The Concept of Genocide by Paul Boghossian

The Concept of Genocide

Paul Boghossian, New York University

[Source / Contact]


The word “genocide” gets a lot of use these days. Whenever a large number of people are killed, it is common for someone to suggest that the event qualifies as genocide and for someone else to dispute that classification.

Recently, the highest UN court ruled that the Srebrenica massacre of 8000 Bosnian Muslims in 1995 was genocide. Most major recent conflicts in Africa have attracted that classification: Rwanda, Darfur. The spread of the conflict to Chad is labeled a “potential genocide.” And so forth.

Are these accusations justified, or have accusations of genocide just become a cynical political tool, as some commentators have been complaining?

My own interest in this issue has autobiographical roots. Like just about any Armenian in the West, the explanation for how I ended up here goes through the near wholesale expulsion of the Armenians from the Ottoman Empire between 1915 and 1923. A people that had lived on the Anatolian peninsula for some 3000 years ended up disappearing from that region almost entirely, mostly in the space of some four months in 1915, leaving behind hundreds of towns, villages and (most historians agree) over one million dead.

Most Armenians insist that what happened to their ancestors in 1915 was genocide. The Turkish Government, while not denying that there were deportations and killings, vehemently denies the label. How should we decide who is right?

My view about this is that, in a sense, in most of these contested cases, the accusations of genocide are justified – but that this is a hollow verdict because it derives from deep flaws in the way the concept of genocide is defined, flaws that it is not easy to see how to remedy.

Genocide Defined

What does the term “genocide” mean?

The first and most important fact about the term “genocide” is that it is not a word of ordinary language. It’s not a word like, say, “kill” or “harm” that has been part of ordinary English for centuries, acquiring a meaning through established usage, a meaning that dictionaries attempt to specify. Rather, it is a made-up word, coined by the Polish jurist Raphael Lemkin in the 1940’s.

Lemkin was moved to introduce the term primarily by two large events – by what Hitler had done to the Jews and by what the Ottoman Turks had done to the Armenians. He thought that these two incidents exemplified a distinctive crime, one worthy of being denoted by a distinctive term, which we lacked.

The word is composed out of two roots: the Greek word for people – genos – and the Latin suffix – cide – for murder. The murder of a people – that is the rough idea that Lemkin wanted to capture with the word “genocide.”

The United Nations recognized the usefulness of the concept that Lemkin had introduced but struggled to make it precise enough so that it could be the basis of international law.

(1) I will regard the meaning of the word “genocide” to be equivalent to the concept genocide. (Italics will be used to indicate that a word is the name of concept, although they will also be used in the standard way for emphasis; context will always make it clear which use is intended.)

(2) That the Armenian case featured in Lemkin’s thinking is occasionally denied. But the evidence that it did is overwhelming. See the article by Khatchig Mouradian.

In 1948, it adopted the Convention on the Prevention and Punishment of the Crime of Genocide which incorporated the following definition of the act of genocide:

Genocide means any of the following acts committed with the intent to destroy, in whole or in part, a national, ethnical, racial or religious group, as such:

(a) Killing members of the group;
(b) Causing serious bodily harm;
(c) Deliberately inflicting on the group conditions of life calculated to bring about its physical destruction in whole or in part;
(d) Imposing measures intended to prevent births within the group;
(e) Forcibly transferring children of the group to another group.

The fact that the term “genocide” is a coined technical term implies that anyone using it must either use it to mean this technical concept or must explicitly supply an alternative definition. One cannot use the word “genocide” without supplying some definition or other, because one cannot rely, as one might with a word of ordinary language, on some common understanding that we all have of that word, whether or not we are able to define it.

The question is: What exactly does the definition supplied by the UN mean?

On the whole, and at first blush, the definition doesn’t seem all that obscure. You commit genocide against a particular group G of the appropriate sort, just in case you kill, harm, etc., members of G, with the intention of destroying that group, either in whole or in part, as such. Perhaps “as such” could use some clarification. I’ll talk about it more later on, but for now let us read it as meaning: just because they were members of that group.

