks like a truism; but is it really true?… the negative fact of a consistent failure to agree
on a three-mile maximum breadth of territorial sea has been in many ways the principal strand in the weaving of a
new law of the territorial sea.”

6.9 The worked example: reservations, 1951–62 (narrate this as a sequence)

1. Orthodoxy: absent treaty provision, a reservation is valid only if accepted by all other parties.
2. 1951 — the Court. Reservations to the Genocide Convention: the old rule rejected — not in general terms, but for that
Convention. The Court said pointedly that the unanimity doctrine, though of “undisputed value”, had not been
“transformed into a rule of law”.
3. 1951 — the Commission goes the other way, reverting to the traditional doctrine as the general rule.
4. 1951–52 — the Assembly’s Sixth Session. The Secretary-General was to conform his Art. 102 practice on the Genocide
Convention to the Court’s opinion, and states were recommended to be guided by it; and for future conventions (after 12
January 1952) he should act as depositary of reservations without passing on their legal effect, merely communicating
them. “This resolution was, as it later appeared, the decisive turning point of the story.”
5. The nature of the split: the “collegiate” school versus the “flexibility” school — a confrontation between two views of
present law. Note the asymmetry: “it was the conservative school that was making suggestions for changes in what they
regarded as the law. The new flexibility school… assumed that they were stating what the present law actually was.”
6. 1962 (Waldock as rapporteur): two changes — the international community’s rapid expansion, making unanimity “less
practicable”; and the Secretary-General’s practice conforming to flexibility, with reserving states counted for entry into
force.
7. The 1962 Report accepts flexibility, except where the treaty provides otherwise, where parties are few, or where the
treaty is a constituent instrument (the organisation’s organ decides — as with the Indian reservation to the IMCO
Convention, resolved by a UNGA Resolution in 1959).
8. Result: (1) the traditional doctrine “no longer has any credibility as a rule of general international law”; (2) flexibility
“has come to occupy a position that nothing now is going substantially to dislodge”. “A striking example of the power of
Article 13(a) procedures assisting and accelerating a change of practice and the opinio juris.”

6.10 Not new — writers were just slow

• Nineteenth-century writers already knew treaties could be material sources of custom. But “for a subject where it is still a
commonplace to cite in the source books the diplomatic dispatches of particular governments engaged in arguing
particular claims, modern writers have been curiously slow to acknowledge the force, as law-making agencies, of
resolutions representing the collective opinion… of a vast majority” of the community.
• Higgins (1963), quoted: custom is deduced from state practice, which includes diplomatic actions and public
pronouncements; with international organisations, “the votes and views of States have come to have legal significance
as evidence of customary laws”; and “the practice of States comprises their collective acts as well as the total of their
individual acts… collective acts of States, repeated by and acquiesced in by sufficient numbers with sufficient
frequency, eventually attain the status of law.” The UN, especially since its trend to universality from 1955, is “a very
clear, very concentrated, focal point of State practice”.

6.11 New states

• “The significance of such a law-changing process for the needs and desires of the new States is obvious.”
• The 1962 Report on Treaties gives as a reason for preferring a convention to an expository code that it “would give all the
new States the opportunity to participate directly in the formulation of the law”, their participation being “extremely
desirable in order that the law of treaties may be placed on the widest and most secure of foundations”.

6.12 Two cautions

• Resolutions may mean nothing. “A resolution of an international body may represent the collective views of States; it
may also represent little or nothing of the sort. Often only time and the interaction of decisions indicating, when taken
together, an unmistakable trend will enable one to sort out the important from the insignificant.”
• The judicial function is weak. The process “will continue to be flaccid and uncertain just so long as the judicial function
remains… so very restricted in scope.” Ross: “The judicial decision is the pulse of legal life.”
• His heterodox qualification: “it is possible to exaggerate the importance of the judicial function in international law” —
(i) we have managed largely without it for the five or six hundred years of modern international law; (ii) in any municipal
INTERNATIONAL LAW - MODULE II | PAGE 27
MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
system much law, “particularly public law which is more closely analogous to the main body of international law than
private law”, is administered not by courts but by political or administrative bodies.
• But the lack of compulsory jurisdiction “is a serious weakness”. And note that the ILC’s two most impressive pieces of
law-making — baselines of the territorial sea and reservations — borrow their principal authority from a judgment
and an opinion of the Court (Anglo-Norwegian Fisheries; Reservations to the Genocide Convention). Rosenne says the
Commission has repeatedly found the absence of compulsory judicial determination a hindrance in drafting.

6.13 Choice of topics

• Two Secretariat surveys of the whole field: 1949, and pursuant to UNGA Res. 1505 (XV) of 12 December 1960.
• Members are part-time — which Jennings thinks “far the best arrangement; I shudder to think what sort of a body it
would become if it carried full-time salaried and pensioned employment for defunct practitioners and judges.”
• “Ripeness” abandoned. Per Cornford, “time… is like the medlar; it has a trick of going rotten before it is ripe.” “The
simple truth is that there are no topics of international law ripe for codification” — all need working up, and that
means law-making.
• Who chooses: initiating progressive-development projects lies with the General Assembly, not the Commission. “So an
important part of the topics… have been chosen or at least sanctioned not just by a legal body but also by a political
body, presumably with an eye to political realities of the moment.”
• Topics worked on: high seas; territorial sea; continental shelf; conservation of high-seas resources; statelessness and
nationality; diplomatic and consular intercourse and immunities; arbitral procedure; treaties; state responsibility; state
succession; special missions; rights and duties of States; formulation of the Nuremberg principles; right of asylum.
• Suggested but untackled: recognition of States and governments; jurisdictional immunities; jurisdiction over
extraterritorial crimes; treatment of aliens; historic waters.
• Defence of the “old-fashioned textbook” list: “it is none the worse for that”; the classical material must be worked first
into a system acceptable to the expanded community. “It is the content of the chapters rather than their rubrics that
needs rewriting.” On the complaint that textbooks ignore international investment and international economic law:
“how can you begin to do anything definitive in this field until you have got straight the basic principles of State
responsibility, that most traditional and yet most elusive, unsettled and intractable problem which lies at the very core of
any notion of an international law? No, the fact is that the old quarries are only just beginning to be opened up.”

