Search form


Drazen Pehar looks specifically at the use of ambiguities in peace agreements. Pehar explains why ambiguities are so often used and why diplomats and others involved in international relations may think it best to eliminate ambiguities from peace agreements altogether. He goes on to demonstrate, however, with numerous examples, that while ambiguities have led to a continuation or re-starting of hostilities in some cases, in many other cases they have provided the only bridge between conflicting parties and allowed for a cessation of violence.

Language and Diplomacy. Ed by J. Kurbalija and H. Slavik (2001)
Drazen Pehar

Resource text:

In this paper I will talk about ambiguous language as it is used in peace agreements. I use the term "peace agreement" somewhat broadly to refer to agreements that resolve conflicts of interest of any kind, not only those that put a close to war.

I will start with a definition and classification of ambiguities and then say a few words about the theoretical explanation of their origin. Next I will explain the basic rationale behind the choice of ambiguous language in a setting such as a peace agreement and the basic reason for dislike of that choice as well. Third, I will present a number of cases from the real world of diplomacy. Fourth, I will try to assess the relative weight and plausibility of a number of arguments both in support of and against the use of ambiguities in peace agreements. Finally, I will add a few reflections on implications that I believe should be drawn from my analysis of "peace-making", constructive, or, as some have called them, creative ambiguities.



Considered per se, ambiguities represent an obstacle to any reflection on language. While the primary aim of language consists of transmitting information, in conveying a piece of knowledge from human being A to human being B, ambiguities seem to run contrary to that aim as they leave a message recipient with a less transparent and less usable kind of data. However, language does not perform only an informative function, but, as Karl Buehler pointed out long ago, at least two more functions.1 Language performs expressive and vocative functions, in addition to informative functions. The expressive function rests on the human need to express feelings, interests, and preferences, while the vocative function rests on the need to influence the feelings, interests or preferences of others. So, although it may not be informative, ambiguous language may thus play an important role, especially when it comes to those subjects where human beings are a central topic of discussion. We often, predictably, find ambiguities playing an essential part in the phenomena of linguistic politeness that combines two conflicting desires: the desire to convey a clear message that, in its authentic form, would insult a human being, and the desire to soften the message to avoid making that human being feel offended or humiliated.2

What is, then, an ambiguous expression, and how does an ambiguous word, sentence or text differ from one that is not ambiguous? Unfortunately, ambiguities are difficult to define precisely because of their ambiguous nature.

R. Munson defined ambiguity in the following way: "An expression is ambiguous when it has more than one meaning and it is used in a situation or context in which it can be understood in at least two different ways."3

I believe that one could and should amend this definition to make it more precise. In order to qualify as an ambiguity an expression must generate not only "at least two different meanings", but also two incompatible and unrelated meanings. It is only then that an expression is truly ambiguous. Metaphors and visual models are often the best way to present difficult intuitions in a more graspable form: the picture that most clearly depicts the common-sense understanding of an ambiguous expression is the optical illusion picture of the "duck-rabbit".4 The "duck-rabbit" picture can be seen as a duck as well as a rabbit, though a normal picture of a rabbit is incompatible with a normal picture of a duck. In the same way, one can read an ambiguity in two incompatible ways. Note that this picture can be interpreted both as a duck and as a rabbit, but, also, with some difficulty, as a third neutral image, which stands equally far from, or close to, both rabbithood and duckhood. The same applies to ambiguities; ambiguities are pieces of language that 1. can be interpreted as meaning A, 2. can be interpreted as meaning B, and 3. cannot be interpreted as A and B simultaneously, but, eventually, as a neutral (re)source, from which, under specific focuses of vision/interpretation, both A and B might at separate times spring. And that is what makes ambiguities a sort of amazement to common sense, just like the "duck-rabbit" picture generates an optical unease.

I have to admit that my definition begs many questions. For instance, if ambiguity carries two different, incompatible and unrelated, meanings, why do we then not consider it sheer contradiction? An ambiguity does not say both "A is B" and "A is non-B". Actually it somehow says both "A is B" and "A is C", but under different angles. It may be interpreted as "A is B" in one light, but then, in exactly the same light, it cannot be interpreted as "A is C". Taken as a whole, as "A is B" and "A is C" combined, an ambiguity is just a possibility of a sentence, not a real sentence. And that is probably the reason we do not consider it a sheer contradiction. But what is this mysterious thing called "a possibility of a sentence"? I do not know. I believe ambiguity is a sort of phenomenon which presents a serious challenge to any theory of mind and language and which demands that we face the outermost limits of our language in the very medium of the language. That is exactly why it is so difficult to grasp the true definition of this linguistic phenomenon.


Let me now propose a typology of ambiguous expressions, which I will close with a few words about a theoretical explanation of their origin.

Ambiguity can emerge at any level of discourse - in words, in a sentence, or in a set of sentences. One can thus distinguish between referential, syntactical, and cross-textual ambiguities, depending on whether they occur in a single word, a sentence, or a set of sentences; a text. I will provide a definition and examples for each.

Referential Ambiguity

Croesus, an ancient king of Lydia, asked the oracle at Delphi to foresee the outcome of his attempt to conquer the Persian Empire. The oracle, as clever as always, issued the following prophesy: "If you attack the Persians, you will destroy a mighty empire."5 In this sentence, the expression "mighty empire" was used in an ambiguous way. The way Croesus understood the expression was not even close to the way the oracle of Delphi intended it. What the oracle meant by "mighty empire" was the empire of Lydia whose king was Croesus himself, not the empire of Persia, as Croesus understood. So, Croesus, acting on his mistaken understanding of the expression "a mighty empire", did destroy a mighty empire, but it was his own. I also believe that Croesus understood the term "destroy" from the prophetic message too narrowly, because the oracle intended it to mean both "destroy" and "self-destroy".

This is a referential type of ambiguity, as the meaning of the expression "mighty empire" in the context of the above sentence does not allow the hearer to refer to, or to pick out correctly, the object which the oracle intended. This is the simplest kind of ambiguity, an ambiguity based on a word. A word usually has many meanings, and some words mean more than others. "A mighty empire" can refer to any one of all mighty empires: it does not specify which one.6

Syntactical Ambiguity

Munson calls a second type of ambiguity "grammatical", while K. Bach calls it "structural".7 I myself prefer the term "syntactical", because this type of ambiguity actually rests on the vagaries of syntactical relations within a sentence.

Sometimes a sentence contains a number of specifications, but their direction - what they specify - is multiple, opening the sentence to several different and incompatible interpretations, making it ambiguous. While referential ambiguity is ambiguous due to an intrinsic quality of the parts of a sentence, syntactical ambiguity is ambiguous due to a relation between the parts of a sentence. For instance, if a sentence has the form "A B C" where C should retroactively specify A or B, and if furthermore C could be understood as specifying both A and B, but not both of them simultaneously, then such a sentence is the classical case of syntactical ambiguity. For instance, the sentence "I am prepared to give the sum of one million US dollars to you and your husband" may be understood in two different ways: 1. I am prepared to give the sum of (1 million $) (to you) and (your husband), or 2. I am prepared to give the sum of (1 million $) to (you and your husband). Another example is as follows: If one says "I hate him more than you", this could mean "I hate him more than you hate him", or "I hate him more than I hate you."

The oracle of Delphi was a mastermind in creating syntactical ambiguities as well. A famous Latin translation of one of its prophecies reads "Ibis, redibis numquam peribis in bello." Two different translations and interpretations may be provided for this sentence. 1. "You'll leave, and you shall never return as you will perish in the war." 2. "You'll leave and return, and you shall not perish in the war." "Numquam" here specifies in too many directions; prima facie it can specify both "redibis" and "peribis", but it cannot specify both simultaneously. However, nothing in the sentence indicates to which verbal phrase the "numquam" qualifier should be allocated.

This kind of ambiguity is what one of the most prominent modern linguists had in mind when he wrote: "ambiguity is a one-many relation between syntax and sense."8

Cross-Textual Ambiguity

Theorists of ambiguity tend to forget that there is another type of ambiguity, which I call "cross-textual". This type of ambiguity rests not on a discrete phrase or sentence, but on a larger body of a text comprising many sentences. This type is thus more complicated than the previous two, involving not only semantics of phrases or semantics of propositions and sentences, but also semantics of texts; semantics of sets of inter-related sentences.

We saw that the syntactical type of ambiguity depends on specifications that have multiple directions, that may be interpreted as specifying several parts of a sentence but not at the same time. Such specifications, however, do not have to be contained within a single sentence, but may be dispersed across a text. Such cross-dispersion may become another source of ambiguity. For instance, if we refer to a "special relationship between two men" by using a number of specifications in a text and refer to it again in later parts of the same text by using a number of new specifications, this may result in the type of ambiguity I call "cross-textual". However, friction between the two sets of specifications is a key precondition for the creation of a cross-textual ambiguity and without it cross-textual ambiguity would not occur at all.

