Choice of Language in Bilateral Treaties: Fifty Years of Changing State Practice


Choice of Language in Bilateral Treaties: Fifty Years of
Changing State Practice'

John King Gamble*
Charlotte Ku * *

I. INTRODUCTION

Treaties and treaty-making deservedly occupy a central place in
the international legal scholarly literature. As the major source of
international obligation, examining the form and content of treaties
can provide tangible and quantifiable information about the behavior
of states and the international environment in which they operate. On
a more immediate level, treaties serve as a means for interstate com-
munication. The language in which treaties are written affects how
widely and deeply treaty obligations are understood and, hence, followed.

Dozens of books and articles have concerned the genre of treaties2

and literally hundreds of pieces have focused on specific treaty-related

1. A discussion in 1982 with Professor Peter H. Rohn, Director of the Uni-
versity of Washington's Treaty Research Center, planted the seeds for this work.
Professor Rohn also commented on portions of an earlier draft of this paper and has
discussed these issues with us. We wish to acknowledge the research assistance of Kyle
L. Hannon, Angela Carone, Lisa Howells, Vicky Kohlmeyer, Sheryl Sivavec, Terri
Thomas-Mack, and Andy Watson.

* Professor of Political Science, The Behrend College, The Pennsylvania State
University.

* * Deputy Executive Director, American Society of International Law.
2. See ARNOLD McNAIR, THE LAW OF TREATIES (1961) for the most compre-

hensive coverage on the genre of treaties in the English language. See also SUZANNE

BASTID, LES TRAITAS DANS LA VIE INTERNATIONALE (1985); FRANCESCO CAPOTORTI, CON-

VENZIONE DI VIENNA SUL DIRITTO DEI TRATTATI (1969); INGRID DETTER, ESSAYS ON THE

LAW OF TREATIES (1976); TASLIM 0. -ELIAS, THE MODERN LAW OF TREATIES (1974);
A. FRANGULIS, THmIORIE ET PRACTIQUE DES TRAITfS INTERNATIONAUX (1934); GYORGY

HARASZTI, SOME FUNDAMENTAL PROBLEMS OF THE LAW OF TREATIES (1973); MYRES

McDoUGAL et al., The Interpretation of Agreements and World Public Order (1967);
PAUL REUTER, INTRODUCTION AU DROIT DES TRAITIES (1972); SHABTAI ROSENNE, DE-
VELOPMENTS IN THE LAW OF TREATIES, 1945-1986 (1989); and SHABTAI ROSENNE, THE

LAW OF TREATIES: A GUIDE TO THE LEGISLATIVE HISTORY OF THE VIENNA CONVENTION

(1973).



IND. INT'L & CoMP. L. REV.

issues s or categories of treaties. 4 Considerable scholarly attention has
been paid to the legal and logistical problems attendant with using
more than one language as official text of a treaty:

If the two can, without violence to the language, be made to
agree, that construction which establishes this conformity is
to prevail. In case they cannot be made to harmonize, other
rules of construction must be resorted to for the purpose of
determining, if possible, the common intention of the parties.'

3. See, e.g., , EFFECTIVE NEGOTIATION: CASE STUDIES IN CONFERENCE DIPLOMACY

(Johan Kaufman ed. 1989); OSCAR SCHACHTER, TOWARDS WIDER ACCEPTANCE OF UN
TREATIES (1971); ROBERT B. STEWART, TREATY RELATIONS OF THE BRITISH COMMON-

WEALTH (1939); M.T.Z. TYAU, THE LEGAL OBLIGATIONS ARISING OUT OF TREATY

RELATIONS BETWEEN CHINA AND OTHER STATES (1917); MARK E. VILLIGER, CUSTOMARY

INTERNATIONAL LAW AND TREATIES (1985); Hans Blix, The Requirement of Ratification,
30 BRIT. Y.B. INT'L L. 356 (1953); Maarten Bos, Theory and Practice of Treaty Interpretation,
27 NETH. Y.B. INT'L L. 3 (1980); Michael Brandon, The Validity of Unregistered Treaties,
29 BRIT. Y.B. INT'L L. 186 (1952); J.E.S. Fawcett, Treaty Relations of British Overseas
Territories, 26 BRIT. Y.B. INT'L L. 86 (1949); John King Gamble, Reservations to
Multilateral Treaties: A Macroscopic View of State Practice, 74 Am. J. OF INT'L L. 372
(1980); Michael J. Glennon, The Senate Role in Treaty Ratification, 77 AM. J. INT'L L.
257 (1983); Manley 0. Hudson, Reservations to Multipartite Instruments, 32 AM. J. INT'L

L. 330 (1938); Donald M. McRae, The Legal Effect of Interpretive Declarations, 49 BRIT.

Y.B. INT'L L. 155 (1978); Martin A. Rogoff, The International Legal Obligations of Signatories
to an Unratfied Treaty, 32 ME. L. REV. 269 (1980); Shabtai Rosenne, The Depository of
International Treaties, 61 AM. J. INT'L L. 923 (1967).

4. See, e.g., HUNGDAH CHIU & DOUGLAS M. JOHNSTON, THE PEOPLE'S REPUBLIC

OF CHINA AND THE LAW OF TREATIES (1972); THE HAGUE CONVENTIONS AND DECLA-

RATIONS OF 1899 AND 1907 (James Brown Scott ed. 1915); ROBERT HUDEC, DEVELOPING

COUNTRIES IN THE GATT LEGAL SYSTEM (1987); NATALIE H. KAUFMAN, HUMAN RIGHTS

TREATIES AND THE SENATE (1990); JOHN V.A. MACMURRAY, TREATIES AND AGREEMENTS

WITH AND CONCERNING CHINA, 1894-1919 (1921); ANTOINE PILLET, LES CONVENTIONS

DE LA HAYE (1918); ROBERT M. SLUSSER & JAN F. TRISKA, THE THEORY, LAW AND

POLICY OF SOVIET TREATIES (1962); KENNETH VANDEVELDE, UNITED STATES INVESTMENT

TREATIES (1992); ROBERT WILSON, U.S. COMMERCIAL TREATIES AND INTERNATIONAL

LAW (1960); Philip Alston, U.S. Ratification of the Covenant of Economic, Social and Cultural
Rights, 84 AM. J. INT'L L. 365 (1990); R.R. Baxter, The Geneva Conventions before the
U.S. Senate, 49 AM. J. INT'L L. 550 (1955); John King Gamble, Post World War H
Treaty-Making: The Task of the Third UN Law of the Sea Conference in Perspective, 17 SAN

DIEGO L. REV. 527 (1980); Pierre-Henri Imbert, La France et les traitis relatfs au droit
de l'homme, 26 ANNUAIRE FRANVAIS DE DROIT INTERNATIONAL 31(1980); Arnold McNair,
The Functions and Differing Legal Character of Treaties, 11 BRIT. Y.B. INT'L L. 100 (1930);
Bernard H. Oxman, The Third U.N. Conference on the Law of the Sea: The Tenth Session
(1981), 76 AM. J. INT'L L. 1 (1982) .

5. SAMUEL B. CRANDALL, TREATIES: THEIR MAKING AND ENFORCEMENT 389
(2d ed. 1916).

[Vol. 3:233



CHOICE OF LANGUAGE

Irconsistency between languages creates many problems. The so-
lution suggested by the Vienna Convention on the Law of Treaties,
however, is not altogether satisfactory:

1. When a treaty has been authenticated in two or more
languages, the text is equally authoritative in each language,
unless the treaty provides or the parties agree that, in case
of divergence, a particular text shall prevail. . . 3. The terms
of the treaty are presumed to have the same meaning in each
authentic text. 4. . . . when a comparison of the authentic
texts discloses .a difference of meaning . . . the meaning which
best reconciles the texts, having regard to the object and
purpose of the treaty, shall be adopted. 6

Of course, many problems arise when the treaty does not have
the same meaning in different languages. Samuel B. Crandall provides
a fascinating discussion of numerous cases where a seemingly minor
difference in wording between languages created serious problems once
the treaties were implemented.7 David P. O'Connell describes a solution
suggested in 1924 by the World Court: "[w]hen a treaty is in two
languages, and there is a discrepancy between them, each party is only
bound by the meaning of the text in its own .language." Both solutions
do little to resolve conflicts. 8

The focus of this article is a different aspect of language in treaties
- the choice of language or languages as official text or texts of bilateral
treaties. Some research has addressed the broader issue of multiple use
of languages in international organizations and multilateral treaties, 9

but bilateral treaties have received scant attention. This inattention
likely stems from the difficulty of examining the treaty practice of more
than 150 states in tens of thousands of treaties, a difficulty now largely
overcome by modern database management techniques. This article
examines a lengthy period of state practice, the half century between
1920 and 1970, in order to describe and understand language choices.
At first blush, this may seem like much ado about nothing. Of course,
it is the content, not the choice of official text, that matters most.