6.14 The absentees (his agenda)

• Outer space — “what lawyers call outer space and what astronomers call inner space” — could be taken on soon.
• The line to quote in any climate question: “Another new problem of the greatest possible importance is the regulation
of experiments which are calculated or, judging by experience so far, miscalculated to change the earth’s environment;
this kind of irresponsibility on an Olympian scale the Commission is not likely to touch with a barge-pole for obvious
reasons, and in any case the problem is so urgent that it calls for straight political action rather than codification of
legal principles.”
• International organisations — including the responsibility of international organisations; Charter law; “co-existence”
(“doubtfully new, for co-existence has been a problem… at least since the crusades”).
• Old but unfinished: air law; war, neutrality and weapons (“of the greatest topical urgency… in the new context of UN
forces”); self-defence and aggression (which the Commission “will hardly wish to test itself on… for some time”);
international rivers (better suited to a code of general principles than a draft convention); acquisition of territory (“very
old-fashioned, but still exceedingly topical”); and the subjects of international law — organisations, public and private
corporations, individuals — “elusive, unsettled, very controversial, yet elementary in the proper sense of that word.”

6.15 Conclusion
• Much remains to do, including outside the Commission — “for writers still have a vital place”.
• “In the procedures which have developed under Article 13(a) of the Charter, we have to hand and actually working a
procedure which is not limited to drafting and proposing but is, within its limits, genuinely law-making; and if we
can bring ourselves to cease following the mirage of international statute law, we may find that there is developing
under our eyes the very machinery of international law-making that we have been looking for all these years.”
Contemporary: (i) He was right institutionally: the ILC now issues draft conclusions and guidelines rather than conventions —
Identification of CIL (2018), Jus Cogens (2022), General Principles (2025), Sea-Level Rise, Subsidiary Means (2025) — formally non-
binding yet cited by courts almost at once. His warning about the Sixth Committee’s political prominence is live in 2026: the Committee
pressed the ILC to understate the role of precedent, and critics reply that the cautious wording misses the rebuttable presumption of
consistency international courts actually operate (Cameroon v Nigeria; US – Stainless Steel; Saipem v Bangladesh). (ii) His “failure to agree can
itself weave new law” has an exact analogue in the collapse of the plastics negotiations (INC-5.2, Geneva, August 2025), where deadlock
over production caps is now shaping unilateral regulation and practice. (iii) His 1963 aside on environment-altering experiments — which
he thought needed “straight political action rather than codification” — was answered in the way he did not expect: through the ITLOS
INTERNATIONAL LAW - MODULE II | PAGE 28
MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
(2024) and ICJ (2025) climate advisory opinions, i.e. by the judicial function whose weakness he lamented, and through advisory rather
than contentious jurisdiction. (iv) His reservations story ended in VCLT Arts. 19–23; the modern frontier — reservations to human-rights
treaties and treaty bodies’ power to sever an invalid reservation (ILC Guide to Practice, 2011) — is exactly the “constituent instrument”
exception he flagged via IMCO.
Exam handles: Codification and progressive development are inseparable; nothing is ever “ripe”. Fitzmaurice: a treaty is a source of obligation, not law
— consent generates the law, the treaty evidences the consent. The gain: conference votes, ILC reports and UNGA resolutions become cogent
MATERIAL sources of custom. Article 13(1)(a) machinery is genuinely law-shaping and can abolish old rules (the reservations story). Stop chasing the
statute-substitute.

7. AUST — MODERN TREATY LAW AND PRACTICE, CHS 1–2

PART A — THE VCLT 1969
Epigraph (Reuter): “The Convention clearly marked the beginning of a new era in the law of treaties.”

7.1 Two terms not to confuse

• The law of treaties = “the body of rules which determines whether an instrument is a treaty, how it is made, brought into
force, amended, terminated and operates generally.”
• Treaty law = the substance of a treaty (the rights and obligations it creates) — “a matter for the negotiating states.”
• The Convention is rightly called “the treaty on treaties”: it does not occupy the whole ground but covers the most
important areas.

7.2 Genesis
• The UNGA established the ILC in 1947; the law of treaties was chosen at the Commission’s first session in 1949.
• Four successive British Special Rapporteurs: Brierly, Lauterpacht, Fitzmaurice, Waldock — tasked with stating the
already well-developed customary law.
• Final draft articles 1966 → Vienna Conference 1968–69 → adopted 22 May 1969 → in force 27 January 1980.
• By end-2006: only 108 parties of about 193 states.

7.3 Why it has never needed amending

• It codified and to some degree developed the law. “Like the United Nations Charter, the intelligence of the Convention’s
drafting has enabled states to continue or modify their practice without distorting or departing from the rules.”
• Many provisions expressly envisage departure. Example: Art. 7(1) requires full powers, then excepts the common
practice of dispensing with them. (He lists dozens of similarly residual articles.)
• So the rules are “largely residual, leaving treaty practice very much in the hands of states — as it should be”. For
practitioners it is their “bible” and “does not need mending, or rather, amending”. But “it is not always wise to depart
from established practice unless it is really necessary.”