This kind of ambiguity is best exemplified with so-called "open-ended sentences" which can be found in legal texts. For example, a chapter in a peace treaty may begin with a precise enumeration of the powers that one entity, for example, a central federal authority, may exercise. But at the end of the chapter an open-ended provision is inserted, which may, for instance, state that "the central federal authority may exercise some other duties as well". This clearly introduces an ambiguity into the list of the powers vested to that authority, as the open-ended provision in a way conflicts with the clear starting provision enumerating the authority's basic powers. If basic powers are clearly spelled out, then why would one need an additional open-ended clause to leave room for the expansion of the powers? Furthermore, does such an open-ended provision imply that the parties to the document already consented to an extension of the list of basic powers? Or have they merely indicated that they would be willing to enter into negotiations dealing with such a possibility without any predetermination as to their final outcome?

It is thus clear that ambiguity is not limited to the realm of phrases, of basic semantic units or sentences, but that it may also be brought into being in a wider context and through an interplay between sentences or sets of sentences. In this case ambiguity must be interpreted as an attribute of a set of sentences or of an entire text, which does not imply any radical departure from already existing practice. In principle we do not hesitate to call a text or a story ambiguous, if we believe that we have reasons to do so.

Now, as to the issue of theoretical explanations of the origins of ambiguity, the most rudimentary theory was proposed long ago by the Greek philosopher and scientist Aristotle.9 His theory says that the number of items that form the vocabulary of any human language is much smaller than the number of realities that the vocabulary-items are supposed to refer to, to make the human language meaningful and functional. In modern terminology, we say that the world is a continuum while the words we use to describe it are discrete. In other words, reality is much more complex than language, its demand is, so to speak, always higher than the supply of words that we use to denote it. It is exactly for this reason that we are bound to draw imperfect and vague boundaries in our language, relative to the realities rich in nuance and distinction that our language refers to. That is why entries, or items, of a vocabulary by necessity fail to perfectly match the items that are the subjects of their reference, and that is also why we use a single language-item to refer to many dissimilar and incompatible world-items. In other words, supply shortage in words makes their average value higher, but it also leaves the consumer needs and demands to an extent unmet. It was through this simple theory of disproportion that Aristotle explained the phenomenon of homonyms, words that have the same sound and spelling but different meanings, and by which he also set the tenets to explain the origins of ambiguous expressions. You will notice that this theory does not apply to syntactical and cross-textual kinds of ambiguity. However, I do not think that Aristotle would find it difficult to broaden the scope of his theory to cover all three types. For instance, he would probably say that the world is paratactic, while language itself is syntactic, and that explains the origin of the syntactical type of ambiguity. In other words, words come in succession, one after another, which may lead to confusion as to which phrase should be coupled with which. The world comes in totality; its items are co-present simultaneously, which leaves us, in principal, no doubt as to which item flows from which. Syntactical ambiguities offer at least two versions of the paratactic world, two equally possible paratactic worlds.

However, this theory also begs many questions. First, its focus is entirely on representational aspects of language. It derives ambiguity from the fact that language represents reality in a less-than-perfect way. Language, however, serves not only the function of representing reality, but many other functions as well. For instance, it establishes relations between people and sometimes is even used to mislead and confuse others, which obviously involves a misrepresentation of reality. Language is thus subject to not only representational pressure, but to other, social and psychological pressures as well. Remember, we often use ambiguity to mould a message in a polite form. So why not include both representational and social-psychological demands on language into an etiology of ambiguous language? Secondly, Aristotelian theory presents the ambiguous aspect of language as a symptom of its insufficiency, of its disability to portray the world as it really is, in short, as a flaw. Contrary to that view, perhaps we should look at ambiguity not as a symptom of the representational insufficiency of language, but as a normal and recurring phenomenon, which we all individually and occasionally produce by adding new meanings to the words we adopt from our linguistic community. Perhaps ambiguity should more aptly be understood as a manifestation of the irreducibly individual character of all language, of our individual capacity to use the shared word pool for novel purposes. In other words, a number of factors, including representational, social and individual ones, may explain the origins of this interesting linguistic phenomenon. I believe that a comprehensive and working theoretical explanation of the origins of ambiguity should incorporate all of them.10


Now, why would one put ambiguous wording in a peace agreement?

Well, mediators, or those drafting such texts, reason approximately in the following way. If two parties have strong and contradictory interests, and if it seems that neither side is ready to concede a part of its maximum demand, and/or if the negotiations are running short of time and the parties can not discuss such concessions in more detail, then the issue of conflicting interests can be resolved by, so to speak, simulating a compromise in a very rudimentary form. The mediators may come up with a formula which is open to at least two different interpretations; which can carry at least two meanings, A and B, one to gratify the interests of party A and another to gratify the interests of party B. Meaning A will thus stand in harmony with the interests, or preferences, of party A, while meaning B will stand in harmony with the interests, or preferences, of party B. Thus, the mediators maintain the integrity and comprehensiveness of the draft, and, at the same time, make a small step towards elaborating, at a later stage, a compromise between the maximum demands by erstwhile conflicting, now negotiating, parties. In other words, ambiguities make sure that, on the one hand, the parties retain their own individual perceptions as to "how things should proceed" and that, on the other, one common language is adopted, which both parties may later equally use. To illustrate this with an example, the Rambouillet mediators started with the premise that interests of Serb and Kosovar-Albanian delegations to the Rambouillet negotiations contradicted one another. The Serb delegation, for instance, wanted to maintain the status of Kosovo as a province with very little, or no, competence in foreign relations, among other things. The Kosovar-Albanian delegation had different interests; to turn Kosovo into at least a fully-fledged republic on equal footing with the other two republics of the Federal Republic of Yugoslavia: Serbia and Montenegro. This status, of course, entails the capacity to run certain aspects of foreign relations independently from the central federal authority. Secondly, the Serb delegation wanted the Rambouillet draft agreement, such as it was presented in Rambouillet, to remain binding in the foreseeable future. The Kosovar delegation had an opposing interest which was not envisaged by the Rambouillet draft: to turn Kosovo one day into a fully independent entity. They therefore wanted to see a revision of the agreement as soon as possible as well as to organise a referendum to check the will of the people of Kosovo vis-à-vis the status of Kosovo within the Federal Republic of Yugoslavia.

Mediators to the Rambouillet process decided to use an ambiguous wording to bridge the gap between the aforementioned interests. The constitution, as the key part of the Rambouillet Draft agreement, stipulated that "Kosovo shall have authority to conduct foreign relations within its areas of responsibility equivalent to the power provided to Republics under Article 7 of the Constitution of the Federal Republic of Yugoslavia." So, the mediators made use of a referentially ambiguous adjective "equivalent", which is not the same as equal, but could be. As to the interim character of the agreement, the mediators used both referential and cross-textual ambiguities to meet the demands of both delegations. First, the draft agreement itself was called "Interim-agreement", to the liking of the Albanian delegation. However, Chapter 8, Article I, 1-3, stipulated that amendments to this agreement should be adopted by agreement of all parties. That meant that without Serb consent the interim agreement could not be changed; and that thereby it would turn into a permanent arrangement. However, in Article I, 3, mediators emphasised that "three years after the entry into force of this Agreement, an international meeting shall be convened to determine a mechanism for a final settlement for Kosovo" which seemed to tilt the balance of the wording again in favor of the Albanian demands. Austrian Ambassador Petritsch, one of the three mediators, did not hide the fact that this paragraph was left open to two contradictory interpretations. At his March 8, 1999, meeting with Serbian President Milutinovic, he clearly stated that "die Interpretierbarkeit einiger wichtiger Textstellen bewusst gewaehlt worden war, um beiden einander ausschliessenden Positionen gerecht zu werden."11

To remind readers, the Serb delegation did not accept the Rambouillet draft agreement, whereas the Albanian delegation accepted it in such an ambiguous fashion that its acceptance was just a bit better than the Serb refusal. Those who dislike ambiguous provisions and ambiguities as a peace-making mechanism would have found this a predictable failure. There is something in ambiguous language that its opponents find difficult to digest and upon which their dislike of ambiguity rests. They say that ambiguities actually do nothing but "paper over" important differences between the parties to an agreement. Ambiguities are a kind of Machiavellian manipulative device that brings but temporary satisfaction to the parties as it deceptively, but not really, meets their demands in full. Such satisfaction is deceptive because both parties have the right to interpret ambiguities in their own irreconcilable ways and that is a right they will certainly, sooner or later, start exploiting. That is also why ambiguous agreements may quickly lead to arguments, and turn into disagreements, as, precisely due to ambiguities, conflicts in interpretation will necessarily break out. Speaking metaphorically, the inclusion of ambiguous expressions in a peace agreement is comparable to reopening a repository of arms to the parties and inviting them to rearm themselves with a kind of intellectual weaponry. This intellectual weaponry, i.e. ambiguity, is highly likely to be employed by both parties as soon as the first step in implementation is taken. For that reason implementation of an ambiguous agreement is very likely to fail.