6. Vienna Convention on the Law of Treaties, May 23, 1969, art. 33, 63
AM. J. INT'L L. 875, 886 (1969) [hereinafter Vienna Convention].

7. CRANDALL, supra note 5, at 389-93.
8. DAVID P. O'CONNELL, INTERNATIONAL LAW 258 (2d ed., vol. 1, 1970). See

Mavrommatis Concessions, 1924 P.C.I.J. (Ser. A) No. 2, at 19.
9. MALA TABORY, MULTILINGUALISM IN INTERNATIONAL LAW AND INSTITUTIONS

(1980).

19931



IND. INT'L & COMP. L. REV.

However, the approach taken here, based as it is on an enormous
amount of state practice, can elucidate a number of important issues:

*To what degree has English taken over the lingua franca1° role
previously played by Latin and French?
oIs there any political dimension to language choice?
*Has the rise of the United States and the Soviet Union as
superpowers after World War II been accompanied by an
increased use of English and Russian as official texts?
*Is the emergence of the Third World as a major force in
the international system reflected in languages used in treaties?

II. THE "CONVENTIONAL" WISDOM ABOUT OFFICIAL TEXT

No doubt the treaty power claimed by all states includes the right
to use their languages as official texts of the treaties to which they are
party. Although the Vienna Convention does not deal explicitly with
the issue, it is inconceivable that any state lacks the power to have its
national language(s) used as an official text of a bilateral treaty. Lord
Arnold McNair stated simply that "parties are free to choose the
language or languages in which a treaty is expressed."' 1

Aside from the assertion that states may use whatever official texts
they wish in their treaties, the most frequent focus of work on language
choice concerns the use of lingua franca:

Until the eighteenth century the common language, or lingua
franca, of diplomacy was Latin. Not only did diplomats write
to each other in Latin but they even conversed in that medium.
Such treaties as those of Westphalia (1648), the Anglo-Danish
Treaty of 1670, and the Anglo-Dutch Treaty of 1674 were
all drafted and signed in Latin and that was in fact the general
practice. 12

Latin was the language of most treaties "[u] ntil about the beginning
of the eighteenth century,"' 13 although Mala Tabory found an exception
in that "towards the end of the fifteenth century, Castilian Spanish

10. The term, linguafranca, will be used here because it is so widely understood.

Some prefer "third language," but that is easily confused with those treaties that have
three different languages as official texts.

11. ARNOLD McNAIR, THE LAW OF TREATIES 30 (1961).
12. HAROLD G. NICOLSON, DIPLOMACY 231 (1939).
13. CHARLES C. HYDE, INTERNATIONAL LAW: CHIEFLY AS INTERPRETED AND

APPLIED BY THE UNITED STATES 1422 (1945).

[Vol. 3:233



CHOICE OF LANGUAGE

was the diplomatic idiom in many courts."1' 4 Eventually, Latin was
replaced by French due to "the political ascendancy of France under
Louis XIV."'' However, others were quick to assert that the position
enjoyed by the French language was at most indirectly related to the
military, diplomatic, and political successes of France. More than half
a century ago, James Brown Scott noted that linkages between national,
economic, and political power and language are at most indirect:

Le grec est devenu la langue du commerce intellectuel quand
la Grace est descendue A l'6tat d'une province romaine. I1 est
devenu mEme la langue d'une civilisation mondiale sous l'em-
pire d'Auguste et pendant les premiers si~cles de notre re,
pendant lesquels ses successeurs ont domin6 le monde, brisant
toute opposition matdrielle, imposant et leur volont6 et leurs
lois. Ni la puissance militaire, ni le prestige immense de
l'Empire ne rdussirent i remplacer le grec de l'Orient par le
latin. '

6

Even less tenable are the numerous assertions on the intrinsic
superiority of the French language. Harold G. Nicolson stated that
French "possesses qualities which enable it to claim precedence over
others for all purposes of diplomatic intercourse." 7 Scott, a francophone
of Ren6 Lvesque (Quebec separatist leader) proportions, believed that
the use of French could overcome the foreign mind-set and avoid
misunderstandings and ambiguity.' Manley 0. Hudson took Scott to
task, albeit politely, pointing out that in the immediate post-World
War I period, English was used increasingly. 19 Perhaps the most am-
bitious claim made on behalf of the French language was:

It is impossible to use French correctly without being obliged
to place one's ideas in the proper order, to develop them in
a logical sequence, and to use words 6f almost geometrical
accuracy. If precision is one of the major virtues of diplomacy,
it may be regretted that we are discarding as our medium of

14. TABORY, supra note 9, at 4.
15. LASSA FRANCES L. OPPENHEIM, INTERNATIONAL LAW: A TREATISE 771 (H.

Lauterpacht ed., 8th ed. 1955).
16. JAMES BROWN Scorr, LE FRAN AIS, LANGUE DIPLOMATIQUE MODERNE 129

(1924).
17. NICOLSON, supra note 12, at 234.
18. ScoTT, supra note 16, at 1.
19. Manley 0. Hudson, Languages Used in Treaties, 26 AM. J. INT'L L 368

(1932).

19931



IND. INT'L & CoMP. L. REv.

negotiation one of the most precise languages ever invented
by the mind of man.20

Writing almost fifty years later, Leslie Green reached entirely the
opposite conclusion: "There is, of course, nothing linguistically special
about either the French or English languages which warrants special
treatment. Indeed, there are no scientific grounds for thinking that any
human language is better, either in general or for some special purposes,
than any other." '21

Regardless of the dominant role played by French, states insisted
that the use of French was a matter of convenience, not a legal
requirement.22 This sentiment was expressed even at the time of the
Congress of Vienna when French was pre-eminent:

Article 120 of the General Treaty of the Vienna Congress of
1815 expressly observes that the fact of the French language
having been exclusively employed in all copies of that treaty
is not to be construed into a precedent for the future, and
that every Power reserves the right to adopt, in future ne-
gotiations, and conventions, the language which it had pre-
viously employed in diplomatic relations. 2

1

Commentators generally agree about when French was overtaken
by English as the model diplomatic language. "French has lost its
dominant position as the diplomatic language and, since 1919, English
has become at least as important. 2 ' The Paris Peace Conference in
1919 appears to have been the first time that English and French were
on an equal footing. 2 In the period since World War II, two dominant
trends have emerged in language choice. First is the increased use of
English where a lingua franca is employed. Second (generally regarded
as the prevailing practice) is the practice of preparing "versions of the
treaty in both languages .... ",26 United States practice early abandoned
French in favor of making English one of the official texts:

20. NICOLSON, supra note 12, at 234.
21. Leslie Green, Are Language Rights Fundamental?, 25 OSGOODE HALL L. J.

639, at 662 (1987).
22. JOHN BASSETT MOORE, A DIGEST OF INTERNATIONAL LAW 181-82 (vol. 5,

1906).
23. OPPENHEIM, supra note 15, at 771-72.
24. EDWARD D. BROWN & GEORG SCHWARZENBERGER, A MANUAL OF INTER-

NATIONAL LAW 129 (6th ed. 1976).
25. OPPENHEIM, supra note 15, at 772.
26. BROWN & SCHWARZENBERGER, supra note 24, at 129.

[Vol. 3:233



CHOICE OF LANGUAGE

Bilateral agreements concluded by the United States with other
countries are usually drawn up in the English language and
in the language of the foreign country, or in some other
foreign language selected by it. The French language has
frequently been the language selected. In a nurfiber of such
agreements the English language has been the only language
of the originals, while in some of the earlier agreements the
English language was not used, but French, Spanish, Arabic,
Turkish, or Portuguese, as the case might be, alone was used.27

John Bassett Moore cited a transmittal letter accompanying the
1785 consular treaty with France written by John Jay recommending
that "in the future, every treaty or convention which Congress might
think proper to engage in should be formally executed in two lan-
guages." ' In fact, the United States and the United Kingdom were
among the first and most insistent upon using their national language
in all their treaties.29

The position of the Soviet Union on language choice showed a
certain ambiguity. On the one hand, Soviet advocacy of the Russian
language because of its "inherent richness and beauty ' 30 is reminiscent
of pleas on behalf of French early in this century. In a different vein,
Academician Korovin declared that "Soviet treaty practice strictly ad-
heres to the principle of complete equality of languages of the contracting
parties. ',31

There are broader implications to these issues. Although linkages
between national, economic, and political power and language are
difficult to prove, the possibility of such linkages should not be over-
looked. Why was Latin chosen as the language of diplomacy? It was
the language of the university, the literati and of the Roman Catholic
Church, whose support for any treaty could make violation costly. A
more cynical explanation would hold that Latin was used to placate
the church in a minor, proforma way to divert attention from the content
of treaties.

