What is Genocide?
From Raphael Lemkin to the Genocide Conventions
"Genocide... does not necessarily signify mass killings although it may mean that. More often it refers to a coordinated plan aimed at destruction of the essential foundations of the life of national groups so that these groups wither and die like plants that have suffered a blight." -- Raphael Lemkin, 19451
The public discourse on genocide seems often either to lack familiarity with the established scholarly and legal definitions or to assume that the reader already has familiarity without explaining. We thus have voices labeling current events undeniably genocidal and others labeling this claim as absurd, with a dangerous assumption that a contrary legal finding to your preference must only result from bias.
So lately I've been reading essays by and about Raphael Lemkin, the law professor who originally proposed the legal framework against genocide in 1933. You read right: 1933. We often discuss the Genocide Conventions as a response to the Holocaust, but that's only partially correct. Lemkin invented the word genocide in the 40s, when people were finally listening to him, but he proposed an international legal framework for the concept soon after the Nazis took power. Lemkin was a scholar of international law and of history, and his proposals took at least as many examples from colonial rule as from his own experiences of antisemitism. The concept of genocide, as created by Lemkin and enshrined in the Genocide Conventions, refers to overarching strategies for the annihilation of groups that usually far precede the naked instrument of mass execution.
The Nazis did not jump directly to gas chambers. In their early years in power, they engaged in strategies to make life sufficiently miserable and unsafe for Jews and other scapegoats that they would emigrate.2 Already at this stage, in 1933, Lemkin wanted any Nazi official who stepped outside of Germany arrested for crimes of barbarism against humanity. He wanted the undermining of a people's cultural, political, and religious foundations of life all recognized on a continuum of a common crime that leads ultimately to the destruction of that people.
If your immediate image of genocide is the Holocaust, and you believe that any current instance qualifies as genocide to the extent that it resembles mass execution by gas chambers, you are incorrect. The Genocide Conventions aim to sanction and arrest a process well before that point. Preventing genocide doesn’t mean having laws to punish a future holocaust; it means having laws that kick in at recognizable stages before such atrocities can happen—and by normalizing the values in those laws, having a global public less willing to tolerate genocide. The point is not to punish Nazis in 1945. The Nuremberg trials did that anyway with no reference to genocide. The point is to stop Nazis in 1933. The postwar world was reeling from one of the worst crimes in all of human history, and Lemkin’s brilliance lay in conveying the broader patterns of which this crime was part.
To start, what does this word “genocide” mean? Words generally have fuzzy boundaries and mean whatever people convey with them. Rare cases exist where words have prescribed technical meanings, though usually these only define an existing fuzzy word for specific contexts, like “energy” in physics. Still rarer words originate and operate continually with a technical definition, such that uses contrary to those definitions are functionally incorrect, rather than merely different.
Genocide is such a term. There are scholarly definitions originating with Lemkin and a legal framework enshrined in international treaty in 1951. Colloquial uses divorced from these are misleading. In serious discussion, you can reference either the scholarly or the international legal definitions (and note their important differences). Anything else, intentionally or not, can make genocide more feasible by undermining the functions of the established definitions. With that in mind, let's follow a few points of Lemkin's scholarship and how that translates into the narrower language of the Genocide Conventions.
Lemkin and the Scholarship of Genocide
Lemkin defined genocide as “a coordinated plan of different actions aiming at the destruction of essential foundations of the life of national groups, with the aim of annihilating the groups themselves.”
If we stopped reading here, we could interpret the word in simple etymological terms: geno-cide is entirely annihilating a people, i.e. killing all its members. But with even a moment’s reflection, this take on the definition is problematic. If we applied it consistently, we would call most genocides in history attempted genocides. We would conclude that even the Holocaust was not a genocide, because the Nazis “only” murdered 2/3 of all European Jews, and probably a similar proportion of Roma,3 and not all of them. Even the utter destruction of many Native American nations at the hands of the United States would only count as “attempted” if any descendants survived. This way of interpreting the definition is obviously stupid.