Genocide and 1915

Read as literally as possible -- and how else should we read a law – how could this concept fail to apply to 1915?

(3) Henceforth, I will abbreviate the list of harms to “kill.” It’s very rare to hear of the label genocide being applied to cases that do not involve killing, even though the definition allows for such cases.

The Armenians formed an ethnic group, so they are of the appropriate type. Some of them were killed. Were some of those who were killed, killed with the intention of destroying the Armenian ethnic group either in whole or in part as that very ethnic group?

It’s hard to believe that the answer to this is ‘No’. The ‘in part’ qualification means that you don’t need to have intended to kill every last one of them. (In fact, if you are being very literal minded, even one Armenian counts as ‘a part’ of the Armenian ethnic group.) So, it looks as though, according to the definition, all you need to have done is intentionally sought to destroy a number of Armenians, just because they were Armenians.

Now, who could reasonably deny that that happened in 1915?

The Turkish Government is fond of saying that there is no evidence that the Young Turk Government of the time ordered anyone to be killed or even harmed. Inference to the best explanation would surely suggest otherwise; but in any case the point is irrelevant. The UN definition doesn’t require the killings to have been committed by the State for them to count as genocide. There is simply no mention of a state in the definition.

So, if we look at the conditions laid down by the technical definition provided by the UN, it’s very hard to see how the term “genocide” could fail to apply to 1915.

Now, there are some who argue that even if the events of 1915 do satisfy the conditions laid down in the technical definition, the term still could not correctly apply to 1915. They supply one of two arguments for this claim.

The first argument goes something like this. The word “genocide” did not exist in 1915. It was first coined by Raphael Lemkin and then codified by the UN Convention in 1948. How it is possible for a concept that was introduced in 1948 to apply to events that occurred in 1915?

(4) That is not to deny the obvious interest of the question what the government’s role was in ordering, organizing and facilitating the mass murder that occurred.

The first thing to say about this argument is that one can’t in general assume that a concept can apply to some event (or object) only if, at the time that the event occurred (or the object existed), there were people around who were prepared to apply that concept to it. For example, I can truly say that there were dinosaurs on Earth 65 million years ago even though 65 million years ago there was no one around who had the concept dinosaur.

Some scholars give evidence of wanting to deny this. When French archeologists working on the mummy of the Pharaoh Ramses II came to the conclusion that Ramses had died of tuberculosis, the famous sociologist of science Bruno Latour questioned their conclusion: “How could he pass away due to a bacillus discovered by Robert Koch in 1882?” he asked. He went on to claim that it would be just as anachronistic to say that Ramses had died of tuberculosis as to say that he had been killed by machine gun fire. In his article he asserted: “Before Koch the bacillus had no real existence.”

I’m not sure whether Latour’s surprising view here derives from the general constructivist claim that nothing can be a particular thing before it is described as being that thing, or whether he thinks that there is something special about bacilli. If it’s only the latter view, he should probably have explained what he thinks is so special about bacilli.

At any rate, I think we can all agree both that the general view is wrong and its particular application to this case: It would seem that I can truly say that Ramses died of tuberculosis even though no one at his time had the concept of tuberculosis.

However, it is true that some concepts are such that, nothing could be said to fall under them at a given time t unless someone is prepared to say at t that something falls under them. Let’s call such concepts contemporaneous-response-dependent concepts. The concept of being hip is arguably response-dependent in this sense: no one can count as hip unless some people think of him as hip. Other candidates would be the concept coronation and the concept priest.

(5) See Bruno Latour, “Ramses II est-il mort de la tubercuose?” La Recherche, 307 (March 1998), 84-85.

So, some concepts are response-dependent and others aren’t. And the question is: Is the UN concept of genocide contemporary-response-dependent in this way, so that it could not apply to some event unless at the time of the event some people were prepared so to describe it?