A critic of the use of ambiguities in peace agreements would also add that their use is an interruption of communication between the parties to a conflict that usually precedes an outbreak of armed conflict between the parties. The parties' inclination to interpret some of the key expressions of their political relationship in a radically different fashion is one of the main causes of such interruption of communication. For instance, interpretative disagreements over the meaning of the key terms of their political relationship preceded the war between Croatian authorities and the Croatian Serb minority, supported by Serbia proper, and probably contributed to its eruption in 1991. For Croatian authorities, for instance, the term "Federal Republic of Yugoslavia" meant an association of free and basically independent republics, whereas for the Serb minority the term meant much more than that. The same applies to ambiguous peace agreements. Just as, prior to an outbreak of war, the crucial terms of political vocabulary become ambiguous and generate misunderstandings and disagreements that then lead to war, an ambiguous peace agreement will itself generate new misunderstandings and add more heat to the parties' already hostile feelings. Ambiguities thus, in all probability, prompt the parties to set in motion a new spiral of physical violence, or at the least put between them a barrier to firmly keep their positions far apart. Under the most optimistic expectations ambiguous "peace" agreements unleash a psychological war of words and perceptions.


Diplomats, negotiators, mediators and decision-makers seem to be blind and deaf to the aforementioned worries. They rely on ambiguities quite often and do so with no concern about the negative effects that, according to the aforementioned critique of ambiguities, an ambiguous peace agreement is likely to produce.

Here is a list of famous texts that contain ambiguous provisions in their key places, with an obvious intent to bridge the differences between major interests of certain parties through an ambiguous formula.

Athenian Constitution

In the sixth century B.C. the Greek poet and statesman Solon wrote a constitution for Athens that was considered a revolutionary turn in organisation of both Athenian social and political life. As Aristotle explains in his book "Athenian Constitution", Solon provided a framework for the resolution of the inter-group conflicts inherent in sixth century Athenian society, leaving an important part of his constitution deliberately ambiguous; open to free interpretation.12 As Aristotle says, some have interpreted Solon's strange decision to do this as implying that Solon primarily wanted to extend the powers of Athenian courts and by implication to strengthen the political position of the "demos"; of the common people of the middle strata of Athenian society. Namely, Solon had opened the possibility for the "demos" to play a larger role at the Athenian courts than they played before. Unfortunately, today it is very difficult to reconstruct the precise differences in interpretation of the ambiguous parts of the constitution by the different strata of Athenian society of that time. Nonetheless, it is clear that there was a conflict of interests and that Solon intended to strike a balance between those interests by including provisions in his constitution that were open to several equally valid interpretations.

Homoousia and the First Nicean Council

The second example comes from the First Nicean Ecumenical Council that concluded with the adoption of the Nicean Creed.13 Deliberations of the Council aimed first at refutation of the Aryan heresy and then at bridging the gaps between various interpretations of the Holy Scripture, and thus at reconciling the differences between various Christian schools and doctrines. So, in a way, the First Nicean Council was an arena of negotiations aimed at producing a mechanism for resolution of an intellectual conflict. One of the most controversial issues that the Council discussed was the issue of interpretation of the relationship between Jesus Christ and his Father, God. The Aryan heresy taught that their relation was one of heterousia, of difference in substance. The Council decided by an intellectual fiat that their relationship was in fact one of homoousia. Now, interpretation of this Greek formula is not an easy matter, since it can be understood in two different ways. Homoousia can mean "identity of substance", but it also means "similarity of substance". The fathers of the Nicean Council deliberately avoided the stronger version of synousia, which means full and complete identity of substance. So, they actually selected a weaker version of the word describing the relationship between the Father and his Son Jesus, to leave the door of their creed open to those who had certain difficulties with the belief in total identity between the divine and the human substance.

W. Wilson's "14 Points"

In 1918 US President Woodrow Wilson presented to the US Congress an American list of basic principles for an international settlement to end World War I. Walter Lippmann delivered an authoritative study of this document in his book Public Opinion and I will in the following mainly adhere to Lippmann's views.14

Lippmann said that the "14 Points" were subject to an infinite number of different interpretations from all corners of the globe. Each group saw in the document what it wanted to see and that is why for the majority of the states involved in World War I the document seemed to be perfectly acceptable. This is also the reason the document succeeded in pulling together a majority of states and channeling their efforts in the direction of a search for terms for a comprehensive peace settlement. It is interesting to note that the US refrained from explaining to its allies its own interpretation of the document until shortly before the allies accepted the truce.

For instance, point 4 on "arms reduction to the levels consistent with domestic safety" is clearly ambiguous, as is point 10 on "autonomy for the peoples of Austria-Hungary". Lippman pays special attention to the ambiguities of Wilson's point 8, which speaks about the issues that plagued the then Franco-German relations. This point stipulates that "all French territory should be freed and the invaded portions restored, and the wrong done to France by Prussia in 1871 in the matter of Alsace-Lorraine, which has unsettled the peace of the world for nearly fifty years, should be righted." As Lippmann sharply notes, Wilson did not say that the Alsace-Lorraine region would be simply returned to France. He did not say that because at that time it was unclear whether the French would have continued to fight had they been offered a plebiscite in this matter. The purpose of the sensitive wording of this point however was to leave the possibility of such an interpretation open. Besides, Wilson has loaded this point with another meaning. Approximately at the same time France had made a secret agreement with Russia to demand that Germany return Alsace-Lorraine under the wider 1814 concept of the region. Wilson happened to know about this secret arrangement, and he did not agree with it. In other words, Wilson left the methods of exact implementation of the "rectification" provision to be worked out in more detail later in the process, leaving enough elbowroom for French interpretation of the methods as well as for its secret ambitions. However, he also clearly pointed out that rectification should aim at the injustice done to France in 1871 and not in 1814. He thus opened quite enough room for his own reading to eventually oppose the wider French interpretation of "Alsace-Lorraine" following from France's secret agreement with Russia.

Yalta Declaration

The Yalta Declaration contains a number of ambiguous formulas that served to bridge the gap between the positions of the US, Britain, and the Soviet Union vis-à-vis the post-World War II global order.15 Its chapter dealing with the Polish question is notorious for ambiguities that Stalin later put to his advantage, for which he was strongly criticised by Churchill, who accused him of violating the Yalta spirit. However, all Stalin later did in regard to Poland was simply to implement the Soviet interpretation of the ambiguously worded Yalta Declaration chapter on Poland. While the British and American interest was to see a free Poland with a democratically elected government, truly representative of the will of its people, Stalin wanted to see an obedient and controllable Poland, which could not be used by another great power as a corridor against the Soviet Union. As the two interests were obviously incompatible, the Big Three were unable to agree on anything but an ambiguous text which allowed for a number of interpretations. For instance, as to the issue of reorganisation of the then Polish "Lublin" government, which was Moscow's puppet, the Three agreed that it would be reorganised "on a broader democratic basis", and include both democratic leaders from Poland itself and from Poles abroad. The text did not literally state that the Polish leaders from abroad should come from the Polish government in exile in London, but Churchill's reading of the text probably assumed that would be the case. As to the holding of free elections in Poland, the Yalta declaration stated that the Polish Provisional Government of National Unity "shall be pledged to the holding of free and unfettered elections as soon as possible". It seems that Roosevelt and Churchill missed one important thing: the text assigned the responsibility for the holding of free elections not to any of the Big Three, but to a yet to be established Provisional Government. This left enough room for Stalin to evade this provision without having to take any blame. Finally, the Poland chapter stipulated that in "these elections all democratic and anti-Nazi parties shall have the right to take part". "Democratic and anti-Nazi parties" is a term vague enough to generate considerable argument over the parties eligible to run in elections. Ambassador Harriman later emphasised that President Roosevelt never bothered too much about wording of provisions, though he, Harriman, often drew the president's attention to loopholes, or vagueness, in various provisions of the Yalta package. Harriman explained this by saying that Roosevelt did not care about other people's interpretations as long as he was able to provide his own. This may be a sound attitude, but what if the authoritative "interpreter" has an untimely death, as with President Roosevelt?