Two parts of Lemkin’s definition become important here. The first and more obvious is that of aim. Genocide aims at the destruction of essential foundations of life, with the aim of annihilating the group. Genocide coordinates strategies that attempt to destroy groups. International law surrounding genocide aims to arrest such strategies and make them fail.
The second relevant part of the definition is equally obvious for anyone who reads Lemkin’s essays, but less obvious from the definition alone. That is the question of what it means to annihilate a group. If you aren’t at least trying to kill every single member of a group, it isn’t genocide, right? Very much wrong!
Lemkin tells us that “Genocide has two phases: one, destruction of the national pattern of the oppressed group; the other, the imposition of the national pattern of the oppressor. This imposition, in turn, may be made upon the oppressed population which is allowed to remain or upon the territory alone, after removal of the population and the colonization by the oppressor's own nationals.”
Wait, if the aim is annihilation, how can the second phase be met if the oppressed population is allowed to remain? Because annihilation of the group does not necessarily mean physically killing all of its members, even as its ultimate aim. Lemkin elaborates that the objectives of genocidal plans “would be disintegration of the political and social institutions, of culture, language, national feelings, religion, and the economic existence of national groups, and the destruction of the personal security, liberty, health, dignity, and even the lives of the individuals belonging to such groups.” He then gives examples of how policies such as dismantling a group’s academic presence or controlling or destroying its art can be genocidal in and of themselves, because they destroy collective identity and culture.
Following Lemkin’s examples, violent assimilation also counts as genocide. Partial physical destruction of a targeted group to force the survivors’ labor also counts as genocide. Rendering a group’s quality of life so wretched that they all must leave counts as genocide—and it’s this very case that first prompted Lemkin’s legal propositions. All of these achieve the goal of the destruction of the group and the imposition of the oppressor’s national aims.
Genocidal regimes will select tactics based on expedience, and they may or may not deem physical destruction necessary. When they do, according to Lemkin, the first technique of the physical destruction of a people is racial discrimination in feeding. Endangering of health is the second technique, and mass killing is the final option. Genocide does not require a single shot be fired if you destroy a people's homes or food supply.
Let’s return to the question of aim, which will translate into the legal concept of intent. Does the eradication of a people need to be an end in itself for actions to count as genocide? What if you’re murdering people for some other goal, and those people just happen to belong to a distinct religion or ethnicity? Using this logic, some historians argue that Belgian rule in Congo doesn't count as genocide because murdering 10 million people wasn't the point in itself, but rather a means to force labor. Besides, they say, most deaths came from plague rather than direct execution.
Meanwhile, Lemkin wrote 120 pages titled Genocide in Belgian Congo because it represented one of the most obvious and egregious cases.4 By his choice of historical examples, Lemkin also demonstrates that genocidal intent doesn't require willfully setting out to exterminate an entire people for its own sake. If your goal is to take territory, wealth, cultural primacy, or something else, and your means of achieving that goal involves destroying the foundations of life for a group, this is genocide. Genocide can be a method as much as a goal—and perhaps it always is. Even the Nazis, as well as contemporary White supremacists, would claim that their goal is a future for the White race, and that the elimination of others is only to secure that goal. Requiring that extermination of a group be explicitly articulated as a goal in itself would return us to an absurd position where practically nothing would ever count, particularly not if perpetrators want to avoid getting caught.
The Belgian Congo example raises another question: the Belgian regime didn’t target any ethnicity or religion in terms meaningful to the Congolese. Rather, it targeted Africans. Some have argued that the Belgian Congo doesn’t count as genocide for this reason: if it’s not targeting a specific ethnicity as understood by those targeted, is it truly aiming to annihilate a people? With the Congo taken by Lemkin as a positive example, yes it does. Perpetrators of genocide will not care about the world as understood by their victims. If the perpetrators are targeting people as Africans or as Black, their lack of knowledge or care for Africans’ own ethnic identities does not exonerate them of genocide.