It would appear not. As we have seen, the main ingredient in the definition is an intention: to harm members of an ethnic group qua members of that group. What concepts would someone need to have in order to have and be able to act on such an intention? Well, he would have to have the concept of an ethnic group and he would have to be able to form the intention of harming someone just because they were part of that very ethnic group. Clearly, someone can have that intention and act on it even if no one contemporaneous with him had the concept of the distinctive crime of genocide. So it does not look as though genocide is one of those contemporaneous-response-dependent concepts that can apply to an event only if there are people around who, at the time of the event, are prepared to apply that concept to that event.

Another unconvincing argument for refusing to apply the word “genocide” to 1915 goes something like this. The UN Convention on Genocide, which defined the word precisely for the first time, was adopted only in 1948. Treaties don’t apply retroactively. So, the Convention, along with the notion that it defines, could not apply to the events of 1915.

This argument is unconvincing because it conflates the question whether the concept of genocide applies to events that preceded its introduction with the entirely different question whether the legal convention that formulated that concept (perhaps for the first time) applies to events that preceded its adoption.

One of the more interesting developments of the process known as the Turkish Armenian Reconciliation Commission involved a decision by the Commission to ask an independent legal body – the International Center for Transitional Justice – to rule on whether the Genocide Convention applied to the events of 1915. The lawyers of the Center had no difficulty quickly reaching a two-part verdict. Of course, they reported, the Convention, qua legal document, does not apply to events that preceded its adoption. It is very rare for a treaty or a law to apply retroactively. However, they noted, the concept that is codified by the Convention clearly does apply: the events of 1915 count as genocide by the criteria formulated by the Convention.

The Turks on the Commission were very unhappy about this ruling and resigned en masse. They thought the ICTJ would rule only on the very narrow and boring question whether a treaty applies retroactively. But once the ICTJ was asked to consider the question, there was no reason they couldn’t also consider the question whether the treaty would have applied, had it already been in force in 1915 – that is, whether the concept that it defines applies to those events.

The Point of the Concept of Genocide

So, here is where we are so far. If we don’t question the validity of the UN concept, I can’t see any difficulty with saying that it applies to 1915.

What does worry me, though, is the validity of the UN concept itself. I think a strong case can be made for saying that that concept is deeply flawed, flaws that make its application to particular cases deeply problematic and that are hard to remedy. This is what I want to discuss in the remainder of this paper.

Now, someone might think: Wait a minute! Isn’t there something funny about complaining about a concept? Surely, one can mean anything one wants by one’s words. A claim can be criticized, of course, for it can be true or false. But in what sense can a concept be criticized? If I introduced a word into the language – say “flurg” – and stipulated that it is to mean, “Murders committed on a Tuesday,” surely that would be up to me. And surely I can then ask of any given murder whether or not it is flurg and get a determinate answer.

In a sense, this is correct. However, when a moral or legal term is introduced into the language it is done with a purpose, a purpose that the term carries with it, either as part of its meaning or as something importantly associated with it. I think we can recognize three such constitutive purposes behind the introduction of the term “genocide.”

(6) For an account of this particular reconciliation process, see the book by its chairman, David Phillips, Unsilencing the Past: Track Two Diplomacy and Turkish-Armenian Reconciliation, (New York and Oxford: Berghahn Books, 2005).

First, it is supposed to name a distinctive phenomenon, something that is not adequately denoted by any of the other terms that we already have in the language. Lemkin believed that in the Jewish and Armenian cases he had observed a distinctive crime, something that deserved to be called the murder of a people, that none of the other terms at our disposal – mass murder, for example – denoted.

Second, it is supposed to name a phenomenon that is, as part of its very meaning, morally wrong. Unlike the notion of killing, for example, but like the notion of murder, it is not supposed to be an open question about any given genocide whether it was morally reprehensible: it’s not supposed to be intelligible to ask: Yes, it was genocide, but was it justified? It’s supposed to follow analytically from the very meaning of the term that, if it was genocide, it was wrong.