UN SC Resolution 242

After the crushing defeat that Israel inflicted on joint Arab forces during the Six Day War in 1967, the UN Security Council agreed on the text of the famous resolution 242. What should be emphasised is that the resolution was a result of bargaining between the powers sitting in the Security Council and that it reflected the deeply polarised political opinion at the United Nations in the period following the war.16

The provision of the resolution which prompted different and incompatible interpretations was the one immediately following the preamble of the text, reading: "establishment of just and lasting peace in the Middle East should include the application of both the following principles:

  • withdrawal of Israeli armed forces from territories occupied in recent conflict;
  • termination of all claims or states of belligerency and respect for...territorial integrity...of every State in the area and their right to live in peace within secure and recognized boundaries."

This resolution uses a somewhat strange English construction "territories occupied in recent conflict", from which the definite article "the" has been omitted. That is why it was possible to raise the question as to whether Israel was actually asked to withdraw from all the territories occupied in the recent conflict, or to withdraw from some, but not all, territories. Another controversy that followed this resolution was due to the French translation of the document, which unlike the English original used the definite article: "Retrait...des territoires occupes lors du recent conflit". So the French version, which together with the English version is an official UN version of the document, suggested that Israel must withdraw from exactly those territories that it occupied during the Six Day War. It is clear that such an interpretation was in harmony with the demands of Arab countries and they did their best to prove its validity. Israel, of course, opposed such an interpretation and it seems that the sponsor of the resolution, Lord Caradon, had no intention of inserting the definite article either. Caradon also emphasised that the fact that the second part of the first provision sheds additional and clarifying light on the first part must be given uppermost consideration. Actually, the second part generates what I called a "cross-textual ambiguity", as it says that all states have the right to live within secure and recognised boundaries, which the boundary before the Six Day War was not, according to Lord Caradon. By implication Israel did not have to withdraw to its pre-Six Day War borders. It seems that the second part of the first provision opened even more room for Israel to interpret UN SC Resolution 242 to its own advantage.

6-Point Agreement

After the Yom Kippur War between Israel and Egypt the very first problem both negotiators and parties to the war encountered was the issue of the status of the Egyptian Third Army, surrounded by the Israeli Defense Force on the eastern side of Suez. After the first phase of peace talks held in October, almost no progress was achieved. The talks continued in Washington in November, while in parallel the UN Security Council issued Resolution 340, which demanded that the Israeli Force withdraw to the lines occupied on October 22, 1973, at 1650 GMT. In that way encirclement of the Egyptian Third Army would end. However, Israel refused to comply with the UN SC Resolution 340. Thereafter negotiations fortunately continued and resulted in an agreement called the "6-Point Agreement", signed on November 11 at Kilometer 101 of the Cairo-Suez road. One of the chief mediators to the agreement was Henry Kissinger, US Secretary of State, who during the negotiations frequently used the term "constructive ambiguity" to explain his negotiating strategy as well as the key purpose of the 6-Point Agreement. For instance, provision B of the agreement contains an ambiguity which served the purpose of making it easier for Israeli negotiators to engage in further talks leading to Israel's compliance with SC Resolution 340. The ambiguous provision was thus the only possible way to ensure that talks continue and arrive at a solution that would save the face of one of the two parties, which would not have been possible if the solution had been arrived at by a quick jump or a fiat.

Provision B says: "Both sides agree that discussions between them will begin immediately to settle the question of the return to the October 22 positions in the framework of agreement on the disengagement and separation of forces under the auspices of the UN." This provision is a shining example of a "syntactical ambiguity" that Egyptian and Israeli negotiators could interpret in diametrically opposite ways, depending on which syntactical links between the parts of the provision they saw fit. Egyptian negotiators interpreted this provision as a clear demand that Israel withdraw its armed force in accordance with UN SC Resolution 340. The syntactical link they saw fit was the one between "return to the October 22 positions" and "under the auspices of the UN". Israeli negotiators however understood the provision simply as calling on the parties to negotiate a "separation of forces" agreement without any specific request to return to the October 22 lines. The syntactical link they saw fit was the one between " settle the question" and "under the auspices of the UN".

Note that the 6-Point Agreement is the first agreement that Israel signed with an Arab country after the outbreak of hostilities between Arabs and Israelis in 1949.17

Shanghai Communiqué

During Nixon's February 1972 visit to China, President Nixon and Chairman Mao adopted the Shanghai Communiqué, 2/3 of which consists of unilateral expressions of each country's specific views of international relations, and the remaining 1/3 of which consists of a number of jointly accepted provisions.18 Within those joint declarations, the part implicitly addressing the issue of the Soviet Union is ambiguously worded. First, the Communiqué says: "each (referring to the US and the PRC) is opposed to efforts by any other establish such hegemony (meaning "a hegemony in the Asia-Pacific region")." It further says "neither is enter into agreements or understandings with the other directed at other states." Here we have a piece of cross-textual ambiguity as the first sentence says that both China and the US agree to resist possible Soviet attempts to establish a hegemony in the region. The pronoun "each", however, expresses this in an ambiguous way, as if China and the US have their own individual views that accidentally coincided. The second sentence says that neither the US nor China are willing to agree on an action directed at other states, including certainly the Soviet Union, which implies that they did not agree on any concrete preventive measure to counter possible Soviet attempts at establishing hegemony. In other words, the first sentence conveys a soft kind of threat to the Soviet Union, while the second sentence weakens the threat by dismissing the possibility of joint US-China action directed at other states. In this way, the Shanghai Communiqué delivered an ambiguous threat to the Soviet Union, a threat in a sort of embryonic form. Both China and the US are likely to have adopted this kind of language to leave enough diplomatic room for their own unilateral build-up or improvement of relations with the Soviet Union.

The Shanghai Communiqué contains an ambiguous provision in its unilateral parts as well. The US inserted into the document the following sentence: "The United States acknowledges that all Chinese on either side of the Taiwan Strait maintain there is but one China and that Taiwan is a province of China." This sentence has been interpreted as the very first expression of American support for the "one-China" policy; for a reintegration of Taiwan into the PRC, although that is not how it should be interpreted in its original context. Actually, the inherent ambiguity of the term "one" was fully exploited. The fact that the PRC and Taiwan agree that there is "but one China" does not imply that they agreed on internal arrangements for the "one China". Actually they disagreed on this. From the very opening of the Taiwan issue until well into modern times, China and the US have been unable to find a formula to ensure that the reintegration of Taiwan takes place. The term "one China", however, helped the US itself express a proper balance between its relations with China and its relations with Taiwan without jeopardising either. The term was a clever tactic that the US employed to both maintain its policy of protecting Taiwan, on the one hand, and to promote its new policy of opening to China, on the other. It was probably the only way for the US to symbolically gratify both its own and China's interests in relation to Taiwan.

Oslo Peace Accords

The Oslo Peace Accords that Israeli and PLO representatives signed in Oslo in 1993 contain many ambiguities as well. Instead of enumerating them in detail, I will simply quote a part of a Financial Times interview with Madeleine Albright and suggest that readers try to identify the ambiguous expressions in the accords themselves. "`Instead of glue it's been sandpaper', she (Madeleine Albright) says of the 1993 Oslo peace accords, a deliberately ambiguous framework for a partnership between Israel and Palestinians."19

Dayton Peace Accords

The Dayton Peace Accords (DPA) ended the war in Bosnia and Herzegovina, redrawing and decentralising its internal structure20.

The Dayton Constitution of BiH (Annex 4) is the annex of the DPA in which one finds the most interesting examples of ambiguities. For instance, in article III, provision 1, the Constitution clearly defines responsibilities of BiH institutions by the method of enumeration (foreign policy, foreign trade policy, customs policy, monetary policy, finances of the institutions of BiH, immigration, refugee and asylum policy and regulation, international and inter-entity criminal law enforcement, establishment and operation of common and international communication facilities). All responsibilities not expressly assigned to the central institutions were placed on the entity level. But article III, provision 5.a, introduced a cross-textual ambiguity into this annex. It reads that "Bosnia and Herzegovina shall assume responsibility for such other matters as are agreed by the entities...or are necessary to preserve the sovereignty, territorial integrity, political independence, and international personality of BiH, in accordance with the division of responsibilities between the institutions of BiH."

Provisions 1 and 5.a together result in an ambiguous text; a cross-textual ambiguity which may be interpreted in two different ways: first, as providing entities with powers that cannot be delegated to the state level unless entities expressly agree to it; and, second, as providing the state with vaguely defined powers that have not been mentioned in the list from article III, provision 1, and for which no express consent by the entities is needed. In other words, while article III, provision 1, seems to reduce the powers of central authorities, provision 5.a seems to open room for a reverse procedure, for the extension of powers in an admittedly less determinate way.