7.4 Scope — the limits

• Applies only to treaties between states (Art. 1).
• With/between organisations → the 1986 Vienna Convention, applying the 1969 rules suitably adapted. Where states
party to the 1969 Convention are also parties to such a treaty, as between them the 1969 Convention applies, not custom
(Art. 3(c)) — though the distinction is “now pretty academic”.
• Constituent treaties and treaties adopted within an organisation are concluded by states, so the 1969 Convention applies
— without prejudice to the rules of the organisation (Art. 5) (e.g. reservations to ILO Conventions).
• Oral agreements are excluded by Art. 2(1)(a) (“in written form”) for clarity, but their legal force is unaffected, and
customary rules still apply (Art. 3(a)). They are “extremely rare”. His example: the Denmark–Finland dispute over
Denmark’s bridge across the Store Bælt (Great Belt) was settled in 1992 in a telephone conversation between the two
Prime Ministers — Finland discontinuing its ICJ case in return for a payment. There is no joint record of the
agreement. (Cf. Eastern Greenland (PCIJ 1933). US law requires oral agreements to be reduced to writing.)
• No retrospective effect (Art. 4). Worked example: the Convention entered into force 27 January 1980; UNCLOS was
concluded 10 December 1982; so for states party to the VCLT on that date, its rules apply between them as regards
UNCLOS. But provisions reflecting custom apply (as custom) to earlier treaties anyway.
• Art. 73 confirms the Convention does not prejudge state succession, state responsibility for breach, or the outbreak of
hostilities — largely governed by custom.
• Bilateral/multilateral: bilateral = two states, multilateral = three or more — but there are bilateral treaties in which two or
more states form one party. Art. 60(1) is the only provision limited to bilateral treaties; Arts. 40, 41, 58 and 60 expressly
address multilateral treaties; reservations and depositaries concern multilateral treaties only.

INTERNATIONAL LAW - MODULE II | PAGE 29

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
7.5 The Convention and custom
• Custom’s two elements per Aust: (1) substantial uniformity of practice by a substantial number of states; (2) opinio
juris — “a general recognition by states that the practice is settled enough to amount to a binding obligation”.
• A new treaty rule followed by non-parties can generate custom binding non-parties inter se and as between them and
the parties — and this can happen even before entry into force (North Sea, paras. 70–85).
• UNCLOS illustration: many provisions went beyond codification, but negotiations proceeded by consensus (even
though the final text was voted on), so during the twelve years before entry into force in 1994 most provisions came to be
accepted as custom — important, since even by end-2006 UNCLOS had only 152 parties.
• “An accumulation of bilateral treaties on the same subject, such as investment, may in certain circumstances be
evidence of a customary rule.” (Contrast Kelly and Panezi, who reach opposite conclusions on exactly this.)

7.6 How much of the VCLT is custom? — the practitioner’s answer

• The question “is not of great concern to the lawyer in his day-to-day work”. In negotiations, “the rules set forth in the
Convention are invariably relied upon, even when the states are not parties to it”. Aust recalls at least three bilateral
negotiations where the other side relied heavily on specific articles without having ratified, “the justification for
invoking the Convention… rarely made clear”.
• It matters only if litigated — “and even then the court or tribunal will take the Convention as its starting — and normally
also its finishing — point”.
• Case law:
• Kasikili/Sedudu Island (Botswana/Namibia) (1999) — applied Arts. 31 and 32 to an 1890 Germany–UK treaty, though the
Convention is not retrospective.
• Namibia (1971) — the rules on termination for breach “may in many respects be considered as codification of existing
customary law”; Art. 60 applied.
• Fisheries Jurisdiction (UK v Iceland) (Jurisdiction) (1973) — Art. 52 (coercion) and Art. 62 (fundamental change) reflect
custom.
• Gabčíkovo (1997) — the treaty pre-dated entry into force, but the Court “brushed aside” non-applicability and applied
Arts. 60–62 as custom, “even though previously they had been considered somewhat controversial”.
• Conclusion: “There has as yet been no case where the Court has found that the Convention does not reflect customary
law.” Unsurprising, since “any codification of the law will inevitably reduce the scope for judicial law making”. “To
attempt to determine whether a particular substantive provision… represents customary international law is now a
rather futile task.” Even procedural rules like Art. 56(2) (minimum notice of termination if the treaty is silent) may now
be custom. As Sir Arthur Watts put it, the modern law of treaties is now authoritatively set out in the Convention. “As
with any court, if the judges consider that the rules in a treaty (or espoused in a learned work) are wise, they will do
their best to find a way of treating them as the law.”

7.7 Can new custom override an old treaty rule?

• Since 1945 law-making treaties have multiplied, so many topics are governed by both custom and treaty.
• Because international law has no hierarchy of sources, custom and treaty being autonomous, it is argued that codified
custom retains a separate existence — “a controversial theory” which “does not reflect the approach… taken by foreign
ministry legal advisers (and dare one say also judges?)”, who “naturally give more weight to a treaty rule than to a
different customary rule”. (The leading authority for separate existence is Nicaragua (Merits), paras. 172–82.)
• But new custom can modify a treaty rule’s operation. Fisheries Jurisdiction (Merits) (1974): since the adoption of the
1958 High Seas Convention, the right to establish twelve-mile fishing zones had crystallised as custom,
notwithstanding that Convention’s freedom-of-fishing provisions.
• “International law does not contain any principle of acte contraire, by which a rule can be altered only by a rule of the
same legal nature.”
• The lost draft article: ILC 1964 draft Art. 68(c) provided that a treaty’s operation may be modified by “the subsequent
emergence of a new rule of customary international law”. It was dropped only because the Commission did not think its
mandate extended to the general relationship between custom and treaty.