Finally, genocide is taken to name not only a distinctive crime but one that is distinctively heinous, deserving of a special measure of censure. Mass murder may be bad; but mass murder done in the context of the targeting of a particular group is supposed to be morally far worse.

When I say that the concept is flawed I mean (a) that the UN concept is ill-suited to satisfy these purposes and (b) that it’s hard to see how any modification of it would do better.

Genocide as a Distinctively Heinous Crime

Let us assume, for the moment, that the UN concept satisfies aims (1) and (2): it names a distinctive phenomenon that is, as part of its very meaning, morally bad. But does it succeed in naming a phenomenon that is distinctively morally heinous? There are two challenges to the claim that it does.

The first has to do with whether it’s true that targeting a particular group really is morally worse than simply killing large numbers of people. It is an interesting and unanswered question in this area what moral principles are needed to underwrite the claim that targeting a group is worse than simply killing a large number of people who happen to be members of that group. Is the idea that, in addition to individual rights, there are group rights as well, so that when you target a group you violate more than the rights of the individuals involved? This might be true, but it would need arguing; and it isn’t obvious to me how that argument should be formulated.

But let’s suppose that a satisfactory case for that claim can be formulated. A different question now arises from the opposite direction. Supposing we can justify saying that targeting a group is morally worse than targeting its members, why is it just ethnic, racial, national and religious groups that count as far as this special moral censure is concerned? Why doesn’t destroying a group that is organized around an ideology, or around class, or around wealth, count equally? Stalin killed millions of kulaks because of their alleged opposition to collectivization. Why shouldn’t that count?

To this objection, some have responded that the definition rightly singles out for special consideration indelible identifications, identifications that one is born with and that one can do nothing about. It is thought to be especially bad to target people because of what they indelibly are, as opposed to what they may have blamelessly become (wealthy, anti-Communist, or what-have-you).

But if we are so concerned with indelibleness, why is gender not included? And why is religion included? You are not born into a religion in the same sense as you are born into the color of your skin or into your gender.

And, anyway, why should indelibleness matter morally in this way? Is it really more morally reprehensible to kill people for what they biologically are than it is to kill them for what they may have blamelessly become? That seems like a peculiar idea.

But if it’s not that, what is the rational basis for this particular selection of the groups that are mentioned in the UN definition, against which this especially heinous crime can be committed?

(7) This point has been extensively discussed in the literature. see, for example, Leo Kuper, Genocide: Its Political Use in the Twentieth Century (New York: Penguin Books, 1981); Pieter N. Drost, The Crime of the State (Leyden: A. W. Styhoff, 1959).

Suppose that we respond to these observations by agreeing that the UN definition is arbitrarily narrow, that we should expand the kinds of groups that are allowed to count. Do we then just say that any group can count? Would residents of Hiroshima and Nagasaki count as a relevant group? Would resident of Dresden? Should we say that what happened on September 11, 2001 was a genocide of the group consisting of those holding jobs in the World Trade Center? (Pace Kuper and Charny, that seems absurd.)

Some people think that the solution is to focus on groups that identify themselves as a group, something that those holding jobs in the World Trade Center presumably didn’t do. But it should be clear that that would still let in an enormous number of groups that no one thinks of as possible objects of genocide – e.g., faculty members at NYU.

Genocide as a Distinctive Crime

So far, I have been worrying about how well the UN concept captures a distinctively heinous crime. Let me now turn to asking how well it does denoting a distinctive crime, regardless of whether it is distinctively heinous.

The problem here derives from the phrase “in whole or in part.” The intention must be to destroy a particular ethnic, national, racial or religious group, either in whole or in part, as such.

What is it to intend to destroy a particular ethnic group, in part? Presumably, it is to intend to destroy a part of it. But the trouble is that a “part” of a particular group can be a very small part. And the definition does nothing to rule out that particular reading.

It is this that explains the UN ruling in the case of Srebrenica. But if it applies in that case, why wouldn’t it apply in a host of other cases, too?