"Good Friday" Agreement

The "Good Friday" Agreement adopted on April 10, 1998, set an institutional framework for resolution of the political conflict in Northern Ireland.21 Prior to the agreement Northern Ireland was torn between two contradictory political ambitions. Unionists favoured maintenance of links between Northern Ireland and the rest of the United Kingdom whereas the Irish nationalists favoured integration of Northern Ireland into the Republic of Ireland. A relative compromise that negotiators achieved on Good Friday 1998 brought partial satisfaction to both political ambitions. The "Good Friday" compromise rests on three key institutions: 1. a Northern Ireland Assembly with an executive composed of "up to 12 members"; 2. a North-South Ministerial Council: an all-Irish body which confirms the link between Northern Ireland and the Republic of Ireland; and finally, 3. a British-Irish Council, including Northern Ireland, the Republic of Ireland, as well as Britain.

The Good Friday Agreement left certain relations vague, which was the main factor that allowed the parties to put incompatible constructions on the deal. For instance, it is clear that nationalists would welcome an extremely strong, or, even better, a dominant North-South Ministerial Council. On the other hand, unionists would prefer that it be washed off the earth altogether. That is why the Good Friday Agreement left the details of the Council deliberately underdetermined, i.e. ambiguous, and both parties were given enough room to project their own interpretation of the details into the ambiguously worded parts of the deal. Now, the details of the North-South Council are described in "Strand Two" of the agreement, while key ambiguities in the Strand can be identified in provisions 8 and 9, including an annex to the Strand.

Namely, the Good Friday Agreement defines two channels through which the North and the South may cooperate. First, through establishment of "implementation bodies" under direct jurisdiction of the North-South Ministerial Council, and, second, through assignment of "areas of cooperation" in which the North and the South would cooperate through existing separate bodies under their separate jurisdictions, rather than through newly established "cross-border" institutions. The agreement furthermore says that there should be at least twelve "subject-areas" of cooperation between North and South; six in each category, meaning six to be covered by "implementation bodies" and six others to be treated as "areas of cooperation". The agreement also provides an annex containing a list of items that may be included in the twelve "subject-areas". It is thus quite clear that the Good Friday Agreement left open the possibility of more than twelve "subject areas" and more than six items in each of the two categories. Furthermore, the list from the annex could be interpreted as a compulsory list, but also as a noncompulsory one, depending on the way one reads the meaning of the auxiliary verb "may" in this context.

Conflicting interpretations of these provisions placed a considerable burden on the negotiation that in December 1998 resulted in a new and precise agreement on "implementation bodies".22 Starting from the same Good Friday framework, unionists demanded that the number of bodies be limited to the original six, while nationalists logically demanded that there be at least a few more. On December 18, 1998, on the basis of the aforementioned provisions of Strand Two, the new agreement determined that there would be six implementation bodies to cover the matters of language, trade, EU programs, food-safety, aquaculture and marine, and inland waterways. But one of the most interesting details of the December 1998 deal was the fact that while nationalists gave up their interpretation of the Strand Two provisions 8 and 9, unionists gave up their interpretation of the Strand One provision 14. In that way unionists conceded to the nationalist demand that there be ten ministerial departments in the Northern Ireland executive. This was a perfect case of a balanced trade of interpretations in which unionists let nationalists decide on interpretation of the ambiguous provision 14 of Strand One, while nationalists in turn let unionists decide on interpretation of ambiguous provisions 8 and 9 of Strand Two.

Hence it was through a fair trade that the parties to the Northern Ireland peace process cleared up several key ambiguities of the Good Friday Agreement. However, one must not forget that original ambiguities in the Good Friday framework actually made it easier for both parties to embrace the deal, launch the peace process and, as some have put it, take the gun out of Irish politics.23


Thomas Franck, one of today's leading theoreticians of international law, subscribes to the aforementioned argument against the use of ambiguities in peace agreements. Both in his book Power of Legitimacy among Nations and in his Fairness in International Law and Institutions Franck proposes a critique of ambiguities the substance of which amounts to the basic argument I mentioned in the second part of this paper.24

Franck stands for the idea of transparent, clear, and determinate meaning of the key norms, rules and provisions in any kind of written, legally binding agreement. He writes: "Textual determinacy is the ability of a text to convey a clear message, to appear transparent in the sense that one can see through the language of a law to its essential meaning."25 In his opinion, ambiguities are harmful because it is hard "to know what conformity is expected" in the condition when a treaty or an agreement contains ambiguous provisions.26 In Franck's opinion, this lack of knowledge of what conformity consists of automatically leads to non-compliance with the ambiguous agreement as it is nearly impossible to justify compliance if parties to an agreement lack a clear idea about the requirements of conformity. Therefore implementation of an indeterminate or ambiguous normative standard is very likely to take the wrong direction. In other words, Franck says that an ambiguous treaty is unable to pull the parties to the treaty towards compliance or towards implementation of that treaty in good faith.

Franck has proposed another argument against the use of vague terms, provisions or norms, in peace agreements. He says that vague norms cannot be considered fair. Fairness of a norm implies that such a norm is understandable to those expected to abide by it. If those who should abide by a vague norm do not fully understand its consequences, including the intricate process of its legal interpretation, then one could not characterise such a norm as fair. That is why semantic and linguistic determinacy can make a norm fair in the eyes of those who should comply with it, while linguistic ambiguities make a norm opaque, difficult to understand and, ultimately, unfair in the eyes of those expected to comply with it.27

There are basically two ways to check the validity of Franck's position vis-à-vis the use of ambiguities in peace agreements. First, we could test it empirically. Such a test would run approximately as follows. One would put on one side all agreements in which ambiguities were used in key provisions. Then one should simply determine the relative percentage of those ambiguous agreements whose implementation failed. We put that target-category on the other side. This however would not suffice. One should also pinpoint those ambiguous agreements whose implementation failed for the exact reason of their ambiguousness. This means that one should form another sub-category including all those ambiguous agreements that failed due to insurmountable obstacles that the successive process of interpretation placed on their implementation. If the percentage of ambiguous agreements that failed due to a failure of interpretation exceeds the percentage of ambiguous agreements whose interpretation delivered satisfactory results and thus helped their implementation, then Franck's position would be acceptable. That would certainly mean that ambiguously worded agreements are likely to generate incurable troubles during their implementation. In other words, select the category of "ambiguous agreements". Then among those select the subcategory of "ambiguous agreements that failed". And then, finally, among those select both sub-subcategories of "ambiguous agreements that failed due to interpretation failure" and of "ambiguous agreements that failed for reasons other than interpretation failure". If the number of members of the former category exceeds the number of members of the latter category, then Franck would be right for purely empirical reasons.

This is how one would envisage a method to check the merits of Franck's position, but I do not believe that such a method is applicable. Even if one could categorise all the ambiguous agreements the implementation of which failed due to insurmountable difficulties in interpretation, the process of interpretation itself might fail for hundreds of reasons. Furthermore, what we usually see in reality are not pure instances of an agreement which failed, or an agreement which was easily, consistently and completely implemented. The majority of cases lay somewhere in between. How would one, for instance, categorise Wilson's "14 Points"? They performed an extremely important function, pulling together many nations and providing them with a source of motivation for a search for peace terms. However, they have not been implemented in full. The 14 Points was thus a semi-successful kind of ambiguous agreement. What can we say about the implementation of Dayton? I believe that the only convenient way to describe the obstacles to the process of its implementation should not include its ambiguous nature, because, properly speaking, those supposed to comply with the Dayton Agreement have not yet recognised its ambiguousness, let alone embraced it. They instead tend to abuse the agreement's ambiguous provisions, to insist on their one-sided interpretation to justify promotion of their outdated policies that brought misery and suffering to the peoples of Bosnia in the recent past.

What I mean to say is that it is nearly impossible to distinguish between pure categories of "ambiguous agreements whose implementation failed because of their ambiguous nature" and "ambiguous agreements both successfully interpreted and implemented despite their ambiguous nature". Additionally, there is always the factor of will and readiness for a compromise. I believe that if and when an ambiguous agreement fails due to an interpretation failure, this may be better explained by the parties' weak desire to strike a compromise; to arrive at a third, mutually acceptable interpretation, than with the inherent ambiguousness of the agreement they originally adopted. When there is a will to both take and give, and to do it in a fair manner, then the process of interpretation is unlikely to fail. The December 1998 negotiation between the Irish unionists and nationalists, who both started from an ambiguous scratch but succeeded to cut a fair and more precise deal, provides the best positive evidence in that regard.