In the modern era, especially in democratic states, public opinion
has become important in the negotiation and implementation of fair

27. GREEN H. HACKWORTH, DIGEST OF INTERNATIONAL LAW 37 (vol. 5, 1943).
28. MOORE, supra note 22, at 180.
29. BISWANATH SEN, A DIPLOMAT'S HANDBOOK OF INTERNATIONAL LAW AND

PRACTICE 458-59 (1979).
30. ROBERT M. SLUSSER & JAN F. TRISKA, THE THEORY, LAW, AND POLICY OF

SOVIET TREATIES 61 (1962).
31. Id. at 60.

19931



IND. INT'L & CoMP. L. REV.

and equitable treaties, "open covenants openly arrived at."3 2 The need
to make treaties accessible to the public-at-large has clear implications
for language choice. Nationalist aspirations, be they in 1648, 1914, or
1993, probably are the most potent political forces operating during
this millennium. Language is a factor in virtually all of these nationalist
movements. The choice of language in treaties can be a manifestation
of important domestic political forces. Furthermore, language choice
can be influenced by external factors such as hegemonic third parties
and the need for a guarantor power.

II. CATEGORIZING STATE PRACTICE

The recent trend in bilateral treaties has been to make the lan-
guage(s) of both party-states official languages of the treaty. 3 For
example, if the United States and Mexico conclude a bilateral agreement
on border problems, both English and Spanish will be official texts.
However, the most interesting cases arise when states opt for language
choices different from this now-dominant pattern. This choice may take
several forms,3 4 the most common of which are:

(i). the parties may opt for a convenient neutral language or lingua
franca, e.g., Bulgaria and Greece employ French; 5

(ii). the parties may decide to use the language of one of the
parties as the only official language, e.g., U.S. and Sweden use English;3 6

(iii). other - there are examples that fit into neither (i) or (ii)
but that combine them, e.g., Indonesia and Japan using Japanese,
Indonesian, and English. 37

Later this article shall deal with the more complicated situation
arising when a language, e.g., English, both plays a lingua franca role
and is the official language of one of the parties.

32. KENNETH N. WALTZ, MAN, THE STATE, AND WAR 95-103 (1959).
33. There are more complicated situations where a state may have more than

one official language, e.g., Belgium, Canada, Finland, India, and Switzerland.
34. Blix and Emerson adopted a scheme for classifying language choice that

consisted of a single authentic text, one of several authentic texts, several authentic
texts, all texts equally authentic and all texts equally authentic (by implication). THE
TREATY MAKER'S HANDBOOK 254-57 (Hans Blix and Jirina H. Emerson eds., 1973).

35. See, e.g., Payments Agreement, Dec. 5, 1953, Bulg.-Greece, 225 U.N.T.S.
145.

36. See, e.g., Exchange of Notes Between the Governments of the United States
of America and Sweden, concerning the denunciation of provisions No. 11 and 12 of
the Consular Convention of June 1, 1910, June 18 and 29, 1920, Swed.-U.S., 2
L.N.T.S. 157.

37. See, e.g., Treaty of Peace, Jan. 20, 1958, Japan-Indon., 324 U.N.T.S. 227.

[Vol. 3:233



CHOICE OF LANGUAGE

It is a daunting task to examine a sizable number of treaties to
determine what practice has actually been followed. Ironically, the
research for this article could have been carried out more easily a
decade ago.38 For purposes of this article, all bilateral treaties contained
in the League of Nations Treaty Series (L.N.T.S.) and in the United
Nations Treaties Series (U.N.T.S.) (those registered between 1945 and
1970) have been examined. This sample does not include all of the
bilateral treaties signed during the period between 1920 and 1970,19

but it does represent such a large number - and not a patently atypical
sample - that we assume that it is representative of all treaties con-
cluded during this fifty-year period. Before examining these treaties,
it is important to get some grasp of the magnitudes involved:

Table 1

TREATIES INCLUDED IN THE STUDY

NO. / %
SERIES YEARS TOTAL BILATERALS NO. / % NO. / %

COVERED NUMBER BET. STATES "RULES" EXCEPTIONS
League 1920-42 4,900 4,410 / 90 2,427 / 55 1,983 / 45
UN 1945-69 10,035 8,128 / 81 7,069 / 87 1,059 / 13

Although Table 1 is fairly straightforward, some elaboration may
prove helpful. This study includes nearly 15,000 treaties, the vast
majority of which were signed between 1920 and 1970. Most of these
treaties were bilaterals between states, however, some were multilateral
treaties and bilaterals between a state and an international organization.
The latter two categories have been excluded to avoid extraneous,
complicating factors. In the case of the L.N.T.S., there are 4,410
bilateral treaties between states (90% x the 4,900 total) of which 45%
(or 1,983) fall into the category of principal interest here: those that
have official texts that are not simply the official language(s) of the
parties. In the case of the U.N.T.S., only 13% of the bilaterals fall
into this category. Broadly speaking, about 15,000 treaties were ex-

38. The Treaty Research Center, established by Professor Peter H. Rohn at
the University of Washington, collected information about most of the world's treaties
and made it possible to answer questions such as those surrounding choice of official
text. Regrettably, the Treaty Research Center is no longer in operation.

39. For example, there is the problem of the "gap"-treaties that were never
registered with the League of Nations or with the United Nations.

1993]



IND. INT'L & COMP. L. REV.

amined and approximately 20% (3,000) of them fell into this "excep-
tion" group of treaties using a language other than the official language(s)
of the parties.

IV. BROAD PATTERNS OF BEHAVIOR

One of the most conspicuous findings, evident from Table 1, is
-the marked increase in the proportion of standard fare cases, i.e., those
using the official language(s) of both parties, from the League of Nations
to the United Nations. For the L.N.T.S., 45% of bilateral treaties
between states did something other than choose the languages of both
parties-what we defined as "exception" nearly was the "rule." In
the case of the U.N.T.S., the figure dropped to 13%. This represents
a marked change in behavior in a relatively short period of time. The
reasons for this change will be examined principally by analyzing the
45% of the L.N.T.S. and the 13% of the U.N.T.S. that constitute
exceptions as defined hereY°

The first question one must ask about these treaties regards the
number of official texts chosen. As Table 2 shows, almost all of these
treaties (95%) use only one language as official text. 4' The others, the
5% using two official texts, are interesting, but represent such a small
group that they can be dealt with separately almost on an individual
basis.

Table 2

NUMBER OF OFFICIAL TEXTS ACCORDING
TO SIGNATURE DECADE

NUMBER OF OFFICIAL TEXTS

ONE TWO > 2 TOTALS

DECADE 1920s 1024 63 0 1087
TREATY 1930s 841 48 0 889
WAS 1940s 112 3 0 115

40. A more satisfactory approach would have been to examine all bilateral
treaties and to compare the "exceptions" to the "rules." This was feasible only for
three "case studies," Canada, China, and Japan, discussed infra at Part V.

41. It must be remembered that this group of treaties contains only the "ex-
ceptions," those that do not have both countries' languages as official texts. If all
bilaterals were included, most would have two official texts.

[Vol. 3:233



CHOICE OF LANGUAGE

SIGNED 1950s 284 13 0 297
1960s 612 32 1 645
TOTALS 2873 159 1 3033

The next question to be addressed is which languages have been
used in these 3,033 treaties. This information is contained in Table 3.

Table 3
FREQUENCIES OF LANGUAGES AS OFFICIAL TEXTS

(TOTAL = 3,194)

ARABIC 1 GERMAN 119
AFRIKAANS 1 ITALIAN 13
CHINESE 1 JAPANESE 4
CZECH. 1 NORWEG. 3
DANISH 4 PORTUG. 26
DUTCH 3 RUSSIAN 6
ENGLISH 903 SPANISH 63
FRENCH 2009 SWEDISH 6

Sixteen different languages were used at least once, but English,
French, and German account for 97% of the cases. Figure 1 shows
the relative proportion of English, French, and German 42 over the fifty
year period. In the earliest decades, German occupied a small, but
significant, position. After World War II, German almost disappeared
as a language in bilateral treaty-making except, of course, in the usual
two language cases and between German-speaking countries.

The relative importance of English and French could hardly be
more striking. Before World War II, French. accounted for approxi-
mately 80% of the treaties. The most recent data available (the 1960s)
show that the frequencies of English and French have reversed with
English now accounting for almost 80% of treaties.