7.8 Where to look

• Best source: the ILC’s Commentary on its final draft articles (YBILC 1966, II); comparative table at ILM (1969) 714;
drafting history in the ILC Yearbooks from 1950; the Conference summary records (since Vienna changed the drafts);
Rosenne’s guide to the travaux with Wetzel’s English collection; accounts by Sinclair and Kearney and Dalton, all
participants.

PART B — WHAT IS A TREATY?

Epigraph (T. S. Eliot, “East Coker”): “the intolerable wrestle with words and meanings.”
INTERNATIONAL LAW - MODULE II | PAGE 30
MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
7.9 The definition
Art. 2(1)(a): “an international agreement concluded between States in written form and governed by international law, whether embodied
in a single instrument or in two or more related instruments and whatever its particular designation.”
• Although expressed for the Convention’s purposes, its elements now represent custom.
• “The difficulty is not the definition itself, but whether a particular instrument or transaction falls within it.”
• Practitioner’s warning: flexibility is the law of treaties’ strength, but most treaties follow standard forms and long-
established procedures, so “one should therefore think very carefully before departing from them”. Trouble arises from
“an honest, but misguided, wish to speed up the process”.

7.10 Element by element

“an international agreement”
• Must have an international character. “Treaty” is a generic term; the Convention does not recognise domestic categories
like inter-state / inter-governmental / inter-ministerial / administrative.
• “Universal” and “regional” have no legal significance in themselves; “international treaty” is tautological and best
avoided.
• Plurilateral = between a limited number of states with a particular interest. Constituent = establishing an organisation.
• Simplified form: “There is no difference in legal effect between a treaty contained in a formal, single instrument and
one constituted by the more informal method of an exchange of notes.” In 1945 there was still doubt, reflected in UN
Charter Art. 102 (“every treaty and every international agreement”); by the 1960s there was none. Because an MOU is
not intended to create international obligations, “it is also a mistake to think it is a treaty in simplified form.”
“concluded between States”
• May be with or between other subjects (especially organisations). But an agreement between companies, or between a
state and a company, is not a treaty. Anglo-Iranian Oil Co. (Preliminary Objections) (ICJ 1952): an oil concession
granted by a state to a foreign company was not a treaty, because the company’s state of nationality was not a party. Even
a clause applying international law “does not make it a treaty”.
• Indigenous peoples. Nineteenth-century agreements with indigenous representatives — the Treaty of Waitangi 1840
(Maori chiefs ceding New Zealand to the Crown), the 1815 US–Sioux treaty, the 1884 Treaty of Protection with the Kings
and Chiefs of Old Calabar — were drawn in treaty form and so described, “but since the land occupied by such peoples
was, at the time, not considered to be a state, such agreements were not treaties, even if they had, and continue to have,
effects in domestic law.” In Cameroon v Nigeria (2002, paras. 200–9) the Court held that, as with some 350 other
“treaties” with chiefs in the Niger Delta, it was not a treaty.
• A treaty need not be expressed to be between states as such — a state is a legal concept, so heads of state, governments,
ministries or agencies act for it.
“in written form”
• No reason a treaty cannot be in a telegram, telex, fax or e-mail, or constituted by an exchange of such communications,
provided it can be reduced to permanent, readable form.
• Absence of signed originals is fine if the “signature” can be authenticated. In September 1998 President Clinton and
Taoiseach Ahern issued a Communiqué on Electronic Commerce electronically, each at a separate terminal using
electronic signatures (uniquely linked to the signatory, identifying him, under his sole control, and linked to the data so
alteration would show). The Communiqué was not a treaty — and the Financial Times, “uncharacteristically for that
excellent newspaper, misdescribed the document as a treaty” — but Aust thinks it “not too fanciful” to envisage full
powers, instruments of ratification or treaties being signed and deposited electronically. His caution: “Given the
numerous mistakes in treaties and treaty procedures made even today, there is no reason to suppose that the latest
information technology will necessarily improve matters.”
“governed by international law” — the decisive element
• Per the ILC Commentary, this embraces the intention to create obligations under international law. No such intention,
no treaty.
• Aegean Sea Continental Shelf (1978) — the Court examined the terms of a Greek–Turkish Prime Ministers’ joint
communiqué and the circumstances of its drafting, and found no intention to submit to the Court’s jurisdiction.
• The evidential rule: “Intention must therefore be gathered from the terms of the instrument itself and the
circumstances of its conclusion, not from what the parties say afterwards was their intention” (Qatar v Bahrain (1994)).
• It also separates treaties from agreements governed by domestic law. Though no form or wording is required (Temple of
Preah Vihear (Prelim. Objections, 1961)), foreign-ministry lawyers deliberately use terminology showing that the
participants “merely wish to record their mutual understandings as to how they will conduct themselves”.