These positions about aim relate to Lemkin, or at least my reading of him, and are one of the areas where the legal questions become much thornier than the clarity in Lemkin’s scholarly and moral discourse. This serves as as an opportune moment to bridge from the scholarship that invented and grounded the concept of genocide to the international legal language that most of the world ultimately signed into law.
The Genocide Conventions
In its broadest strokes, genocide translated reasonably directly from a concept invented by a Jewish law professor into a term defined by international treaty. But that treaty disappointed Lemkin, because it left out parts of his thought that he deemed important for prevention of future atrocities. Here are the key passages of the Conventions:
Article II: In the present Convention, genocide means any of the following acts committed with 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 or mental harm to members of the group;
(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.
Article III: The following acts shall be punishable:
(a) Genocide;
(b) Conspiracy to commit genocide;
(c) Direct and public incitement to commit genocide;
(d) Attempt to commit genocide;
(e) Complicity in genocide.
The Conventions didn’t start out so narrow! The 1947 draft includes linguistic and political groups among the targets, and it specifies that forced emigration, destruction of books or religious sites, or deprivation of housing or livelihood all constituted acts of genocide. In short, that draft catches the Nazis red-handed in 1933. Unfortunately for the Conventions, that language would also paint the major postwar powers as guilty of their own obvious crimes. To avoid culpability, the Soviet Union got political groups removed from the listed targets,5 while the United States narrowed the Conventions’ focus to physical destruction to avoid scrutiny of its Jim Crow laws, and it switched “purpose” in the 1947 draft to “intent”—which will raise thorny questions.
We thus have a legal definition which focuses on physical destruction of ethnic groups and omits the cultural components which Lemkin viewed both as vital warning signs and as crimes in themselves. Moreover, political groups became valid targets. Lemkin mentioned the Roman destruction of the temple in Jerusalem in 70 CE as an example of genocide that loomed large in his early thought; under the treaty as passed, Emperors Titus and Hadrian would have more wiggle room to argue that they were merely responding appropriately to violent rebels. Modern regimes have predictably taken advantage of this latitude, to the extent that an argument can be made that the Conventions enable rather than discourage mass violence against civilians by providing a legal standard that the unjust need never fear as long as they are not recklessly stupid in crossing certain lines and discussing them openly.
While understanding the Conventions as harmful is an extreme position, the scholarly and legal uses do have a clear divergence. We thus have a division between definitions among social scientists who build from Lemkin and definitions by legal scholars who work with precedent from cases brought before the ICJ. However, these groups are aware of each other and do not operate in isolation, but instead participate in ongoing conversation on how best to read and enforce the Conventions, and on what strategies civil society and governments should pursue.
To the extent that case law sided with the interests of the postwar superpowers, this argument of the Conventions protecting the unjust could hold water. To the extent that case law looks to the essays that created and defined the term genocide for the original public meaning6 of terms used in the Conventions, such as “intent to destroy,” we have an international legal framework which narrows which actions constitute the actual crime, and which might not be as effective at catching Nazis in 1933, but which can still have a preventative function.
Even with the narrower legal definition, you could in theory have a genocide with no casualties. The most obvious means would be by forcible transfer of children. Alternatively, if a regime willfully destroys the foundations of life for a group and the international community intervenes to restore them before casualties mount, genocide still applies. Notice also the inclusion of complicity and incitement among punishable acts. The framers of the Conventions were not stupid. They understood that a government can simply fail to punish genocidal incitement, and that people down the chain will get the memo. This is why incitement to genocide is punishable in itself, and why failure to punish it can be taken as evidence of genocidal intent. Moreover, case law has ruled that “conduct not constituting acts of genocide may be considered when assessing genocidal intent.” So Lemkin’s cultural, political, and economic considerations aren’t componential to the crime, but they do become factors in determining whether the crime has occurred (and remain crimes in their own right under different treaties).