No satisfactory way exists to account in the data for an official
text acting both as lingua franca and as the national language of one of
the parties. This problem is most acute for the United States and
English. The United States is clearly the most prolific treaty maker in
the world4 3 and also insists on the use of English in virtually all of its

42. The incidence of other languages is too small to be represented on the
graph.

43. PETER H. ROHN, TREATY PROFILES 244-45 (1976).

19931



IND. INT'L & COMP. L. REV.

PERCENTAGE OF
Figure 1

TREATIES BY LANGUAGE BY DECADE
(TOTAL = 2,943)

1920's 1930's 1940's 1950's 1960's
DECADE

* German

[ French

* English

treaties. How can one determine the extent to which the dominance
of English since World War II is due to the power and influence of
the United States, to the ascendancy of the English language or whether
the two factors are inseparable? It wotld be fascinating to try to resolve
this matter by carefully controlling for each factor, perhaps even quan-
tifying differences between languages. For example, consider the cases
of Romania and the Netherlands. Since Dutch and Romanian are used
by relatively few people, there is virtually no chance that either will
be used as a lingua franca. One might assume that Romania would
prefer to use French as a lingua franca while the Netherlands would opt
for English because of relative linguistic proximities, Romanian to
French and Dutch to English.

[Vol. 3:233



1993] CHOICE OF LANGUAGE 245

Figure 2
PERCENTAGE OF TREATIES BY LANGUAGE BY DECADE

(EXCLUDES TREATIES WITH OFFICIAL TEXT SAME AS NATIONAL LANGUAGE

OF EITHER PARTY)

(TOTAL = 1601)

907

80

70

60 s

DECADE 50

40-

30-

20-:

10"-

0
1920's 1930's 1940's 1950's 1960's

PERCENT OF TREATIES

E French

El English

Figure 2 provides one way of addressing the problem of the in-
separability of the lingua franca and dominant party roles played by
languages. The figure includes those treaties that use only one language
as official text and for which that language is the national language of
neither party. Thus, any possible direct effects the United States would
have on the greater use of English, France on the greater use of French,
etc., have been removed. The dominance of English over French is
not as marked as in Figure 1, but the fifty-year shift is strong and
clear. During the 1920s, French accounted for 95% of these treaties
and English only 4%. By the 1960s, English stood at 56% while French
had dropped to 41%.

The large quantity of treaties included in this study limited the
amount of information collected about each treaty except, of course,
for the most basic attributes: signature date, force date, parties, lan-
guages, and a general categorization of the subject matter (i.e., political,
diplomatic, economic, cultural, military, and humanitarian). One could



IND. INT'L & CoMP. L. REV.

disagree about where a particular treaty should fall, but the categories
were applied uniformly.

Table 4 must be interpreted with great care due to the small
numbers, especially in the case of German and the co-existence of two
different factors, topic category and time. Since French dominated the
pre-World War II years, it is difficult to determine if the subject matter
of those treaties is due more to the era or to a tendency for the French
language to be used more when certain subjects are negotiated. Overall,
the clearest finding is a negative one. There is not much difference
among French, English, and German for the major topic categories.

Table 4

FREQUENCIES OF LANGUAGES BY TOPIC
(TOTAL = 2,908)

DIP.

CUL.

ECON.

HUM.

MIL.

POL.

TOTALS:

English
294
34%
26

3%
395
46%

8
1%

29
4%

114
13%

866
100%

French

988
51%
25
1%

844
44%
3

<1%
35

2%
36

2%
1931

100%

German
60
54%

1
1%

39
35%
0
0%

11

10%
0
0%

111
l00%

While Table 2 shows the infrequency of treaties with more than
one official text, Table 5 illustrates changes over time. Except for the
atypical decade of the 1940s, there is remarkable consistency.

Table 5

PERCENTAGE OF TREATIES WITH MORE THAN
ONE OFFICIAL TEXT

DECADE:
% WITH MORE
THAN ONE TEXT

1920s 1930s 1940s 1950s 1960s
5.8% 5.4% 2.6% 4.4% 5.0%

[Vol. 3:233



CHOICE OF LANGUAGE

Another way to approach the treaties with two official texts is to
examine language dyads, i.e., pairs of languages. With sixteen different
languages represented, a total of 120 possible dyads exists. Most of the
dyads were never used. This is hardly surprising-one would not expect
to find Chinese and Irish as official texts in an agreement other than
one between China and the Republic of Ireland. It might be interesting
to look at very small dyads, for example, Chinese/English, 4 Italian/
Russian, 5 and English/Irish, 46 each of which occurs only once.4 7 How-
ever, if any generalizations are to be made, the focus should be on
those dyads that occur fairly often. There are three of these that account
for 70% of the treaties in this multi-text category:

English/French: 43 occurrences
French/Portuguese: 18 occurrences
French/Spanish: 39 occurrences

Looking first at the English/French dyad, more than half of these
treaties have the United States or the United Kingdom as one party.
Evidently, English was used to accommodate them along with French
in its role as linguafranca. Typical examples here are the 1923 Agreement
between the United States and the Netherlands,4 8 the 1930 Agreement
between the United States and Greece,4 9 and the 1967 Agreement
between the U.K. and Yugoslavia.50 Although the overall incidence of
multi-texted treaties is quite constant over the fifty-year period, the
treaties involving the United States and the United Kingdom occurred
mostly before World War II, indicative probably of a wider linguafranca
role for English after World War II. The French/Portuguese dyad
consists of treaties between Portugal or Brazil and European countries.

44. Customs Treaty, Nov. 12, 1928, China-Nor., 87 L.N.T.S. 381.
45. Agreement on Economic and Technical Cooperation, June 2, 1961, U.S.S.R.-

Italy, 457 U.N.T.S. 263.
46. Exchange of Notes Constituting a Trade Agreement, Dec. 26, 1951, Ir.-

Switz., 558 U.N.T.S. 306.
47. This discussion considers only "exceptions," e.g., a treaty between China

and the United Kingdom with English and Chinese both as official texts would not
be counted.

48. Exchange of Notes Regarding Copyright on Musical Works, Oct. 2, 1922
Apr. 3, 1923, U.S.-Neth., 21 L.N.T.S. 175.

49. Treaty of Conciliation, June 19, 1930, U.S.-Greece, 136 L.N.T.S. '399.
50. Agreement Between the Postal Administration of the UK and the Admin-

istration of Posts, Telegraphs, Telephones of the Socialist Federal Republic of Yugoslavia

for the Exchange of Money Orders, June, 30, 1967, U.K.-Yugo. 642 U.N.T.S. 325.

1993]



IND. INT'L & CoMP. L. REV.

Typical examples, all of which are found in the L.N.T.S., are the
1921 Exchange of Notes between Portugal and Sweden5" and the 1932
Exchange between Brazil and Sweden. 52 The French/Spanish dyad is
similar to the French/Portuguese, except that it spans the entire fifty-
year period. Examples are the 1922 Exchange between Spain and
Sweden 3 and the 1961 Exchange between Chile and the Netherlands.14

V. THREE CASE STUDIES: CANADA, CHINA AND JAPAN

If a comprehensive system were available to examine treaties, it
would be possible to answer virtually any question about them including
uses of official text. For example, it is possible to construct a model
of language choice and test hypotheses about languages. Since no system
now exists, one has to be satisfied with carefully drawn subsets of
treaties, such as the group of 3,000 "exceptions" already discussed.
Another way of defining subsets is to examine the treaty behavior of
individual countries. In some instances, this examination has already
been done, most notably with the United States and Germany. It is
clear that the United States insists on English in virtually all of its
treaties-there was no appreciable change in this posture during the
fifty-year period covered in this study. The case of the German language
is interesting in that one can see German playing a fairly substantial
role through the 1930s, followed by its virtual disappearance thereafter,
as the single language of bilateral treaties, save for those instruments
between two German-speaking countries.

The three countries selected for "case studies" provide very dif-
ferent contexts for viewing the interplay between language choice and
broader sociopolitical forces. In each case, we have examined all bilateral
treaties during at least the fifty-year period,5 5 not just the "exceptional"

51. Exchange of Notes Constituting an Agreement Concerning the Exchange
of Notification with Regard to Person of Unsound Mind, Sept. 20, 1921, Port.-Swed.,

7 L.N.T.S. 143.
52. Exchange of Notes Constituting an Agreement Regarding Reciprocal As-

sistance to Brazilian & Swedish Nationals Suffering from Mental Disease, Jan. 27,

1932, Braz.-Swed., .178 L.N.T.S. 119.
53. Exchange of Notes Concerning the Application of the New Spanish Customs

Tariff to Merchandise Ordered in & Coming from Sweden, Dec. 29, 1921, Spain-
Swed., 9 L.N.T.S. 57.