INTERNATIONAL LAW - MODULE II | PAGE 31

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• MOUs. Their existence and importance “has until recent years not been well known outside government circles… Most
are never published.” Published examples: the MOU on Port State Control in the Caribbean Region 1996; and the UK–
Jordan MOU on deportations (2005), with similar ones with Libya (18 October 2005) and Lebanon (23 December 2005),
“rightly described as no more than ‘diplomatic assurances’”.
• Synonyms: political agreements, gentlemen’s agreements, non-legally binding agreements, de facto agreements.
Diplomats say “MOU”. But the name does not determine status — some treaties are also called MOUs. By defining a
treaty as governed by international law, the ILC excluded MOUs from the draft articles.
• “The fact that some treaties, bilateral or multilateral, have little substance, ‘hard’ obligations or enforcement
mechanism, does not affect their treaty status” (example: the ICESCR 1966).
“a single instrument or two or more related instruments”
• Covers exchanges of notes (an initiating note and a reply), which “play an increasingly important role”.
• Qatar v Bahrain (1994) — a double exchange of letters: Saudi Arabia sent identical letters to Qatar and Bahrain
proposing settlement procedures; each accepted; Saudi Arabia announced they had agreed to go to the ICJ. “This
complicated scheme was necessary because of political sensitivities”, but the texts were agreed in advance, so although
three states were involved there were only two parties.
• The Algiers Accords 1981 (creating the Iran–US Claims Tribunal) comprised (1) two Algerian Declarations — General,
and on Settlement of Claims — plus the two states’ Undertakings, together a treaty; and (2) an Escrow Agreement with
the Federal Reserve Bank of New York, Bank Markazi and the Central Bank of Algeria.
• Parallel exchanges (A–B; A–C; A–D) can make a part-bilateral, part-multilateral treaty. Make clear who the parties are:
four states could be four parties, or two (A and B+C+D). Where there are two, clarify whether one member of the
collective party can terminate by withdrawing.
• An exchange of notes on interpretation may be subsidiary and simultaneous; it need not itself be a treaty, though
sometimes it is.
“whatever its particular designation”
• Naming is “unsystematic”, now “even more confusing, inconsistent and changeable than in the past”, driven by
organisational practice or political preference.
• The rule: “what is decisive is whether the negotiating states intend the instrument to be (or not to be) binding in
international law. Thus, just as one should never judge a book by its cover, one should never assume that the name given
to an international instrument automatically indicates its status.”
• Treaty/agreement/convention are safe assumptions — but check the text. “Most other names are problematic.” The UN
Charter 1945 and the CIS Charter 1993 are treaties; the OSCE Charter of Paris 1990, the Russia–US Charter of
Partnership and Friendship 1992 and the OSCE Charter for European Security 1999 are MOUs.

7.11 A treaty need not be signed

• The definition never mentions signature, and Arts. 12–13 show it is unnecessary. A treaty may be constituted by third-
person diplomatic notes, which by long practice are initialled, not signed, and are expressed to be between a ministry
and an embassy, not individuals.
• Sometimes unsigned is politically preferable. The Decision of the Heads of State and Government at the Edinburgh
European Council, 12 December 1992, on Denmark’s Maastricht problems, is not in customary treaty form but is treated
as a treaty and registered as such. “Some of the less confident leaders were reluctant to be seen signing it. They were
skilfully advised that signature was not necessary.”
• Routine examples: FAO Conference treaties are not signed but subject to acceptance (equivalent to ratification); the
Refugees Protocol 1967 could only be acceded or succeeded to, though the Refugees Convention 1951 could be signed;
and the treaty establishing the CTBT Preparatory Commission (1996) was effected by a resolution of the signatory
states.