So what about genocidal intent? Intent is hard to prove, whether for individuals or for governments. The most seminal case law requires specific intent, a dolus specialis: deliberately performing the acts forbidden in the Conventions with the specific and conscious aim of genocide. However, legal arguments ever since have made the case for a switch to conditional intent, which would allow for guilty verdicts on negligence or opportunism, and knowledge-based intent, where knowing that the impact of an action will be genocidal and taking it anyway is sufficient, regardless of whether genocide was anyone’s aim. These options clearly follow Lemkin more closely, and we don’t know what standard the ICJ might adopt in future cases.
If you'd asked me before I read about the scholarly and legal underpinnings of genocide what did and did not count, I might have answered based on rough historical comparison to the best-known genocides that came to mind. And I would have been wrong. Hopefully this has been as informative for you as it was for me.
What does this mean for Israel in Gaza?
Obviously, South Africa’s case in the ICJ against Israel’s actions in Gaza prompted this reading. I watched the entirety of both South Africa’s prosecution and Israel’s defense and realized I should know more. So how does what we’ve learned apply?
First, let’s talk about what Israeli violence in Gaza is not. It is not a holocaust. It is not an attempt at direct murder of the entire civilian population. If Israel were dropping bombs to maximize direct civilian casualties, the death toll would not be in the tens of thousands after six months. It would be in the hundreds of thousands, at least. Some argue that Israel’s actions are not genocidal because direct murder of the entire civilian population is clearly not the aim.7
As we’ve seen above, this argument is incorrect. Genocide can be direct mass murder, but more often it is not. So let’s run through the legal criteria one by one:
(a) Killing members of the group
Clearly Israel has killed civilians. Moreover, it has happened disproportionately to other conflicts. More civilians have died in six months of Israel’s war in Gaza than in two years of the Russian invasion of Ukraine, and more children have died in Gaza than in all other global conflicts worldwide in the last four years.8
(b) Causing serious bodily or mental harm to members of the group;
See above; casualties in injuries are also high.
(c) Deliberately inflicting on the group conditions of life calculated to bring about its physical destruction in whole or in part;
With this criterion, the case shifts most strongly from a war run incompetently or cavalierly, which Israel can argue regarding the body count, to a calculated strategy of destruction.
Israel has destroyed at least 60% of homes in Gaza.
It has destroyed vital infrastructure for healthcare.
It has inflicted forced starvation on an entire civilian population.
(d) Imposing measures intended to prevent births within the group;
Destruction of healthcare infrastructure arguably could qualify, along with the denial of food, aid, and resources.
(e) Forcibly transferring children of the group to another group.
I am not aware that Israel has pursued this strategy in Gaza. Instead, its policies seem to have made orphans of many children and to have killed others.
This isn’t an exhaustive list, but when we drop the idea that genocide requires Israel to attempt direct mass murder and focus instead on the text of the law, South Africa’s case becomes clear. However, we’ve left out the hardest to prove component: genocidal intent.
With regards to intent, Israel’s far right has not done the government any favors:
“Erase Gaza from the face of the earth.” “Bomb without distinction.” “No electricity, no food, no water, no fuel – everything is closed. We are fighting human animals and we must act accordingly.” If we side with the argument that the Conventions protect all but the recklessly stupid, Netanyahu’s far-right coalition has blundered into the recklessly stupid.
With South Africa’s logic for presenting the case, if your government has failed to punish incitement, and then performs actions that meet the criteria for genocide, then genocidal intent can be inferred—and indeed must be, or else the Conventions are useless against all but the most profoundly stupid of regimes.
Meanwhile, one of Israel’s key defenses in the ICJ hearings was that regardless of whether several high-ranking government officials may have incited genocide, those are not the officials with control over military strategy, and the IDF demonstrably does not have direct orders for genocide.
However, what’s very difficult to contest from the letter of the law is that the Palestinians’ right to protection from genocide was and is plausibly at risk. This is what the ICJ ruled, and it’s different from ruling whether genocide plausibly already had happened. The justices thus overwhelmingly voted for provisional measures for prevention of genocide, most of which Israel should already have been doing anyway. Hopefully, if that ruling didn’t make sense to you before, it does now. Arguing against the ruling mostly seems to require ignoring the text of the treaty and appealing instead to a sense of outrage.