54. Exchange of Notes Constituting an Agreement Concerning the Abolition

of the Travel Visa Requirement, Apr. 7, 1961, Chile-Neth., 453 U.N.T.S. 239.
55. Depending upon the country examined, it was necessary to extend the fifty-

year timespan to accommodate more recent political events.

[Vol. 3:233



CHOICE OF LANGUAGE

cases that have been the focus so far. This increases the information
base and provides a different view on the results.

A. Canada

Canada represents an interesting case study because of a marked
change in official policy, i.e., from the mid-1960s onward, bilingualism
was taken seriously as official policy by the Canadian federal govern-
ment. Thus, a change might be expected in official text languages to
occur sometime in the 1960s. A Royal Commission report published
in 1967 went so far as to tabulate numbers of bilateral treaties between
1928 and 1965. The Commission found that the English language
dominated Canadian treaty-making. 6 However, the report distinguished
between agreements and exchanges of notes in a somewhat confusing
way." In addition, some attention has been given to the matter of
language of official text in the scholarly literature. Allan Gotlieb's study,
which understates the dominance of English in earlier Canadian treaties,
commented:

So far as Canadian practice is concerned, there is a tendency,
in contemporary usage, to pay very strict attention to Canada's
bilingual character, perhaps resulting in more stringent prac-
tices than are generally found internationally. At one time,
the Canadian practice appears to have been to conclude bi-
lateral agreements which were authentic in one of the two
official languages, and in particular, English with English-
speaking countries, and French with French-speaking coun-
tries. Exchanges of notes .followed in general the usual inter-
national practice of being in only one language, English with
English-speaking countries and French with French-speaking
countries.

5 8

Professor Anne Marie Jacomy-Millette dealt somewhat more with
political forces at work in determining official texts. She wrote that,
earlier in Canadian history, English was the language of most Canadian
bilateral treaties.5 9 The policy has changed in more recent years:

However, there has recently been a new tendency in this
respect, in line with the bilingual policy introduced at the

56. REPORT OF THE ROYAL COMMISSION ON BILINGUALISM AND BICULTURALISM,

BOOK I, THE OFFICIAL LANGUAGES 64-65 (1967).
57. Id.
58. ALLEN GOTLIEB, CANADIAN TREATY-MAKING 70-71 (1968).
59. ANNE MARIE JACOMY-MILLETTE, TREATY LAW IN CANADA 60 (1975).

19931



IND. INT'L & COMP. L. REV.

federal level, first by the government of Mr. Pearson, and
since June 1968, that of Mr. Trudeau. Generally the official
text of bilateral treaties involving Canada is now in two ver-
sions, French and English, equal in authenticity, even where
the agreement was concluded with English-speaking countries
such as the United States. Use of both official languages for
the text of agreements is also the rule with the majority of
exchanges of notes or letters. 60

According to Jacomy-Millette, before the 1960s, it was not uncommon
to find English used in exchanges of notes even with French-speaking
countries.

61

In order to be certain of finding treaty-based evidence of official
bilingualism, the article extends coverage through the mid-1970s. Since
Canada is an active treaty maker, there is a sizable number of treaties
from which to draw conclusions: 453 with signature dates ranging from
1907 through 1977.62 First the overall picture and then a few interesting
examples will be discussed. The most telling finding is that 95.6%
(433) of these treaties have English as an official text, whereas only
32.7% (148) have French. Figure 3 shows the proportion of treaties
falling into each language category: English-Other; English-French;
English only; and French only-as well as the absolute numbers of
treaties in each language category (the small numbers atop the bars).

For Canada, the most telling question is the degree of change in
use of official texts over time. French has not quite achieved parity
with English in terms of frequency of use, but, as Figure 3 shows, the
change in the more recent time period is pronounced. Changes might
be particularly apparent in examining Canadian treaties with France
and the United States where, one would surmise, pressure would be
substantial to defer to the language of the other party. The number
of Canada/France bilaterals is too small to permit generalization. How-
ever, in the case of the United States, there is such a volume of treaty
activity 6l that trends, if there are any, should be discernible.

60. Id.
61. Id. at 48.
62. Signature date is the most accurate measure of policy of language since it

is unaffected by delays in registering the treaty for publication in the L.N.T.S. or the
U.N.T.S. Many of these delays are inexplicable.

63. PETER H. ROHN, TREATY PROFILES 80-81, 244-45 (1976). There were 113
bilaterals for the period between 1946 and 1965, the largest dyad for both countries.

250 [Vol. 3:233



CHOICE OF LANGUAGE

z

Figure 3
LANGUAGE COMBINATIONS IN CANADIAN

BILATERAL TREATIES

Fr. only 20

Eng. only 250

130

61

CJ)

0 Eng.-Fr.C)
m

Eng.-Other

0 10 20 30 40 50 60 70 80 90 100
PERCENT

* Post-1965

E3 Pre-1966

Table 664

USE OF ENGLISH AND FRENCH
CANADA/U.S.A. BILATERALS

Englis French
signed before 1966 128 2
signed after 1965 48 36

The contrast shown by Table 6 could hardly be greater. Before
1966, a bilateral treaty between Canada and the United States had less
than a 2% chance of having French as one of the official texts. After
1965, and the introduction of clear new policies on bilingualism, three
quarters of the treaties used both English and French. These findings

64. Use of percentages in tables like this one would be confusing because of

double counting, i.e., a treaty using both English and French would count in both
the French and English columns.

1993]

1



IND. INT'L & COMP. L. REv.

are clear evidence of political forces affecting behavior manifested in
international law.

Now that bilateral treaties as a whole have been described, this
article will discuss some of the individual treaties. Before 1950, most
Canadian treaties used only English as official text, unusual and notable
because, at this time, French dominated as a treaty language. French,
however, was used in a few instances. For example, treaties with French
speaking countries occasionally included both French and English as
official texts, e.g., the 1922 Convention between Canada and France ,65

the 1937 Agreement between Canada and Haiti, 66 and the 1946 Ex-
change of Notes between Belgium and Canada. 67 There are, literally,
only several examples where French is the sole official text. These seem
to be of two varieties, i.e., treaties with France such as the 1947
Exchange on War Damage Compensation,6 and a few cases where
Canada conformed to normal practice at the time, e.g., the 1926
Agreement with Norway. 69 This limited use of French should be seen
in the context of dozens of bilaterals with the United States, almost
none of which used French. In fact, it is not until 1962 that the first
bilateral between Canada and the United States with both English and
French as official texts was signed. 0

The post-1965 situation finds both English and French used in
most treaties. The exceptions are a dozen or so bilateral treaties with
the United States where only English is used. This period saw a
significant increase in treaties with three official texts. If one party
insists on more than one official text, that seems to open the door to
official texts for all languages of both parties. For example, the 1966
Convention for the Avoidance of Double Taxation71 between Canada
and Norway has English, French, and Norwegian as official texts. In

65. Commercial Convention, Dec. 15, 1922, Can.-Fr., 21 L.N.T.S. 38.
66. Agreement, Apr. 23, 1937, Can.-Haiti, 224 L.N.T.S. 59.
67. Exchange of Notes Constituting an Agreement Concerning Settlement of

War Claims, July 13, 1946, Belg.-Can., 230 U.N.T.S. 159.
68. Exchange of Notes Constituting an Agreement on War Damage Compen-

sation, May 5, 1947, Can.-Fr., 231 U.N.T.S. 81.
69. Agreement Concerning Exchange of Postal Parcels, Apr. 30, 1926, Can.-

Nor., 51 L.N.T.S. 203.
70. Interim Agreement Relating to General Agreement on Tariffs & Trade,

Mar. 7, 1962, Can.-U.S., 436 U.N.T.S. 3.
71. Convention for the Avoidance of Double Taxation and the Prevention of

Fiscal Evasion with Respect to Taxes on Income, Nov. 23, 1966, Can.-Nor., 604
U.N.T.S. 295.

[Vol. 3:233



CHOICE OF LANGUAGE

the 1967 Cultural Agreement," Canada and Belgium use Dutch, Eng-
lish, and French. This practice is different from that of the 1920's,
when they likely would have used French only, an official language in
both countries. Later, the article discusses the implications of this
"inflation" in the number of languages.