7.12 The named forms

• Memorandum of Understanding. “Only by studying carefully all the terms… can one determine its status.” History:
began “in a small way after the Second World War, three being concluded in the 1950s in connection with the Treaty of
Peace with Italy”, probably for a less formal appearance. Modern treaties called MOUs: the 1992 MOU on Nuclear-
Related Dual-Use Equipment; the two 1991 MOUs on Avoidance of Overlaps… in Deep Sea-Bed Areas; the 1995 MOU
on the Inter-Organisation Programme for the Sound Management of Chemicals; the 1999 Memorandum on Marine
Turtles off the Atlantic Coast of Africa. “An even more misleading name for a treaty is the Provisional Understanding
regarding Deep Seabed Matters 1984”, while the Paris MOU on Port State Control 1982 is unquestionably an MOU. The
US and EU conclude many bilateral treaties called MOUs, as does the UN (e.g. the ill-fated Iraq–UN MOU 1998 on
weapons inspections). “But the UN Treaty Handbook is quite wrong in appearing to regard MOUs… as treaties” — the
UN Treaty Section analyses every instrument submitted for registration, whatever its name.
INTERNATIONAL LAW - MODULE II | PAGE 32
MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
• The abuse point: “Occasionally, an international organisation or a state will be asked by a state (and may even agree) to
call a treaty a Memorandum of Understanding so that the requesting state can try to avoid a constitutional requirement
that it put treaties to its legislature for prior approval.”
• Exchange of notes. Same ambiguity. Drafting signals: a treaty says the exchange “shall constitute an agreement
between our two Governments”; an MOU says it “records the understanding of our two Governments”. Usually minor
or supplementary — “but a few have been substantial”: the UK–US Lend-Lease Agreements 1940–1, “under which the
former was lent sixty badly needed destroyers in return for leases of bases”, were exchanges of notes.
• Protocol. Nowadays generally a supplementary or amending treaty, but also used for documents annexed to and
integral to a treaty (CFE Treaty 1990, Art. I(3)). “Use (and read) with care.”
• Optional Protocol. Closely associated with, usually adopted with, a treaty; “Optional” means a party to the main treaty
need not join. Most commonly on nationality or dispute settlement (both attached to the VCDR 1961).
• Unusual names: Compact, Solemn Declaration, Administrative Agreement, Protocol of Decisions, Platform, Concordat,
Agreed Minute, Terms of Reference (the International Nickel Study Group Terms of Reference 1986, then tin and
copper), Declaration (the Maroua Declaration (Cameroon–Nigeria) 1975, in Cameroon v Nigeria).
• The best story: a 1992 Lithuania–Russia treaty on the withdrawal of Russian forces was called “Timetable”, its
substance being only detailed tables of military assets. It had been an annex to a treaty on the same subject; there was a
political problem over signing the treaty, so the Timetable alone was signed and later registered as a treaty. A
supplementary treaty to the CTBT 1996 is called simply “Text”.
• Inscrutable titles. The China–UK Joint Declaration on the Future of Hong Kong 1984 “is clearly a treaty, and was
registered as such… The title was chosen for political reasons.” Conversely, since 1989 Argentina and the UK have
issued “Joint Declarations” and “Joint Statements” on the Falkland Islands (fishing, oil exploration), sometimes with
substantial provisions and occasional treaty language (“agree”, “shall”), but not in normal treaty form — and they
provide for circulation as General Assembly documents, implying no intention to register, and none have been
registered. “While Argentina continues to maintain its claim… it may be reluctant to enter into treaties with the United
Kingdom about the territory, even though it did so before the Argentine invasion in 1982.”
• In the distant past a Final Act occasionally constituted a treaty; no longer.
• Colloquial names. The Chicago Convention 1944 = the Convention on International Civil Aviation. The Dayton
Agreement 1995 was adopted at Dayton but signed in Paris; the Landmines Convention 1997 is also the Ottawa
Convention, conceived and opened for signature there but adopted in Oslo; the Locarno Pact 1925 was negotiated at
Locarno but signed “in a posh room in the Foreign Office in London, which has been known ever since as the Locarno
Room”. “And the media has much to answer for. A treaty is now very rarely called a ‘Pact’, but sub-editors regularly
use this four-letter word as convenient shorthand… The ‘Locarno Pact’ was a series of treaties. The famous ‘Kellogg–
Briand Pact’ 1928 was in fact the General Treaty for the Renunciation of War.” Avoid “the Hague/Geneva/Vienna
Convention” unless clear from context.

7.13 Treaties governed by, or referring to, domestic law

• States can contract under domestic law — e.g. bulk commodity purchases, possibly under a third state’s law.
• A state leasing land for an embassy will usually use a domestic-law lease, even if provided for in a treaty. (“Contrary to
popular belief, the land on which a state’s embassy stands is not territory of that state.”)
• Loan treaties may provide that the lender state’s law governs the contractual arrangements.
• None of this alters treaty status. Grants under the US Foreign Assistance Act 1961 are made under agreements governed
by US federal law, not treaties.

7.14 Two Latin terms

• Pactum de contrahendo: “Latin phrases should always be approached with caution: they can conceal more than they
reveal.” Variously used for an agreement to conclude a treaty; to include certain clauses in future agreements; or to
become party to an existing treaty — the celebrated undertaking by Poland in the Treaty of Versailles to accede to the
Berne Convention on Railway Transport. It should not describe an agreement to negotiate (a pactum de negotiando),
“even though such an agreement must be carried out in good faith” (Railway Traffic between Lithuania and Poland, PCIJ
1931). Example: NPT Art. VI — each party “undertakes to pursue negotiations in good faith on effective measures
relating to cessation of the nuclear arms race… and to nuclear disarmament”.
• Modus vivendi: used for a treaty intended to be temporary, and for MOUs, particularly on fisheries.
Contemporary: (i) His claim that VCLT participation hardly matters has held — around 116 parties, no tribunal has found a substantive
provision non-declaratory, and the ICJ applied Arts. 31–33 as custom again in the Advisory Opinion on the Right to Strike under ILO
Convention No. 87 (21 May 2026), where whether an unstated right was protected turned entirely on ordinary meaning, context and
object and purpose. (ii) The MOU/treaty line has become politically decisive exactly as he foresaw: the UK–Rwanda migration
arrangements were converted from an MOU into a treaty (December 2023) because the UK Supreme Court had treated the earlier
instrument’s non-binding character as bearing on risk — his own UK–Jordan, Libya and Lebanon deportation MOUs are that litigation’s
INTERNATIONAL LAW - MODULE II | PAGE 33
MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
direct ancestors. (iii) His point that states request the MOU label to avoid parliamentary approval is now a mainstream accountability
critique, applied to executive agreements and political commitments in trade, migration and security (including AUKUS arrangements
and post-Brexit UK–EU instruments); progressive commentary treats it as a hollowing-out of legislative scrutiny. (iv) His suggestion that
accumulated investment treaties may evidence custom is being unwound by the ECT withdrawals and UNCITRAL WG III — a network
can contract as fast as it grew, weakening any inference from numerosity. (v) The BBNJ Agreement (2023) — sixtieth ratification
September 2025, in force January 2026 — is a textbook re-run of his UNCLOS point: a consensus-negotiated text shaping practice well
before universal participation.
Exam handles: Treaty = (a) international agreement, (b) between states, (c) written, (d) governed by international law, (e) one or more instruments, (f)
whatever its designation. The decisive element is (d) — intention, from the terms and the circumstances of conclusion, not later assertions (Aegean Sea;
Qatar v Bahrain). The name is never decisive (Hong Kong Joint Declaration = treaty; OSCE Charter of Paris = MOU). The VCLT is residual, non-
retrospective (Art. 4), inter-state only (Art. 1), silent on succession/responsibility/hostilities (Art. 73) — but applied as custom anyway.