So, that is why the method of empirical testing of Franck's argument is not an absolutely reliable method to discuss and resolve the issue of the use of ambiguities in peace agreements. There is another, purely deductive, method to help one debate this issue. I will present more elaborated views along the lines of this method in the text that follows. I will start with a discussion of Franck's cons in the format of a deductive argument, in which he framed his own view.

The reader should remember that Franck's first argument says that ambiguousness leads to an easy justification of non-compliance and that is why one should use only precise and transparent language in legal documents, including peace agreements. This argument, as it stands, is not conclusive because non-compliance could be justified only if it is absolutely unclear what the provision with which one party wants to justify non-compliance means. However, ambiguous provisions generally mean two things, and in principle carry two meanings: one, which one party finds more convenient, and another, which the other party finds more convenient. That is why it is impossible to claim that ambiguous provisions carry an absolutely unclear meaning.

When it comes to ambiguously worded provisions, if party A opts for implementing an ambiguous provision X in the way they find more suitable (Xa), then this does not imply their non-compliance with the provision. The party actually complies with the agreement, but does so under the interpretation they deem logical. If, however, the party opts for implementing provision X in the way that the other party finds more suitable (Xb), then this does not imply non-compliance with the provision either, as they are complying with the provision in the sense that the other party prefers. Finally, if A opts to refrain from implementation until the process of interpretation has been complete, then this would certainly represent the most rational and fair strategy, and should not be considered a case of non-compliance either. This is why the situation which Franck envisaged and in which parties allegedly avoid compliance and justify their non-compliance with reference to the vagueness of a provision of the agreement they adopted, is a logical impossibility. This is why Franck's first argument does not apply.

Franck's second argument says that a party must understand a norm fully in order for the norm to be characterised as fair. Otherwise, and in the case when a norm is vaguely worded, one could not characterise it as fair. This argument itself suffers from several weak spots. First, Franck seems to forget that we, in general, understand very well the consequences of ambiguous provisions as well as their possible readings. Those adopting such provisions are, as a rule, fully aware of all legal means, arenas and authorities through and by which it is possible to defend one's own interpretation. That is why again this situation, which Franck counts on, applies rarely if ever.

Secondly, since an ambiguous provision X is open to basically two incompatible readings, a consequence which party A would possibly deem unfair is one in which implementation of provision X would run along the reading Xb, not Xa. But what if the implementation of another ambiguous provision Y runs along the reading Ya and not Yb, compensating for party A's feeling of injustice and restoring the balance between parties A and B? This situation thus strongly suggests that a possible sense of unfairness has nothing to do with an ambiguous provision taken per se. A proper description of the first situation would say that party A considers implementation of X unfair because the ambiguousness of X is replaced with a strict and disambiguated meaning favorable to the other party's interests. The feeling of unfairness thus has something to do with the disambiguation of ambiguities, or otherwise with a disparity, or unfairness, in the process of parallel interpretation of several ambiguous provisions of an agreement. It has nothing to do with ambiguities taken per se.

That is why both Franck's arguments against the use of ambiguities in peace agreements are inconclusive. Franck perhaps had in mind norms opaque to such an extent that one could not discern any logical or possible meaning in them. It is only then that he could claim a certain validity for his argument. In real life, however, we are unlikely to meet norms of this kind. Both parties and mediators to an agreement are usually aware of the main interpretations opened by a relatively vague provision as much as they are aware of their consequences. They are also well aware of the conduct of the parties that would stand in harmony with those interpretations, which implies that they are also aware of the fact that it would not be possible to justify non-compliance merely by referring to allegedly total opacity of an agreement's terms.

We may now conclude that Franck's arguments against the use of ambiguities in legal norms and documents, including peace treaties and draft agreements, do not hold.

Now I will turn to the general con against the use of ambiguities that I formulated in the section on basic rationale and dislike, and explore it in more detail. As the argument says, an ambiguous agreement is very likely to generate severe intellectual conflict between its parties, as each party will interpret the agreement to their own benefit, contradicting the interpretation by the other party. That is why one should resist temptation to "paper over" differences between parties by drafting an ambiguous formula. This argument against the use of ambiguities is valid, but only under one additional condition. Namely, ambiguous agreements are likely to generate hermeneutic conflict if and only if their parties insist on their own, unilateral interpretation of an ambiguous provision and do not recognise ambiguity qua ambiguity. If they recognise an ambiguous provision for what it actually is, a sentence or a text open to several incompatible interpretations, the argument over interpretations would in all likelihood give way to the relationship of a joint cooperative effort in the search for a third impartial reading of the provision. Recognition of ambiguity qua ambiguity would imply recognition of the other party's right to interpret it in its own way, and would thereby automatically reduce potential for severe hermeneutic conflict. This conditioning of the above general argument against the use of ambiguities provides two extremely important lessons. First, the general argument as such is not valid, because it is only under the condition of the parties' unilateral insistence on their partial interpretation of ambiguities that such an argument, only imperfectly, applies. In such a case the argument against ambiguities targets only those ambiguities that are not recognised as such. This is why the general con, the reason for a basic dislike of ambiguous peace agreements, does not at all represent a conclusive argument against the use of ambiguities in peace agreements, and all those keen to represent it as such are deeply mistaken. However, one must admit that there is a risk in the use of ambiguities. Parties may initiate an argument over interpretation, which then may cause a serious rupture in their relations. It is true that such an argument would not occur without the use of ambiguities, but it is also true that the risks they entail lend very limited and weak support to the general conclusion that we should not use them at all. In other words, if a usage of a form entails possible risks, but at the same time opens a more promising path, then the fact that there may be a risk in the usage cannot not be taken as a springboard to bury the form itself tout a fait. It is a pure case of non sequitur. True, there is a risk that the use of ambiguities may lead parties down the wrong path. But there is also the possibility that the parties to an ambiguous agreement may engage in a joint search for a third, creative and constructive interpretation, which would reconcile their contradictory readings of the peace provision. Both possibilities are equally open and they both spring from the same language.

The negative attitude towards ambiguous agreements may have its source in particular historical experiences. If one focuses on the Yalta Declaration, for instance, or if a Palestinian diplomat focuses on UN SC Resolution 242, he or she will probably emerge with a very negative view of ambiguous agreements in general. If one focuses on the Good Friday Agreement, however, then he or she will probably take a generally positive view of ambiguous agreements. Whichever position one takes, it must not be based solely on a particular historical case, but on a rational and argumentative debate of pros and cons. To demonstrate how such a debate ought to look, I will present a few more pros and discuss them in more detail.

First, if an ambiguity makes it easier for negotiating parties to accept an agreement and therewith put a close to a war, or to a situation of increased friction or hostility, this should be taken as an argument supporting the use of ambiguities. Even if an ambiguous provision may later generate a conflict in opinion, the fact that the relationship of physical hostility gave way to the relationship of merely verbal conflict must be taken as a sign of progress. In terms of empirical evidence, this argument rests on cases such as Wilson's 14-Points, which established a broad consensus to end World War I by defining a number of points on which the majority of states vaguely agreed. This also applies to both the Dayton Peace Accords and the Good Friday Agreement. However, one must not forget that there are inherent risks in the use of ambiguous wording and that a number of additional criteria must be met before the parties can take full advantage of an ambiguous provision. Such criteria include the parties' readiness to accept tradeoffs in interpretation, to make further concessions and to engage in a common search for a third interpretation.

Second, one could compare the use of ambiguities with the practice of reservation in international treaties, and say that both somehow depend on the imperfect nature of international actors. Both the practice of reservation and the use of ambiguities rest on the purely pragmatic idea of using whatever means are available to ensure that a text of a treaty is accepted. This, again, is not a perfect pro for the use of ambiguities, because one could perhaps rightly argue that neither reservation nor ambiguity contributes to making truly significant progress in inter-state relations. Let us take Yalta as an example. There was no progress after Yalta, a critic of ambiguities would say, because the Big Three adopted an ambiguous document and one of them, Stalin, decided to unilaterally implement his own reading of the text. However, this does not accurately reflect the historic record of post-Yalta tensions in international relations. In addition to ambiguous language, many other factors played an active role in the emergence of the cold war. They include the death of the chief interpreter of the document from the US side, a newly-emerging disagreement over the issue of German reparations, and an obvious lack of will on the American part to take more tangible action as a response to Stalin's treatment of the Polish question.