The court also ruled that South Africa may proceed with its case. If Israel had taken the opportunity to change course when the ICJ first delivered its emergency measures, that case would be weaker. Given that Netanyahu’s regime instead doubled down and created conditions for starvation for nearly the entire Palestinian population in Gaza, the legal case for genocide has legs, even with the high bar of special intent.9
The case may hinge on the connection between incitement and policy, and on whether the open statements of Israel’s governmental authorities plus its collective strategies suffice to infer intent to destroy Palestinians in part. If the ICJ shifts from special intent toward knowledge-based intent, Netanyahu is in bad shape. In this case, his allies will probably claim that Israel is being treated unfairly compared to precedent, when in reality this shift in understanding intent has been decades in the making, and is definitely closer to aim as Lemkin understood it. If the ICJ sticks with special intent, we still might see a conviction, but it will be much more difficult.
Of course, a strong case is not a conviction. There’s plenty of room for Israel and its leadership to avoid a guilty verdict with arguments of incompetence or deferred responsibility. There’s also the question of whether supplying Israel with weapons violates section III-e of the Conventions, Complicity in Genocide.10 We’ll have to wait and find out.
Back to Lemkin and the 1947 Draft
That’s the legal answer with the treaty. Meanwhile, many social scientists work from Lemkin’s definition, rather than from the legal one, and many legal scholars argue for case law to interpret the treaty more in alignment with Lemkin. After reviewing Lemkin’s writings and the path toward the final treaty, I count myself in this camp. Though his language needed refining, Lemkin was basically correct in 1933. The 1947 draft of the Conventions got it right. The concessions made for sake of the powerful gave us a treaty that still enables preventative measures, but less effectively than what we could and should have had.
Here’s the odd part: even our limited treaty might be doing what Lemkin intended. Reportedly, he was willing to see vital parts thrown out as long as we got something, since he saw states holding each other accountable as only one part of what such a treaty should do. Instead, by countries passing domestic laws in accord with with the treaty, he hoped for global publics less willing to tolerate genocide.
This, on the American Left, is what we are seeing right now. From the plain text of the treaty, Israel might be found in violation of the Genocide Conventions, and it almost undeniably meets Lemkin’s definition. But Israel’s leadership has openly denigrated the ICJ ruling, and a final court decision will take years. Simply waiting for an ICJ decision is not what anyone who sincerely opposes genocide should argue for a public to do, ever. Based on this understanding, American protest voters have made clear to Biden that ignoring the United States’ laws against providing arms that are involved in war crimes comes at an electoral cost.
Without Lemkin, and without the Genocide Conventions, we would still be facing a crime without name. Protestors are as much the proximate goal of naming this crime as the Conventions. If protestors and an international legal framework together can prevent genocide in its earlier stages, Lemkin’s ultimate goal will be met.
I’m using preventgenocide.org to link direct quotes from Lemkin where possible, because its lack of updates since 2015 is in this case an advantage. It contains the plain text of important essays and laws, doesn’t seem to be going anywhere, and has minimal connection to current political arguments.
The Nazis even cut a deal with Zionists, not because they were allies in any form, but because the Nazis wanted the Jews out (and also wanted political and economic leverage in Palestine). Nazi policy in the 1930s focused on forcing emigration. Outright mass murder came in the 40s.
The Holocaust killed approximately six million Jews and millions of others. A common misconception is that we have a decent estimate of about 5 million others. In reality, we don’t know, and we never will. Did the Nazis kill half a million Roma, or 1.5 million? The scale boggles the mind. Perpetrators of mass violence, particularly genocide, intentionally cover their tracks and make the scale of death difficult to track.