When, in the 1970’s and 1980’s, Armenian gangs targeted Turkish diplomats for assassination, they were targeting individuals who constituted a “part” of the Turkish ethnic group; and they were targeting them precisely because they were a part of that group. In that sense, then, they seemed to have had the intention to destroy a particular ethnic group, either in whole or in part, as such. But proponents of the concept wouldn’t want to say that those Armenians were committing genocide against the Turks.

(8) See Israel Charny, “Toward a Generic Definition of Genocide”, in G. A. Andrepoulos, ed., Genocide: Conceptual and Historical Dimensions. (Philadelphia: University of Pennsylvania Press, 1994, p. 64-94).

Or take the case of the Jewish “revenge” gangs who targeted Germans after the end of the Second World War. They, too, would appear to satisfy the definition. But it would once again be deemed absurd to say that the Jews have committed genocide against the Germans.

It might be thought that the solution here is simple and clear: the definition should be modified to read “in whole or in substantial part,” or something along those lines, thus ruling out the sorts of case that I have just now described, which involved a very small number of fatalities.

But while this modification might deal with those cases, it would leave other problem cases intact. The September 11 attackers were clearly targeting Americans because they were Americans. It would appear that their intent was to destroy the American national group in part, and as such. But no one is prepared to count September 11 as genocide.

The intuition that fuels the thought that the Armenian and Jewish cases exemplify a distinctive crime clearly recoils from the thought that September 11 also exemplifies it. But how is that captured by the UN concept? Surely, 3000 people count as a “substantial” part of the American national group. Or are we to think that an important moral difference resides in the distinction between 3000 and the 8000 of Srebrenica?

So the “in part” clause is highly problematic.

On the other hand, it is not as though we can simply eliminate it. The trouble is that requiring that the intention be an intention to eliminate the absolute entirety of the group means that what happened to the Jews and the Armenians is not genocide after all, for the Nazis were willing to spare some Jews and the Turks did spare large numbers of Armenians (in the sense that they could have killed them if they wanted to but chose not to do so, settling instead for their forced expulsion from the boundaries of the envisaged Turkish Republic).

The only solution that I can see would be to set some sort of arbitrary percentage threshold, so that some mass killings count as genocide if they exceed that threshold, but not otherwise. But that seems absurd.

The problem that we face here is that there is a gross and seemingly irremediable vagueness in the notion of “killing a people.” It is clear enough what it is to kill a person. But it is not so clear what it is to kill a group, unless you are willing to so restrict the notion that it involves killing very last member of the group.

Most concepts are vague, so it is not a sufficient indictment of a concept to point out that it is vague to some extent. But there are limits to how vague a concept can be and still be useful, especially when it is involved in charges of such moral and legal gravity.

Genocide as Analytically Immoral

Let me turn, finally, to the “as such” condition. The intention must be to destroy a particular ethnic group, as such. What does that mean?

In philosophy, the phrase “as such” is usually used to mean “under that very description.” If we read it in this way, the idea would be to emphasize that, in the case of genocide, the intent must be to destroy a particular ethnic group, under that very description, and not under some description that is co-extensive with it.

Let me take a moment to illustrate what I am talking about here. Suppose you form the intention, for whatever reason, to detain for questioning the first ten people to walk off a given plane. Suppose that the first ten people to walk off the plane are in fact all Irish and indeed the only Irish citizens on the plane. It would be wrong to describe you as having formed the intention to detain all the Irishmen on the plane. Your intention was to detain the first ten people coming off the plane; these ten just happened to all be Irish.

(9) There is some discussion of this in the literature, but it doesn’t seem to me to put its finger on what is genuinely problematic about this clause. See Martin Shaw, What is Genocide? (Polity, 2007, p. 81-96, for a general discussion.)

When people form intentions they form them under certain descriptions and not others; we cannot accurately report on those intentions unless we use pretty much the same descriptions they use and not just descriptions that happen to be co-extensive with the descriptions that they used.

Now, the trouble with this reading of the “as such” condition is that it seems redundant: given that we are talking about intentions, it is an obvious truth about intentions that they are description-sensitive in this way. There would be no need to emphasize the point in the way that the UN definition does.