Third, I would say that ambiguous provisions perform another useful function. They make the conflict of interpretation predictable. In other words, start from the premise that the parties to an agreement will continue fighting politically even after they sign a treaty. However, this process of political fight will be more channelled, more orderly and predictable if one knows in advance which provisions of the jointly adopted text will give rise to a conflict in opinion or interpretation. I am ready to admit that this does not imply that the very process of "post-conflict" conflict will be absolutely predictable, because the post-conflict relations between erstwhile enemies do not depend solely on the text of their agreement. However, to the extent that they do depend on the text, they, paradoxically, will be more predictable, and better channelled, when the text is ambiguous than when it is not. Again, reliance on the text may be minimal by both parties, if they took the text as a fig leaf to provide them with a respite; a cease-fire, after which they intend to gather forces and resume fighting. This, however, has nothing to do with the use of ambiguities itself. In other words, I believe that the post-conflict situation is not a situation of perfect peace and friendship; that a traumatic energy remains to aggravate relations between the signatories. Ambiguous provisions may be thought of as channels to direct this energy and to allow it to be acted out in a more or less predictable fashion. That is why their use may serve an outstandingly important function. That is also why the oft-repeated argument by Palestinian representatives against the Israeli insistence on vagueness and ambiguousness of treaties they sign, which the former view as a mere nitpicking, may be too short-sighted. The fight over interpretation is infinitely better than physical violence and, combined with other, more popular methods of relief such as punishment of war criminals, or truth-finding, may considerably help relieve the burden of traumatic energies accumulated through the past violence.

This leads me to the fourth and final argument in partial support of the use of ambiguous language in peace agreements. Speaking strictly, there is no such a thing as a peace agreement, which, as it is usually defined, resolves a conflict and turns hostile relations between former adversaries into a straightforward relationship of peace, cooperation and understanding. Instead there are shades of grey and the process of conflict transformation slowly takes root. The erstwhile enemies gradually learn to cope with their differences, and how to prevent them from jeopardising the areas of overlapping interests, including interest in peace. For this reason verbal conflict over interpretation of an ambiguous provision may teach erstwhile enemies two important lessons. First, verbal conflict, a free expression of one's own interest, which contradicts somebody else's interest, is no wrong as long as it takes place in a polite manner and with due regard to the codes of civility. Second, the best way to move away from the state of war is through a slow accumulation of pros and cons, in the form of logical, well-founded arguments, aiming at a third reading, a third interpretation, to which an ambiguous provision may have, however indirectly, already pointed. "Peace is what we have when creative conflict transformation can take place non-violently":28 such a concept of peace leaves more than enough room for the concept of constructive ambiguity.


I will draw a number of implications from the arguments and examples that I thus far provided. First, there are fine differences between types of ambiguous expressions which have direct bearing on the issue of their interpretation. For purely pragmatic reasons, referential ambiguities do not seem to cause too many troubles in the process of interpretation. The parties read two different interpretations into a word or a phrase, but all they need to do to resolve such an ambiguity is to couple and cross-fertilize their readings. The language they originally adopted is the same for both. Cross-textual ambiguity may be much more difficult to tame. Each party identifies with a part of the text and tends to dismiss the other parts, i.e. textual components of a cross-textual ambiguity, as an aberration from the original intention of the text. Besides, it is truly intrinsically more difficult to clear up an ambiguous text than it is to clear up an ambiguous word or phrase. That is why it was probably easier for the parties to the Good Friday Agreement to disambiguate their text than it is for the parties to the Dayton Peace Accords. Furthermore, the ambiguities in the former treaty were presented in a quantified form; they have the shape of numerical values that are relatively easy to manage or to mentally manipulate. So, the first conclusion from this brief reflection reads that a) one ought to prefer the use of referential over syntactical, and the use of either over the cross-textual kind of ambiguity; and b) one ought to quantify ambiguities as much as possible. For instance, do not speak only about the competence of certain layers of government, but try as well to express them in a numerical, metrical fashion; try to define an approximate number of bodies to represent certain interests, for example.

My second reflection is epistemological. The subject of "peace-making" ambiguities offers an extremely fertile intersection for two disciplines: linguistics and the theory of diplomacy. I believe there is a lot of opportunity here for mutual borrowing and mutual inspiration. For instance, the Vienna Law of Treaties contains an explicit chapter on methods of interpretation. Unfortunately, the means of interpretation it proposes are, in terms of linguistics, poor, vague and completely outdated. No serious linguist would accept that the process of disambiguation runs exactly along the four methods this highly important convention proposes: travaux preparatoires, inter-related texts, true intentions, and common sense-interpretation. What methods would modern linguistics propose? This is one idea, one, so to say, project proposal. Second, in my opinion, a theory of diplomatic ambiguity should place emphasis on methods of ambiguation, as well as on the process of disambiguation, an emphasis which could give new impetus to studies in non-applied linguistics as well. Let us start with the premise that ambiguous provisions provide some benefits in the particular context of treaty making. We should then try to better understand the processes whereby a mediator or a negotiator generates ambiguities. We should try to better understand the mental processes that produce and facilitate an ambiguous verbal outcome. Third, the processes of disambiguation have been extensively studied by means of computer simulation and automatic computing, and I believe that here one has an opportunity to blend not only two, but three disciplines: information sciences, linguistics, and the theory of diplomacy (or the theory of treaty making).

Now, a third set of reflections. Measured by the use of ambiguities in peace agreements, there is no significant difference between so-called "low-context" and "high-context" cultures. Both low- and high-context cultures use ambiguous language to bridge the gap between negotiating parties. We have Wilson's 14 Points side by side with the Oslo Accords; Chinese cross-textual ambiguities in the Shanghai Communiqué together with American referential ambiguities in the very same document. We have both the Dayton Accords and the Rambouillet draft agreement drafted by American negotiators. We have the Yalta Declaration drafted jointly by representatives of high- and low-context cultures, under the assumption that Russian culture is low-context, with which I do not quite agree. In other words, there is no direct and positive correlation between the use of ambiguous provisions, on the one hand, and types, or kinds of culture, on the other. For me, this is good news for diplomacy. Negotiators from different cultures rely on similar means to arrive at the text of an agreement. This means that, at least when it comes to the use of ambiguities, there may be a common diplomatic culture, a common culture of drafting an agreement. In other words, there is no pre-determined cultural barrier to hugely affect one's attitude towards an ambiguous proposal.29

My fourth set of reflections concerns the role of the mediator offering an ambiguous formula to two parties involved in a conflict. I believe that a mediator should explain to the parties the meaning and purpose of ambiguous wording: this does not mean that the mediator must disclose his or her own interpretation but rather that the mediator should explain why the ambiguous provision was proposed. The mediator must clearly point out, first, that an ambiguous provision is ambiguous; second, that both parties have an equal right to provide their own interpretation; third, that when it comes to the process of interpretation, both parties will enjoy an equal status and have the opportunity to present their arguments in favor of a specific interpretation; and fourth, finally, that the issue concerned was deliberately left open, but that the preferable course of interpretation should aim at a middle ground between the two incompatible interpretations that the parties take with the view of safeguarding their specific interests. I believe that a mediator must provide this explanation, because unless he does he runs the risk of losing both parties. If parties realise that an ambiguous formula has been proposed and that the chief negotiator has been silent about it, and if they have not had time to develop a relationship of trust with the mediator, then what would stop them from suspecting that the ambiguous formula may be a secret safe-guard for the other party's interest and a disguised killer of their own interests? This also implies that unless the parties sufficiently trust a mediator, no explanation will be of help. This however has general validity and does not apply solely to the situation of an ambiguous offer.

Fifth, ambiguity certainly serves the important function of setting a very rudimentary basis for future relations between erstwhile adversaries. However, its effectiveness depends to a high degree on the adversaries' attitudes and behavior during the process of interpretation. Disambiguation of ambiguity can take place only in one of the following two ways: 1. If the parties themselves arrive at a third formula, a third reading, through an exchange of arguments and additional post-peace agreement talks. This is the ideal case. 2. If there is a third party, arbitrator or mediator sufficiently trusted and respected by both parties to the peace process, to deliver a third, fair, and well-measured interpretation. Sometimes neither way is open. For instance, many of the problems that beset the process of implementation of the Dayton Peace Accords have been caused, in my opinion, not only by the local parties' unwillingness to strike another deal or to give up their narrow interpretations of the agreement. The cause of these problems lies also in the polarised opinion of the international community vis-à-vis the appropriate ways to disambiguate the ambiguities of Dayton. It is too often the case that the US and EU member countries have different, sometimes even contradictory, ideas on the direction which the Dayton implementation should ideally take. This may prove fatal to the state of Bosnia and Herzegovina, because the international community plays a decisive role there.