PART III — SYNTHESIS

S.1 The seven authors on one page
Custom is… Treaty is… Cure for indeterminacy
Akehurst real and workable — practice + practice; a material source of custom better doctrine
public statements where opinio juris exists
Kelly a fiction; “non-empirical” paper the only legitimate law-making form abolish custom; use treaties and defined
norms deduced by elites (ritual + consent) supermajority processes
Kennedy the soft pole, corrupted by the hard pole, corrupted by good- none — indeterminacy is the point
consent-talk faith talk
Jennings “hot-house forced custom” a source of obligation, not law; a recognise the machinery we already
accelerated by Art. 13(1)(a) material source of custom have
Panezi one of three equal, autonomous, one of three general principles as the reservoir,
independent sources esp. economic law
White how equity entered maritime the vehicle for express ex aequo et tie equity to concrete circumstances
delimitation bono (never used)
Aust fills the VCLT’s gaps and can even the practitioner’s whole world the question is largely futile in practice
modify treaty rules

S.2 The four fault-lines (use these to structure any answer)

1. Consent v. community — can a state be bound without agreeing? Kennedy: unanswerable. Kelly: must be consent.
Akehurst: the persistent-objector compromise. Jennings: relocate law-making to collective political acts.
2. Induction v. deduction — Kelly’s charge is that CIL is deduced and dressed as induced. Akehurst is the sophisticated
inductivist. Panezi and White show 38(1)(c) is frankly deductive and unembarrassed — which is why general principles
are the honest home for reasoning custom performs dishonestly.
3. Who participates — acts-only empowers the strong; declarations-count empowers the many; treaty-making empowers
negotiators; general principles empower judges and experts. Identifying whose power a doctrine entrenches is the
Module’s real work.
4. The status of non-binding text — UNGA resolutions, ILC drafts, declarations, MOUs, “diplomatic assurances” all sit in
one zone. Akehurst: practice if declaratory in terms. Jennings: material sources. Kelly: aspirational, excluded. Aust:
distinguish by intention. Kennedy: this is why the argument never ends.

S.3 Master case table

Case Proposition Reading
Lotus (1927) Presumption of liberty (own territory, concurrent jurisdiction); omissions Akehurst; Kennedy
need opinio juris; absence of protest significant
The Wimbledon (1923) Treaties affecting third parties (Suez/Panama) evidence permissive custom Akehurst
Åland Islands (1920) Self-determination in “a certain number of treaties” is not yet a positive rule Akehurst
Serbian Loans (1929) Estoppel needs clear representation + change of position White
Free Zones (1932) Rebus sic stantibus; third-party stipulations not lightly presumed Kennedy
Eastern Greenland (1933) Little sovereign activity suffices against a weaker claim; oral agreements Akehurst; Aust
River Meuse (1937) Clean hands / “he who seeks equity must do equity” as a general principle White

INTERNATIONAL LAW - MODULE II | PAGE 34

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
Case Proposition Reading
Trail Smelter (1938/41) Deduction by analogy from US domestic law with no relevant state practice Kelly
Chorzów Factory Reparation as a general principle Panezi
Nuremberg IMT (1946) Resolutions and unratified treaties evidence custom; Hague Regulations Akehurst
declaratory by 1939
Reservations to Genocide Unanimity not “transformed into a rule of law”; object and purpose test Jennings; Kennedy
Convention (1951)
Anglo-Norwegian Fisheries Persistent objector; acquiescence; straight baselines; economic interests + Akehurst; Kennedy;
(1951) long usage White; Kelly
Asylum (1950) Regional custom must be proved; major inconsistency fatal; Montevideo Akehurst; Kelly
preamble non-declaratory
Haya de la Torre (1951) Law distinguished from courtesy and good neighbourliness Akehurst
Anglo-Iranian Oil (1952) A state–company concession is not a treaty Aust
Rights of US Nationals in Custom in correspondence and conference records; ex gratia payments Akehurst
Morocco (1952)
Nottebohm (1955) National laws as practice (genuine link); omissions as practice Akehurst
Right of Passage (1960) Bilateral custom; 125 years; lex specialis over general custom Akehurst
Temple of Preah Vihear No prescribed form if intention clear; 50 years’ acquiescence; plea of error White; Kennedy;
(1961/62) Panezi
South West Africa (1966) Moral principles count only if given “sufficient expression in legal form” Akehurst
North Sea (1969) Two-element test; “extensive and virtually uniform”; equidistance not all
inherent; natural prolongation; equitable principles; reservations
Barcelona Traction (1970) Shareholder protection and equity’s limits; delegates’ positions as practice White; Akehurst;
Panezi
Namibia (1971) VCLT Art. 60 as codified custom Aust
Fisheries Jurisdiction Arts. 52, 62 as custom; new custom modifies a prior treaty rule; protests Aust; Akehurst
(1973/74)
Nuclear Tests (1974) Unilateral declarations bind: intention + good faith; objective interpretation Kennedy; White;
Akehurst
Aegean Sea (1978) Joint communiqué not an agreement — intention decisive; interpret law as Aust; Panezi
it is today
Tehran Hostages (1980) Clean hands in dissent (Morozov) White; Panezi
Tunisia/Libya (1982) Equity is “a general principle directly applicable as law”; the result White
predominates
Gulf of Maine (1984) No systematic definition of equitable criteria White
Libya v Malta (1985) Custom deduced by analogy from the EEZ regime Kelly
Frontier Dispute (1986) Equity ≠ arbitrary fairness; cannot modify a colonial frontier; equity ≠ White
equality
Nicaragua (1986) Two elements restated; custom retains separate existence; clean hands Kelly; Aust; White
(Schwebel, diss.)
Qatar v Bahrain (1994) Double exchange of letters as a treaty; intention from terms and Aust
circumstances
Gabčíkovo (1997) VCLT Arts. 60–62 as custom; estoppel (Weeramantry) Aust; White
Kasikili/Sedudu (1999) Arts. 31–32 applied to an 1890 treaty Aust
Arrest Warrant (2002) Clean hands in dissent (Van den Wyngaert) White
Cameroon v Nigeria (2002) Niger Delta chiefs’ agreements are not treaties; the Maroua Declaration is Aust
West Rand Central Gold Treaty stipulations are weak evidence of “international common law” Akehurst
Mining [1905]
Paquete Habana (1900) Comparative national law; inconsistency before/after 1815 Akehurst
Sabbatino (1964) Judicial restraint in declaring CIL Kelly