Furthermore, if we read the “as such” clause simply as meaning “under that very description,” the definition of genocide would leave it entirely open whether a particular act of genocide was or was not justified, in contravention of the second of the constitutive purposes with which the term was introduced.

Consider the firebombing of Dresden, and suppose it is debatable whether that firebombing was justified. The residents of Dresden constituted a substantial part of the German national group. They were picked as targets under the description Germans. The act of firebombing and thus killing and maiming large numbers of them was intentional. So, the three elements of the definition of genocide are satisfied. But it is left open whether the act was morally justified under the circumstances.

When we look at how the “as such” condition is actually employed, though, it is clear, I think, that it does not just mean “under that very description.” As is already made clear by the travaux preparatoires to the UN convention, the point of the “as such” condition was not to insist on the obvious but to lay down a condition regarding motive: for genocide to occur there must be an intention to destroy a particular group just because it is that very group. That is the point of the “as such” condition. It is not merely to say: the intention must be to destroy a particular ethnic group under the description “that very ethnic group” but, rather, that it must be an intention that is arrived at for a particular reason or motive: to destroy a particular ethnic group just because it is that very ethnic group.

Clearly, these two readings are not equivalent; we can illustrate once more by the case of Dresden. The residents of Dresden were intentionally bombed under the description German residents of a German city. But they were not bombed just because they were Germans. Rather, they were bombed in part because they were Germans and in part because there was a war going on between Germany and the Allied Powers.

So what does it mean to say that the intention must be to destroy a particular ethnic group just because it is that very group?

On first pass, this seems easy enough to understand. For genocide to occur, when you set out to inflict harm on the members of the group, you must be moved to act by the fact that they are who they are -- Armenian or Jewish, whatever.

On second thought, though, the matter is not so clear.

The problem is that to say that you need to have a certain motive is ambiguous between saying that that motive must be part of your overall motive and saying that it must be the whole of it. And there are problems no matter which way the condition is disambiguated.

Suppose we pick the weaker reading, so that harming the group just because it is that very group simply has to be part of the overall motive. This would have the unwanted effect of classifying Dresden as genocide, since part of the motive for selecting that target was, of course, that it was a German city full of Germans.

This pushes us to the stronger reading according to which, in the intended sense, an intention is genocidal when someone is moved to inflict harm on certain individuals merely because of their membership in a given group. It’s not enough that their membership in that group enters into it; it’s that their membership alone explains why they are being attacked. And, of course, as previously emphasized, Dresden wasn’t like that. The Allies did not bomb that city just because it was German, but because there was a war going on.

The trouble with this stronger reading is that it is nearly never the case that the sole explainer for why someone acts to kill or harm large numbers of people is just the identity of those people.

It is very unlikely that the Armenians were killed just because they were Armenian, but also because, as a large Christian civilization that had lived on the Anatolian plateau for millennia, they stood in the way of the project of building a Turkic Muslim nation-state out of the remnants of the Ottoman Empire. Never mind if this is right as an explanation. The point is (a) that there were likely to be many other motives beyond just that of identity and (b) that acknowledging that fact doesn’t make the crime any less heinous, any less “genocidal.”

So, the “as such” condition seems very problematic as well. It could mean either “in part because of whom they are” or “merely because of whom they are.” The former seems too weak a condition and the latter too strong.

Some have suggested a more specific reading of the “as such” condition – as implying hatred for the group in question. This has the effect of assimilating genocide to what is now known as a hate crime. But we can deploy a slightly fanciful case to show that that is not a good requirement.

Imagine a dictator who rules over an enormous empire comprising many different ethnic groups. Every once in a while he chooses to demonstrate to his subjects that he is their undisputed master by picking some ethnic group at random and destroying it. He spins, so to speak, a wheel of misfortune, and whichever group the wheel happens to stop on he sets out to destroy. It would be natural to class this along with the other cases as genocide, but the dictator may have had nothing against them; it was just his way of showing the others who is boss. So it would be wrong to say that he destroyed them because he hated them or just because they were that very group. He destroyed them because that was the group that the wheel landed on and it had to land on some group or other.