Finally, tolerance of ambiguity occupies an important seat among social virtues. It is widely, and justifiably, believed that societies whose members display an ability to tolerate an ambiguous state of affairs fare both economically and psychologically much better than societies whose members are lacking in such ability. Individuals tolerating ambiguity also tend to tolerate risks, to cope more easily with emotional or intellectual friction and conflict, and to refrain from jumping to premature conclusions when evidence is inconclusive. They need no general, all-embracing theories to achieve certain objectives; they are able to achieve them through a "slow and gradual accumulation of almost imperceptible nuances",30 adhering to pragmatism as their philosophy.31 Those tolerant of ambiguity do not believe in a black and white image of human affairs, but find shades of grey more attractive and enjoyable. That is why one will hardly ever find them caught in the dangerous logic of "zero-sum" games. If tolerance of ambiguity represents a value worth striving for, then why would one oppose the use of ambiguous wording in peace agreements? And yet both theoreticians and practitioners of diplomacy should understand that, though the use of ambiguous language is supported by many plausible arguments, such language nonetheless entails considerable risks and produces a desirable outcome only if a number of additional criteria have been met. One should avoid laying too much emphasis on these risks, but should not discount the factors affecting the atmospherics of the process of disambiguation either.

Hence, ambiguous peace agreements are to be treated in the very same way that those tolerant of ambiguities treat ambiguities themselves. They are to be tolerated in an ambiguous fashion, used as a last resort and employed to the best of their capacity, with all the caution they deserve. Bearing in mind that the human being is a born decipherer, with an innate tendency to reduce a complex structure to a simpler token, two seemingly contradictory stands are to be reconciled: the stand that ambiguity may play an important role in versatile settings and the stand that our innate reductionism does not cohere well with such open-ended structures. As a stubborn advocate of both compromises and ambiguities, I would suggest silencing our reductionist instinct for a while to let ambiguities play out their part to gratify our reductionism in a smooth, controlled, and truly creative manner.


1. Buehler's theory in: Karl Buehler, Sprachtheorie: die Darstellungsfunktion der Sprache (Jena: Fischer, 1934). As to the link between ambiguous messaging and international relations, and diplomacy, one should start with Robert Jervis, The Logic of Images in International Relations (New York: Columbia University Press, 1989), 113-138.

2. See Penelope Brown and Stephen Levinson, "Universals in Language Usage: Politeness Phenomena," in Questions and Politeness: Strategies in Social Interaction, ed. Esther N. Goody (Cambridge: Cambridge University Press, 1978), 56-310. For psychotherapeutic effects of ambiguous language, see Paul Watzlawick, The Language of Change (New York/London: W.W. Norton & Company, 1978), 78-90.

3. Ronald Munson, The Way of Words (Boston: Houghton Mifflin Company, 1976), 74. Munson, on page 73, offers another definition of ambiguity: "An expression is ambiguous when it can be understood in more than one way, and we aren't certain which way is intended." This confirms my point concerning the intrinsic vagueness of the concept.

4. The "duck-rabbit" has been reprinted in thousands of books on the elementary psychology of perception. For those who have not yet seen it, see, for instance, C. J. Adcock, Fundamentals of Psychology (Harmondsworth: Penguin Books, 1964), 131.

5. The History of Herodotus (Chicago/London/Toronto: Encyclopaedia Britannica Inc., 1952), 11.

6. For a sub-division into contrastive and complementary ambiguity, see U. Weinreich, "Webster's Third: A Critique of its Semantics," International Journal of American Linguistics 30 (1964), 405-409. However, I am not sure that complementary ambiguity is real ambiguity.

7. See Kent Bach, Ambiguity, Routledge Encyclopedia of Philosophy website; available at (Accessed on September 15, 1999).

8. Geoffrey Leech, Semantics (Harmondsworth: Penguin Books, 1987), 79; See also G.W. Turner, Stylistics (Harmondsworth: Penguin Books, 1973), 100-102; and a discussion about the syntactical kind of ambiguity in Steven Pinker, The Language Instinct (Harmondsworth: Penguin Books, 1994), 192-230.

9. Aristotle, "Peri ton sophistikon elenhon, 165a 10-20," in Aristotelis Topica et Sophistici Elenchi, recensuit brevique adnotatione critica instruxit, ed. Sir D. Ross (Oxford: Oxonii, 1958). For a more comprehensive and complex interpretation of Aristotle's view of ambiguities, see Robert Wardy, The Birth of Rhetoric (London/New York: Routledge, 1998), 130-132.

10. For more details, see Lexical Semantics: The Problem of Polysemy, ed. James Pustejovsky and Branimir Boguraev (Oxford: Clarendon Press, 1996).

11. "Interpretability of some important parts of the text was introduced (chosen) deliberately in order to do justice to both, mutually exclusive, positions." For Kosovo, Petritsch and Rambouillet, see Wolfgang Petritsch, Karl Kaser, and Robert Pichler, Kosovo-Kosova (Klagenfurt- Wien: WieserVerlag, 1999), esp. 327-328.

12. Aristotle, "Athenaion Politeia, 9 2, p. VIII," in Aristotelis Atheniensum Respublica, recognovit brevique adnotatione critica instruxit, ed. F.G. Kenyon (Oxford: Oxonii, 1958). Kenyon's English translation of the Athenian Constitution is also available on the Internet at

13. My thanks to Jovan Kurbalija for drawing my attention to the case of the 1st Nicean Council through one of his issues of DiploNews, available at

14. See Walter Lippmann, Public Opinion (New York: Free Press Paperbacks, 1997), 133-138; 1st edition published in 1922 in New York by Macmillan.

15. My account is based on the following books: Winston Churchill, Triumph and Tragedy (Boston: Houghton Mifflin Company, 1953); Arthur Conte, Yalta ou le partage de monde (Paris: Robert Laffont, 1966); W. Averell Harriman and Elie Abel, Special Envoy to Churchill and Stalin, 1941-1946 (New York: Random House, 1975).

16. See a number of reference documents at the website of the Foreign Ministry of Israel:, especially Statements Clarifying the Meaning of U.N. Security Council Resolution 242.

17. This account is based on James O.C. Jonah, "The United Nations and International Conflict: The Military Talks at Kilometer Marker-101," in Mediation in International Relations, ed. Jacob Bercovitch and Jeffrey Z. Rubin (London: Macmillan, 1992), 176-205.

18. See Henry Kissinger, White House Years (Boston: Little, Brown and Company, 1979), 783-784. See also Paul Wolfowitz, "Remembering the Future," The National Interest, Spring 2000, 44-45.

19. Financial Times, 7 May 1998, 14: Interview with Madeleine Albright.

20. For the text of the Dayton Accords, which awaits official translation into a language common to all Bosnians, see Robert Hayden considers the part of the Dayton constitution I discussed here "confusing", which implies that he did not recognise its ambiguous nature. For details, see Robert M. Hayden, Blueprints for a House Divided: The Constitutional Logic of the Yugoslav Conflicts (Ann Arbor: University of Michigan Press, 1999), 123-139.

21. See "Goodbye to all that," The Economist, 18 April 1998, 11, and "From process to procession," The Economist, 18 April 1998, 15-18.

22. See Dominic Cunningham, "Peace process gets kiss of life in new deal," Irish Independent-Online, 18 December 1998,; and William Graham, "Assembly deal breakthrough," Irish News, 18 December 1998,

23. See also John Loyd, "Ireland's Uncertain Peace," Foreign Affairs, September/October 1998, 109-122.

24. In the following argument, I will focus on Thomas M. Franck, Fairness in International Law and Institutions (Oxford: Clarendon Press, 1997), 30-34.

25. Ibid., 30.

26. Ibid., 31.

27. Ibid., 33.

28. Johan Galtung, Peace by Peaceful Means (Oslo/London: PRIO/ Sage, 1996), 265. For the concept of conflict transformation, see also Peter Wallensteen, "The Resolution and Transformation of International Conflicts: A Structural Perspective," in New Directions in Conflict Theory, ed. Raimo Vaeyrynen (London: Sage, 1991), 129-151.

29. This is a highly specific assessment, and perhaps should not be treated as a critique of the school of cultural relativism in the theory of negotiations, as formulated by one of its prominent advocates, Raymond Cohen. See Raymond Cohen, Negotiating Across Cultures (Washington: United States Institute of Peace Press, 1991).

30. This is a paraphrase of Henry Kissinger's sentence that "Japanese leaders will make adjustment by the accumulation of apparently imperceptible nuances", from Diplomacy (London: Simon & Schuster Ltd., 1994), 827.

31. Or Bart Kosko's philosophy of "fuzzy thinking".

Scroll to Top