B. China

China presents a vastly different situation from that encountered
with Canada. One of the earliest Chinese treaties, the 1689 Treaty
between China and Russia, was written in Latin, the norm for that
time, and was negotiated through intermediaries (Catholic priests)."
Any discussion of Chinese treaty-making must consider the nature of
Chinese treaty relations as they developed in the 19th century. These
relations were characterized by Western powers and Japan using treaties
with China to maintain diplomatic, commercial, and political privilege.
These foreign powers assumed responsibility for implementing treaties
within China.7 4 Given this context, language choice surely was influ-
enced. However, interpreting the use of official texts in 19th century
treaties presents certain difficulties. Treaty series often did not specify
official texts. Further, a widespread practice of 20th century treaties-
including an explicit textual statement about authentic texts-was not
used in 19th century Chinese treaties.

Even given the above limitations, a reasonable estimate can be made
that the Chinese language was used in about one third of 19th century
Chinese treaties. For example, the 1842 Treaty between China and Great
Britain7 5 and the 1844 Peace Treaty76 between China and the United
States used Chinese and English. Many other treaties used only English,
e.g., the 1872 Treaty between China and Japan" and the 1894 Convention
with the United States. 78 A few treaties followed the mode of the times,
using French as lingua franca, e.g., the 1844 Treaty of Friendship, Com-

72. Cultural Agreement, May 8, 1967, Can-Bel., 637 U.N.T.S. 217.
73. Treaty, Oct. 1689, China-Russia, 18 C.T.S. 503.
74. See TREATIES AND AGREEMENTS WITH AND CONCERNING CHINA, 1894-1919

(John V.A. MacMurray ed. 1921); M.T.Z. TYAU, THE LEGAL OBLIGATIONS ARISING

OF TREATY RELATIONS BETWEEN CHINA AND OTHER STATES (1917).
75. Treaty, Aug. 29, 1842, China-U.K., 93 C.T.S. 465.

76. Treaty of Peace, Amity & Commerce, July 3, 1844, China-U.S., 97 C.T.S.
105.

77. Treaty of Commerce & Navigation, 1868, China-Japan, 144 C.T.S. 139.
78. Convention Regulating Immigration, Mar 17, 1894, China-U.S., 180 C.T.S.

1993]



IND. INT'L & COMP. L. REV.

merce and Navigation with France.7 9 One of the very few explicit textual
references to languages and language issues was contained in the 1899
Agreement between China and Germany, a treaty that appears to have
been written in German: "Article IV. All correspondence between the
Customs Office at Tsiutan and the German authorities and German
merchants shall be conducted in the German language .... Corre-
spondence in Chinese shall be likewise permitted." '80

The period between 1920 and 1980 saw the end of an era of Western
influence in China followed by a civil war and, ultimately, the bifurcation
of Chinese treaty-making. Thus, it is important to compare pre-1949
treaty behavior with that of Beijing and Taipei after 1949. Figure 4
provides a broad view of language use in Chinese treaties.

Figure 4
USE OF LANGUAGE IN THE TREATIES OF

PRE-1949 CHINA, TAIWAN, and P.R.C.

90-
-o 80
:' 70 5

0-
r60

050
h 40

S30m .
120-

flo-
UM0

Pre-1949 Taiwan P.R.C.

* Eng. only

Both

Both & Eng.

79. Treaty of Friendship, Commerce & Navigation, Oct. 24, 1844, China-Fr.,
97 C.T.S. 375.

80. Agreement Respecting the Establishment of a Maritime Custom Office at
Tsintaw, Apr. 17, 1899, China-Germany, 187 C.T.S. 383.

[Vol. 3:233



CHOICE OF LANGUAGE

Figure 4 uses three main language categories - English only;
Chinese and the language of the other party; and Chinese, the language
of the other party, and English. Several points stand out about the
period before 1949 (summarized in the leftmost portion of Figure 4).
First, there is not a single treaty with Chinese as the only official text.
Although the same may be true for Hungarian, Finnish, Urdu, and
Irish, Chinese was used by more people than any other language in
the world. Thirty-two from a total of 75 treaties have only one official
text and more than 90% of the official texts are English. English was
used as the only official text with many different parties including
Japan,"' the Soviet Union, 82 the United Kingdom,83 and Italy.8 4 The
other unilingual treaties used French and commonly were with
Switzerland.85

Somewhat more than half (43) of these 75 treaties have two or
more official languages. The pattern is very clear in showing that
virtually all of this subset uses English as one of the official texts.
If the treaty is between China and an English-speaking country, 86

the official texts are Chinese and English. If the party is not an
English-speaking country, English is still used, making a total of
three official texts. This pattern holds across the entire range of
languages including treaties with Japan,8 7 Austria,8  Portugal, 89

81. Treaty for the Settlement of Outstanding Questions Relative to Shantung,
Feb. 2, 1922, China-Japan, 10 L.N.T.S. 309.

82. Agreement for the Provisional Management of the "Chinese Eastern Rail-
way," with Declaration for the Regulation of the Situation Created for the Russian
Citizen in Pursuance of the Principle of Equal Representation of Citizenship of the
Two Countries on the Staff of the "Chinese Eastern Railway," May 31, 1924, China-

U.S.S.R., 37 L.N.T.S. 193.
83. Agreement Between the Post Office of North Borneo and the Post Office of

China for the Transaction of Money Order Business, July 17, 1936, China-U.K., 173
L.N.T.S. 343.

84. Exchange of Notes of the Settlement of the Claims for Damages Arising out

of the War, July 30, 1947, China-Italy, 12 U.N.T.S. 377.
85. See, e.g., Agreement with Regard to the Traffic in Narcotic Drugs Between

the two Countries, Apr. 12, 1927, China-Switz., 66 L.N.T.S. 427.
86. See, e.g., Parcel Postal Agreement Between China & Ceylon, Dec. 8, 1922,

China-U.K.-Ir., 137 L.N.T.S. 319.
87. Agreement Concerning the Exchange of Inserts, Letters & Boxes Between the

Two Countries, Dec. 8, 1922, China-Japan, 20 L.N.T.S. 233.
88. Treaty, Oct. 19, 1925, China-Aus., 55 L.N.T.S. 9.
89. Preliminary Treaty of Amity & Commerce, Dec. 1928, China-Port., 107

L.N.T.S. 93.

1993]



IND. INT'L & COMP. L. REV.

Sweden, 90 Thailand, 91 Ecuador, 92 and the Netherlands.9 3

Another view of Chinese treaty-making in this earlier period comes
from inquiring into what portion of Chinese treaties, signed between
1920 and 1948, had Chinese as an official text. The answer is 56%,
compared to 84% for English as an official text. Notwithstanding
arguments in favor of English as an international language, one can
imagine how this dominance of English over Chinese could create
feelings of neocolonialism and linguistic hegemony.

The post-1949 period provides the opportunity to compare the
behavior of two Chinese states, each with many treaty commitments.
The practice followed by the People's Republic of China (hereinafter
China or PRC) showed a marked change. All 41 treaties of the PRC
used Chinese as one official language. The norm was Chinese and the
language of the other party. There are a few tri-lingual treaties. A
1951 Agreement between China and Poland used Chinese, Polish, and
Russian. 94 Perhaps reflecting the deterioration in Sino-Soviet relations,
a 1960 Treaty between Czechoslovakia and PRC used only Chinese
and Czech. 95 In a very few cases, English was used, but only along
with Chinese and the language of the other party, e.g., a 1974 Agree-
ment on Maritime Transport with Denmark with Chinese, Danish,
and English as official texts. 96 Possibly, the use of English in this treaty
resulted from a combination of a relatively sympathetic Western party
and little knowledge of Danish in the Chinese foreign ministry.

The use of official languages by the Republic of China (hereinafter
Taiwan) also shows a dramatic change from the pre-revolution period.
Taiwan, like the PRC has adopted Chinese as an offical text. Some
exceptions exist, such as French being the sole official text9 7 and 17
treaties, mostly with the United States, in which only English is used.
Unlike the PRC treaties, however, Taiwan often used English as a
third language along with the national languages of both parties, e.g.,

90. Treaty Regulating Tariff Relations Between the Two Countries, Dec. 20,
1928, China-Swed., 107 L.N.T.S. 81.

91. Treaty of Amity, Jan. 23, 1946, China-Thail., 161 U.N.T.S. 127.
92. Treaty of Amity, Jan. 6, 1946, China-Ecuador, 7 U.N.T.S. 233.
93. Air Transport Agreement, Dec. 6, 1947, China-Neth., 43 U.N.T.S. 185.
94. Agreement Concerning Cultural Co-operation, Apr. 3, 1951, P.R.C.-Pol., 304

U.N.T.S. 187.
95. Consular Treaty, May 7, 1960, P.R.C.-Czech., 402 U.N.T.S. 209.
96. Agreement on Maritime Transport, Oct. 21, 1974, Den.-P.R.C., 991 U.N.T.S.