Lemkin was less effective at writing empathetic accounts of Africans than of other peoples, as he often took the accounts of colonizers and perpetrators at face value in ways that he simply doesn’t for others. The details of the account are thus not his best work as an historian, as he accepts shifting of blame for certain atrocities onto the Africans rather than the Belgians. Even with these errors, he unflinchingly labels the Belgian Congo a genocide.
The focus on national groups is a limitation in the language of genocide that also partially belongs to Lemkin. The Nazis also persecuted trans people and murdered about 200,000 disabled people. Neither of these are a focus in Lemkin, and they definitely are not genocide under international law. We don’t have a ready word for attempting to eradicate people who differ by disability, sexuality, gender, or other differences that aren’t understood in ethnic or religious terms.
Following similar rhetoric to the Nazis, contemporary right-wing movements in the United States and abroad have selected trans people as scapegoats. Outlawing medical care, outlawing trans people from existing in public, outlawing even the discussion of their existence in schools, along with many other measures, clearly correspond to Lemkin’s enumerated strategies of genocide. But because trans people, disabled people, and others are not protected groups under the Genocide Conventions, it doesn’t count. “Trans eradication” can describe the aim of these movements, but “trans genocide” would require a quite liberal reading of Lemkin. We might argue that the legal definition should include plans to annihilate a larger variety of vulnerable groups under the rubric of genocide, but currently, it does not.
A focus on original public meaning is a hallmark of American conservative legal interpretation. It’s not a global norm, and it’s hotly contested in the United States. I disagree with American conservatives on the exclusive correctness of this hermeneutic, and disagree even further over whether the conservative Supreme Court practices it with consistency or intellectual honesty. However, referencing Lemkin’s writings serves as an excellent example for why original public meaning can lead to conclusions that protect more rights than approaches which outright ignore it.
By the same token, some who have heard expert opinions that Israel’s actions amount to genocide have incorrectly concluded that direct mass murder is the goal and that direct comparison to the Holocaust is appropriate. “Genocide is genocide,” runs this reasoning. If reading Lemkin teaches us anything, it is that while all genocide is an abomination, genocides can be quite different from one another in their forms and severity. Comparison is necessary, but facile equation of genocides is potentially harmful.
Some critics like to doubt the Health Ministry’s numbers because it is “Hamas-run.” The health ministry in Gaza is actually co-run with Fatah, with many of its employees reporting to superiors in the West Bank. Also, they follow documented methods for reporting casualties according to identified individuals recorded dead; they don’t just report out whatever numbers they like. (This is also why they can’t just revise down in that much-maligned outlier where their numbers were higher than other estimates). Because casualty numbers do not include thousands of people who are missing because their bodies are most likely buried under rubble, they almost certainly undercount. Attempts to debunk the casualty numbers have been based on willfully deceptive math. A more serious critique of the Health Ministry’s numbers finds that because Israel has rendered most hospitals and morgues non-functional, the Ministry has increasingly relied on media reports rather than certified deaths, which results in decreased capacity to track the deaths of combatants. This means that the total number of deaths is almost certainly higher, but that the proportion of women and children deaths is probably lower—though it would have to be much lower to cease being a majority. As in most conflicts, more precise numbers of dead will be tallied by historians, but there’s no grounded reason to believe that overall casualties are any lower than stated, and many to believe that they are likely higher.
Some readers might notice that I haven’t mentioned Hamas even once. Hamas is irrelevant to the question of whether Israel’s actions in Gaza constitute genocide. Whether Israeli military response in Gaza is warranted, whether Hamas itself has genocidal aims, and whether Hamas also violates international law are all irrelevant to this question. An opponent who violates international law does not suddenly void the need to follow international law. How to prosecute Hamas under international law is its own question. Palestine’s representative to the United Nations has recognized the International Criminal Court, so Hamas leadership is under independent investigation by the ICC.
Nicaragua has raised a case against Germany on these grounds. The ICJ has ruled that it doesn’t currently have the authority or cause to order Germany not to send arms, but it will still hear the case based on its merits. Future charges against the United States and its leadership are possible as well.