What is the “as such” condition doing there, anyway? Can’t we just eliminate it? Why do we need a clause specifying motive?

I believe that the reason for the inclusion of the “as such” condition has to with the need to satisfy the second of the three desiderata specified for the concept: namely, that the concept mark out a distinctive phenomenon that is, by virtue of its very meaning, morally reprehensible.

The point is that the mere targeting of a group is not in itself necessarily a bad thing. Suppose that that group has attacked you or has sided with your enemies. Then attacking them in return would be justifiable self-defense. So, if you want a notion of the destruction of a group that is reprehensible as a matter of its very meaning, you must somehow include in its definition that the destruction of the group was not morally justifiable.

One way to do that, of course, is simply to put that very phrase into the definition: Genocide is the act of…committed with the intent of destroying a particular …group, in whole or in part, without moral justification.

The trouble is that the minute you invoke the notion of “moral justification” explicitly, it triggers peoples’ relativistic impulses: who’s to say what is and is not morally justifiable?

From this point of view, the framers of the UN convention came up with an ingenious solution. Since one can only justify attacking a group of people because of something they have done and never simply because of who they are, the “as such” condition ensures that the defined phenomenon is necessarily reprehensible, even if it doesn’t explicitly state that it is.

The trouble is that this way of satisfying the second desideratum runs into the problems we have outlined.

(10) This example was suggested to me by Kit Fine.


So, each element of the UN concept’s intention clause seems deeply problematic. Together, they add up to a very serious critique of that concept, showing that it hasn’t succeeded in satisfying the aims that underlie its introduction.

Now, I am hardly the first to make at least some of these points. But many commentators stop only long enough to note some problems and then proceed to say something pragmatic like: well, every definition has problems; or, we need to use something and UN definition is the best thing around; or, so many countries have ratified this Convention by now – and do you have any idea how much work that took – it’s simply too late to do anything about it. And so forth.

I myself can’t see taking such a relaxed attitude. It makes me very uneasy that accusations of such gravity are conducted in terms of a concept that is as vague and as problematic as the UN concept is.

Some theorists have tried to modify the UN concept. Although I don’t have the space to discuss them here, every one of the proposals that I’ve seen runs into very severe difficulties, some of which are worse even than the problems we have just been exploring for the UN definition.

What should we do? I don’t know that I have an answer. But I think we should be careful not to do something that many in the Armenian community seem to want to do – and that is to frame the issue that divides them from the Turkish

Government as resting exclusively on the applicability of this special label ‘genocide’ to what happened in 1915. The word is too fragile a reed to sustain so much weight.

Even without the availability of the concept of genocide, we can still point out that in 1915 over a million Armenian men, women and children were either intentionally killed or died during mass deportations that were conducted with wanton disregard for life. We can observe that there was no conceivable moral justification to sanction the Ottoman Government treating some of its subjects in this way. We can add that it not only brutalized and dehumanized them, but also confiscated their lands and possessions, and attempted to destroy their centuries-old culture so as to make it seem that they had never lived in those lands in the first place. And that to this day its successor, the Government of the Republic of Turkey, engages in an elaborate and expensive campaign to deny and cover up the fact that all these events occurred.

(11) See, for example, Charny, op. cit., 75. F. Chalk and K. Johansson, The History and Sociology of Genocide, (New Haven and London: Yale UP, 1990), p. 23. Helen Fein, Genocide: A Sociological Perspective, (NY: Sage Publications, 1993), p. 24.

What I think we should resist is the temptation to capture all this in one neat word.

(12) I am grateful to audiences at the Ohanessian Lecture at the University of Minnesota, the Remarque Institute for European Studies at NYU, the New York Institute for the Humanities and to David Velleman, Eric Weitz, Melis Erdur and David Phillips for comments on earlier versions of this piece.