15.
97. Exchange of Notes Constituting a Provisional Agreement Concerning Civil

Air Transport, July 19, 1963, P.R.C.-Belg., 564 U.N.T.S. 23.

[Vol. 3:233



CHOICE OF LANGUAGE

treaties with Spain,98 Turkey, 99 and the Republic of Korea. 100 Taiwanese
treaty activity diminishes substantially after the mid-1970s. However,
there is some indication of a shift back towards the use of English as
the only language, likely a result of feelings of isolation resulting from
the widespread recognition of the PRC as the official representative of
China.

Overall, the use of official texts in Chinese treaties seems to reflect
nationalism, but differently than might be imagined. Similarly, the
forces of nationalism are manifest in the treaties of the PRC and of
Taiwan. This is evidence of the dominance and durability of nationalism
over more transient political ideologies.

C. Japan

The international legal scholarly literature, at least that available
to us in North America, mentions little about Japanese policy or
behavior in choice of official text languages in bilateral treaties. Adams
provided a general description of the behavioral side of choosing official
text:

English is the most frequently used language as an official
language occurs, first, where it is the only language stated as
being official and, second, in instances where three languages
are designated and English is designated to prevail in the
event of a dispute. These latter agreements usually are written
in Japanese, the language of the co-signatory and English. In
some instances, English and Japanese are recognized as being
equally authentic, most notably in treaties with the United
States, the United Kingdom and Canada. Finally, in instances
where English is not used as the official language, French is
the most prevalent language used. 10 1

The practice of Japan in treaty languages should be especially
interesting for two reasons. First, it can be a test bf whether the changes
in the use of English and French hold up in a non-Western context.
It is entirely possible that the dominance of French as lingua franca in
diplomacy and treaty-making is a remnant from the age of Euro-centric

98. Treaty of Amity, Feb. 19, 1953, P.R.C.-Spain, 181 U.N.T.S. 81.
99. Cultural Convention, Feb. 12, 1957, P.R.C.-Turk., 282 U.N.T.S. 125.

100. Treaty of Amity, Nov. 27, 1964, P.R.C.-Korea, 555 U.N.T.S. 3.
101. L. JEROLD ADAMS, THE THEORY, LAW AND POLICY OF CONTEMPORARY JAPANESE

TREATIES 41 (1974).

1993]



IND. INT'L & CoMP. L. REV.

international law. Perhaps these behaviors did not extend beyond Eu-
rope and its colonial umbrella. Second, it would be interesting to
examine whether Japan's subjugation following World War II is re-
flected in treaty behavior. One might hypothesize that the Japanese
language, as a symbol of militarism, would have been used less, es-
pecially immediately after 1945.

Table 8

NUMBER OF OFFICIAL TEXTS BY TIME PERIOD

ONE TWO >2
UNTIL 1945 56 (79%) 10 (14%) 5 (7%)
1946-1959 66 (47%) 65 (48%) 8 (5%)
1960-1974 69 (43%) 79 (49%) 13 (8%)

Clearly, there has been a substantial change in Japanese practice.
In the pre-war period, about four-fifths of Japanese treaties had only
one official text. Since World War II, almost half have had two official
texts. The use of English and French in those treaties with only one
official text has changed significantly over the three time periods
considered.

Table 9

USE OF ENGLISH AND FRENCH IN UNILINGUAL TREATIES

English French
until 1945 34 (61%) 22 (39%)
1946-1959 53 (80%) 10 (15%)
1960-1974 67 (97%) 2 (3%)

The diminution in the use of French is comparable in degree to
that observed for the entire world. The difference stems from the fact
that French never achieved the widespread use in Japanese treaties it
enjoyed in most of the world.

The anticipated change in language behavior after World War II
is not seen in these data, at least not in an unequivocal way. The
Japanese language was used in a much higher percentage of these

treaties in the immediate post-war period than in the period between
1920 and 1945. One way to focus on any changes that may have
occurred is to examine United States-Japanese bilateral treaties. In all
except the first time period, there were large numbers of these treaties

[Vol. 3:233



1993] CHOICE OF LANGUAGE 259

from which to draw conclusions. Not surprisingly, all of these treaties
are of one of two language types, English only or English and Japanese.

Figure 5
USE OF ENGLISH AND ENGLISH/JAPANESE IN
BILATERAL TREATIES BETWEEN JAPAN AND

THE UNITED STATES
80-

mug
•,e70 NEW;-"q,' :

01m :
Zn50- . ....

040-

m 30 . K .........
> IN

2

until 1945 1946-1959 1960-1974

I English

fl Eng./Jap.

Conclusions based on this information must be tentative. Few
treaties existed in the pre-war period. But this began to change in the
wake of World War II with Japanese being used as one official text
in most treaties. The most recent time interval shows some shift back
towards English - the amount of that shift is small (4%) and would
be unremarkable were it not for half of the most recent treaties, i.e.,
6/13, signed from 1971-1974, using English exclusively. One possible
explanation is that the wider use of the Japanese language was a point
of honor immediately after the war that eventually gave way to ex-
pediency, as Japan participated more actively in a global economy
dominated by the English language.

Some examples of specific Japanese treaties that fall into these
categories follow. During the first time period, 1919-1945, most Jap-
anese treaties had only one official text, usually English. Not surpris-



260 IND. INT'L & COMP. L. RFV. [Vol. 3:233

ingly, English was used in treaties with the United States1 0 2 and the
United Kingdom.10 3 English was the only official text in treaties with
Poland, 0 4 the Soviet Union, 0 5 Cuba, 1'6 and Thailand. 0 7 French was

used mostly in treaties with European countries, including the Neth-
erlands,10 8 Norway,"' 9 and Finland. 10 During this period, virtually all
Japanese treaties with China had three official texts, invariably using
English as a third language."'

The post-war treaties show less variety with only one dominant
mode, i.e., Japanese and the language of the other party. There are
a few cases where French remains the sole language, e.g., the 1953
Note with France ' 2 and a 1956 Exchange of Notes with Italy." 3 There
are many examples of English as the only official text. Most of these
are with the United States, but English also was used exclusively in

102. See, e.g., Treaty Concerning the Yap Island and the Other Islands Under
Mandate Situated in the Pacific North of the Equator and Exchange of Notes Relating
Thereto, Feb. 11, 1922, U.S.-Japan, 12 L.N.T.S. 201; Exchange of Notes Constituting
an Agreement Regarding the Development of Certain Portions of the Great Lakes-St.
Lawrence Basin Project, Oct. 14, 31 & Nov. 17, 1940, U.S.-Japan, 101 L.N.T.S. 63.

103. See, e.g., Exchange of Notes Respecting the Obligation of Article 8 of the
Treaty of Commerce & Navigation, June 14, 1924, Japan-U.K., 28 L.N.T.S. 538;
Convention for the Prevention of Smuggling of Intoxicating Liquors, Nov. 7, 1940,
Japan-U.K., 203 L.N.T.S. 267.

104. Treaty of Commerce & Navigation, Dec. 7, 1922, Japan-Pol., 32 L.N.T.S.
61.

105. Convention Embodying Basic Rules of the Relations Between Japan and
the Union of Soviet Socialists Republic, Together with Protocols A and B, Declaration,
Exchange of Notes, Annexed Note and Protocol of Signature, Jan. 20, 1925, Japan-
U.S.S.R., 34 L.N.T.S. 31.

106. Exchange of Notes Constituting a Provisional Commercial Agreement, Dec.
21, 1929, Japan-Cuba, 111 L.N.T.S. 13.

107. Treaty Concerning the Continuance of Friendly Relations Between the Two
Countries and the Mutual Respect of Each Other's Territorial Integrity, June 12,
1940, Japan-Thail., 204 L.N.T.S. 131.

108. Treaty, Oct. 12, 1921, Japan-Neth., 12 L.N.T.S. 239.
109. Exchange of Notes Constituting an Arrangement Concerning the Exchange

of Notifications with Regard to Persons of Unsound Mind, Oct. 23, Nov. 6, 1923,
Japan-Nor., 33 L.N.T.S. 265.

110. Exchange of Notes Regarding the Abolition of Passport Visas Between the
Two Countries, Feb. 25, 1928, Fin.-Japan, 71 L.N.T.S. 467.

111. See, e.g. Agreement Concerning the Exchange of Correspondance Between
the Two Countries, Dec. 8, 1922, P.R.C.-Japan, 20 L.N.T.S. 205.

112. Peace Treaty, Apr. 25, 1953, Japan-Fr., 187 U.N.T.S. 41.
113. Exchange of Notes Constituting an Arrangement Concerning the Abolition

of Visas on Passports, Jan. 11, 1956, Japan-Italy, 267 U.N.T.S. 175.