INTERNATIONAL LAW - MODULE II | PAGE 35

MIDSEMESTER SPEEDRUN HRUTHIKA ADDLAGATTA
Case Proposition Reading
Filartiga (1980) Torture as customary human rights law Kelly

S.4 Contemporary dossier (to August 2026)

• Climate. ITLOS AO (21 May 2024) — GHG emissions are marine pollution under UNCLOS. ICJ AO, Obligations of
States in respect of Climate Change (23 July 2025) — unanimous; customary duty to prevent significant harm to the
climate system and to cooperate; climate treaties are not lex specialis excluding custom; breach = internationally wrongful
act (cessation, guarantees of non-repetition, full reparation); new fossil-fuel exploration licences carry the greatest
exposure; CBDR-RC informs due diligence. UNGA res. A/80/L.65 (20 May 2026) — tabled by Vanuatu, 90 co-sponsors,
adopted 141–8–28.
• ILC as quasi-legislator. Identification of CIL (2018); Jus Cogens (2022); General Principles (2025, dropping “civilized
nations”); Subsidiary Means (2025) + the Sixth Committee row over precedent; Sea-Level Rise.
• Treaty-making — wins and failures. WHO Pandemic Agreement (May 2025); BBNJ (in force January 2026); the collapse
of the plastics negotiations, INC-5.2, Geneva, August 2025; the Ozone/Montreal–Kigali model Kelly relies on.
• Economic law. ECT withdrawals (2023–25); UNCITRAL WG III on ISDS reform; WTO Appellate Body paralysis and the
MPIA; sovereign-debt architecture debates after Sevilla (2025).
• Adjudication and accountability. South Africa v Israel provisional measures (26 January 2024 onwards); ICC warrants
(Putin 2023; Netanyahu and Gallant 2024) and US sanctions on ICC personnel (2025); ICJ Right to Strike AO (21 May
2026); Chagos AO (2019) and the UK–Mauritius agreement (22 May 2025).
• Use of force. Russia–Ukraine; the January 2026 US strikes in Venezuela and the Art. 2(4) debate — Kelly’s warning that
CIL lets powerful states “conjure up exceptions to fundamental norms”.

S.5 Four exam skeletons

“CIL is a fiction maintained by those who benefit from it.” Open with Kelly’s authority/legitimacy frame and empirical v. non-
empirical CIL. Concede the best points: deduction dressed as induction (Trail Smelter, Libya v Malta), unavailable practice, the
objector rule as counter-reformation. Answer with Akehurst: practice properly defined; opinio juris as public statement; change by
declaration is preferable to change by breach. Add Jennings: the “fiction” is Art. 13(1)(a) machinery doing real work. Use Kennedy
to explain why neither side can win. Illustrate with the ICJ Climate AO and the distributive twist (the Global South using the
customary form). Conclude on participation.
“Is Article 38 a statement of the sources of international law?” Kennedy (the word “source” never appears; addressed to judges)
→ Jennings (Fitzmaurice’s axiom explains why 38(1)(a) lumps general and particular conventions together) → Panezi (three equal,
autonomous, independent sources; hierarchy shown only by jus cogens) → White (equity proposed and rejected in 1920, entering
via 38(1)(c)) → Akehurst on the “notoriously defective” drafting of 38(1)(b). Conclude: the real modern statement of sources is now
the ILC’s draft conclusions, not Art. 38.
“What role should UNGA resolutions play?” Jennings (collective acts as a concentrated focal point — Higgins; the reservations
story) → Akehurst (only if declaratory in terms; voting figures, separate votes, abstentions; the NIEO preamble as counter-example)
→ Kelly (bootstrapping; Eleanor Roosevelt on the UDHR; the Harper letter; Tunis and Bangkok Declarations) → Kennedy (the “too
instant / too implied” objections cancel). Illustrate with A/80/L.65, and with Res. 2625 and 3314 as Charter interpretation rather
than custom.
“Equity is not a source but a technique of adjudication.” White supplies the whole answer: the 1920 rejection; infra/praeter/contra
legem; the never-used Art. 38(2); estoppel and clean hands; Frontier Dispute (equity is not arbitrary fairness); Tunisia/Libya (a general
principle “directly applicable as law”); Charney and Nelson on flexibility shading into ex aequo et bono. Add Panezi (equity in
Mosler’s third category with good faith and estoppel) and Kennedy (equity as the paradigm soft register). Close with the three-
stage delimitation method and CBDR-RC as proceduralised equity.

INTERNATIONAL LAW - MODULE II | PAGE 36