CHOICE OF LANGUAGE

treaties with Sweden,' 14 Turkey," 5 Czechoslovakia," 6 and Argentina." 7

VI. CONCLUSIONS

This survey of language choice in bilateral treaties has produced
results that should be of interest along several dimensions. It provides
a precise description of broadly based state practice in the bilateral
aspect of treaty-making. Three prevalent generalizations about treaty
making are confirmed:

1) there is no international law standard about use of official
texts;
2) English now dominates treaty-making as French once did;
and
3) modern practice tends to use the language of both parties
as official texts.

Although all three are true, they are too broad to describe, much
less to explain, state behavior. The absence of an international law
standard hardly obviates the need to understand practice. Neither of
the other two assertions is an ironclad rule-it is important to know
the degree to which English now dominates and what portion of recent
treaties use the languages of both parties. English is now the diplomatic
language, although French has been more resilient than many would
have believed. The issue of the utility of French language ability
resurfaced during the informal negotiations that, ultimately, led to the
selection of Boutros Boutros-Ghali as UN Secretary-General.' 8

In many instances, choice of language reflects political forces, most
notably, nationalism or international status in this aspect of treaty
behavior. German has all but disappeared as a lingua franca. The Soviet
Union and the Russian language are the exception-there was virtually
no tendency for the rise of the U.S.S.R. as a superpower to be

114. Convention for the Avoidance of Double Taxation and the Prevention of
Fiscal Evasion with Respect to Taxes on Income, Dec. 12, 1956, Japan-Swed., 318
U.N.T.S. 309.

115. Exchange of Letters Constituting an Arrangement Concerning the Mutual
Abolition of Passport Visas, Nov. 5, 1957, Japan-Turk., 318 U.N.T.S. 411.

116. Treaty on Commerce, Dec. 15, 1959, Japan-Czech., 383 U.N.T.S. 277.
117. Exchange of Notes Constituting an Agreement Concerning Reciprocal Waiv-

ing of Passport Visas, Dec. 26, 1961, Japan-Arg., 451 U.N.T.S. 71.
118. See Paul Lewis, Ajricans Pressing Bid for U.N. Post Say Next Chief Must Be

From Their Continent and Warn of Clash at Assembly, N.Y. TIMES, Oct. 22, 1991, at All;
Judith Miller, A Born Secretary General, N.Y. TIMES, Nov. 22, 1991, at A12.

19931



IND. INT'L & COMP. L. REV.

accompanied by increased use of the Russian language. In the case of
Canada, domestic policy changes in the form of official bilingualism
explain the behavior. The treaty practices of China and Japan indicate
that both national pride and convenience are factors in language choice.

It is hoped that this research will stimulate further efforts to show
how information basic to international law-treaty attributes in this
case-can be a meaningful indicator of state behavior. A disappointing
development in academic international law circles over the last quarter
century is the abandonment of international law by mainstream political
science.11 9 One of the main reasons offered for this change is inter-
national law's inability to use the quantitative techniques widely applied
in political science since the 1960s. These treaties provide one concrete
example of how phenomena of interest to international lawyers can be
meaningful indices of national behavior of the sort in which political
scientists should be very interested.120 In fact, the care with which inter-
national law experts are taught to approach their work increases the
likelihood that the data collected will be meaningful, sensitive indicators
of behavior.

The patterns in language choice uncovered by this research can
be viewed in a broader context. There is an interesting and rapidly
expanding trend of literature in political science and psychology dealing
with language problems, especially within multilanguage societies. A
major problem is balancing the competing values of simplicity and
efficiency against the often emotional feelings individuals and whole
groups have for national languages. The logical solution is an easy-to-
learn artificial language. The advantages of an artificial language are
significant:

1. learnable (because of grammatical and lexical regularity);
2. powerful (having true-to-nature terminologies, logical struc-
tures, and freedom from idiomatic restrictions); and,
3. fair (having no native speakers).' 2 1

The most successful of these, Esperanto, never achieved any level
of proficiency by more than one person in a thousand, 2 2 and has never

119. For a wide ranging discussion, see Charlotte Ku (chair), Bridging the Gap
Between Political Scientists and Lawyers, 81 ASIL Proceedings 1987 381, 381-394 (1990).

120. Twenty years ago, Harold Lasswell made these same points, but almost
no one listened. See Michael Barkun & Wesley B. Gould, INTERNATIONAL LAW AND

THE SOCIAL SCIENCES (XV-XX by Harold D. Lasswell 1970).
121. Jonathan Pool, The World Language Problem, 1 Univ. of Wash.: Language

and Soc'y Papers # LD2 (1990), reprinted in RATIONALITY AND SOCIETY 1991.
122. PETER G. FORSTER, THE ESPERANTO MOVEMENT 1-3 (1982).

[Vol. 3:233



CHOICE OF LANGUAGE

been used in either the L.N.T.S. or the U.N.T.S. It appears that the
international community will have to endure continued use of more

official texts-on the part of states. One of our most striking findings

is the huge increase in average number of texts used from the League

of Nations to the United Nations. The statistics may be sterile, but

they paint a clear picture. L.N.T.S. treaties averaged about 1.6 official

texts, while U.N.T.S. treaties average about 2.0 texts per bilateral
treaty. Faced with the combination of perceptions of linguistic

imperialism 123 and feelings that every national language has "a basic

right to its possession and, if necessary, its defence,"124 there probably
will be further increases in the number of languages used. It is possible

to develop theoretical solutions to try to curb the cumbersome use of
several official texts. One commentator suggests a tax or subsidy ac-

cording to ease or difficulty in operating with languages. 25 Although

an interesting idea, it is fraught with political and operational problems.
Professor Manley 0. Hudson, in a seminar he conducted at Har-

vard Law School in 1932, asked each class member to examine official
texts used in the treaties of her/his country for the period January 1,
1921 - December 31, 1930.116 Although the class analyzed the treaties

of only twelve countries,127 their findings are consistent with those
discussed here. Hudson realized the shortcomings of his study noting

that "the limited number of treaties considered does not permit a final
and general conclusion to be drawn. 1 ' 28 Sixty years later, the research
for this article has met Professor Hudson's objection.

Many possibilities exist to broaden the applicability and increase
the elegance and parsimony of this research. There are prospects for

using the broad contours developed here to examine more concrete
matters:

*How has the World Court dealt with issues of language and
official text;
*Do third languages used in treaties relate to substantive treaty
provision, e.g., the guarantor role of third parties;

123. TABORY, supra note 9, at 5.
124. PETER L. BERGER, FACING UP TO MODERNITY 161(1979).
125. Jonathan Pool, The Official Language Problem, 85 AMER. POL. Sci. REV. 495,

498-502 (1991).
126. Hudson, supra. note 19, at 370.
127. Id. at 370-72 - China, Colombia, France, Germany, Hungary, Italy,

Japan, Peru, Poland, Switzerland, the United Kingdom, and the United States.
128. Id. at 372.

1993]



IND. INT'L & COMP. L. REV.

*Does language choice affect the probability of treaty
compliance?

Language choices in certain multilateral treaties might be exam-
ined. One could learn little from general multilateral treaties, since
almost all of them now employ all six languages of the United Nations
as official texts; but plurilateral treaties are more interesting. For ex-
ample, the five Scandinavian countries often negotiate pentalingual
treaties. 129 Putting aside intraregional politics, it would be much simpler
if English were used exclusively, especially since all five use English
often in their other treaty relations. This exclusive use of English would
seem all the more rational since, if one exceeds three languages, com-
plexities and problems seem to escalate. 130 Maybe it is time to consider
radical solutions like a linguistic difficulty supplement or tax.

There can be no doubt that huge amounts of resources are used
for everything from foreign office personnel to library budgets because
of growing sensitivity about the use of national languages. Maybe this
is another area where some curbs on nationalistic forces might be
considered. An alternative view holds that the widespread use of bi-
lingual and trilingual treaties might be a small price to pay if it were
to improve, even marginally, the likelihood of agreement on substantive
treaty provisions. Given the large number of possible languages which
could be used in treaties, coupled with the increasing number of tech-
nical subjects covered by treaties, language choice remains an important
consideration as the world strives for greater equality and order through
increased clarity and compliance in treaty relations.

129. See, e.g., Convention Respecting the Application of the Accident Insurance
Laws of the Various States to Cases where an Employer in Lieu of the Contracting
States Carries on a Business or Employs Workers in Another of the Said States, March
3, 1937, 182 L.N.T.S. 127.

130. Heinz Kloss, Types of Multilingual Communities, in EXPLORATIONS IN SOCIO-
LINGUISTICS 7 (Stanley Lieberson ed. 1966).

[Vol. 3:233