Hide table of contents

As I argued here, one away to improve international law and foreign policy would be to promote more recognition of a right to self-determination. This should be of interest to those who are interested in the more general EA cause areas of improving international institutions and reducing S-risks. Self-determination may reduce S-risk because inflexible norms of territorial sovereignty allow autocratic governments to freely inflict mass suffering within their country, as we see and have seen in many times and places in modern history.

Actually fleshing out the details of when self-determination should be granted or denied is a more difficult problem. What I provide here is a general moral framework for the issue. I propose a set of criteria which should count in favor of or against self-determination efforts and comment on how to draw an overall judgment from the various criteria. Along the way I evaluate three case studies of self-determination efforts, to show how to apply the framework in practice and to produce some conclusions for foreign policy and activism. Then I discuss further what we might do to make the international system more accommodating of self-determination. This points the way to more concrete legal and policy changes.

Introduction to the case studies

I evaluate three modern examples which are highly relevant for current policymaking: Artsakh, Crimea, and Taiwan. First, here are basic[1] descriptions of the examples.



During the Soviet era, Nagorno-Karabakh was an autonomous oblast within Azerbaijan - that is, it was formally part of Azerbaijan, but it had a high degree of autonomy to manage its own affairs. The population of Nagorno-Karabakh was mostly ethnically Armenian, with a substantial Azerbaijani minority. Despite their autonomous status, Armenians in Nagorno-Karabakh were dissatisfied living under Azerbaijani authority. They lobbied to unify their territory with nearby Armenia, but these requests were denied by the Soviets. Starting in the late 1980s, as the Soviet Union dissolved and Azerbaijan and Armenia gained their independence, the Armenians of Nagorno-Karabakh pressed harder for unification with Armenia, which was again denied by Azerbaijan. This provoked spiraling violence between Armenians and Azerbaijanis culminating in a ferocious war with many atrocities. The Armenians of Nagorno-Karabakh, fighting with the help of Armenians from Armenia proper, emerged victorious. They pushed out Azerbaijani rule and expelled pretty much all Azerbaijani civilians not only from Nagorno-Karabakh but also from some surrounding Azerbaijani territories. The combined land of Nagorno-Karabakh and these surrounding territories formed the Nagorno-Karabakh Republic.

Throughout the 1990s, 2000s and 2010s, Nagorno-Karabakh existed as an unrecognized republic with de facto independence, supported by Armenia, and fighting periodic skirmishes with Azerbaijan as the latter aimed to recover all of its former land. The Nagorno-Karabakh Republic eventually changed its name to the Republic of Artsakh.

In 2020, Azerbaijan launched a full-scale invasion of Artsakh and defeated the militaries of Armenia and Artsakh. Azerbaijan captured all of the territories surrounding Nagorno-Karabakh and about a third of Nagorno-Karabakh before Russia forced a peace settlement. In the process, Azerbaijan forcibly displaced the (now entirely Armenian) population of these territories.


The remainder of Artsakh still exists as an unrecognized state with de facto independence sitting on two-thirds of the original territory of the Nagorno-Karabakh autonomous oblast. It is an enclave entirely surrounded by Azerbaijan, it is separated from Armenia by just five kilometers in the Lachin Corridor, and it is protected by Russian peacekeepers. It is populated almost entirely by ethnic Armenians.

The territory which Azerbaijan took from Artsakh in 2020 is now populated only by Azerbaijani settlers, some of whom had been displaced during the First Karabakh War.


Common Armenian aspirations are to secure Artsakh from further Azerbaijani encroachment, achieve international recognition of Artsakh's independence or formally unify it with Armenia, and possibly recover some of the territories which were lost in 2020 (though this last goal looks totally implausible).

Common Azerbaijani aspirations are to forestall international recognition of Artsakh and capture the remainder of its territory.

International involvement

Russia has played a major diplomatic role and deployed troops for peacekeeping. Turkey has sent mercenaries and probably drone operators to assist Azerbaijan's war effort. A variety of countries including Russia, Turkey and Israel have sold or given weapons to warring parties.

UN Security Council resolutions in the early 1990s urged both sides to stop fighting and urged Armenia to withdraw from the occupied territories. International rhetoric has typically asked both sides to refrain from warfare and atrocities. None of this rhetoric seems to have been impactful.



Crimea was an autonomous region within Russia during the early Soviet era. In 1954, Soviet leadership transferred Crimea to Ukraine, without Crimean consent, despite the fact that the majority of Crimeans were ethnically Russian. When Ukraine gained independence in 1991, Crimea continued as an autonomous region within the new state, but many Crimeans wished to unify with Russia instead.

Many Ukrainians outside Crimea desired a closer economic relationship with the European Union, even though that could provoke Russia into reducing trade with Ukraine. In 2014, pro-EU Ukrainians overthrew President Yanukovych and established a new government which readily cooperated with the EU. Crimeans tended to oppose this revolution and began to more strongly demand secession of Crimea from Ukraine. Tensions turned to violence. Russia very quickly deployed its military to occupy key sites in Crimea and aid separatist militias. Russian troops seized Crimean government buildings and oversaw an emergency session where the Crimean parliament voted for a new Russia-friendly government and scheduled a referendum on whether Russia could annex Crimea.

The Ukrainian government declared all this to be illegal, but the Russian military prevented them from interfering with events on the peninsula. The referendum proceeded despite the absence of international recognition or reputable election observers. The official result of the referendum was 97% in favor of joining Russia with 83% turnout. Russia subsequently annexed Crimea into its territory.


Crimea is de facto part of Russia, but the international community generally asserts that it lawfully belongs to Ukraine.


Common Russian and Crimean aspirations are to maintain the territorial status quo, gain international recognition for Crimea's new status, and lift the associated sanctions placed on Russia.

A common Ukrainian aspiration is to retake Crimea.

International involvement

The United States and the EU have passed fairly strong economic sanctions on Russia, though this was partly in response to other Russian activities in eastern Ukraine. The UN General Assembly approved a resolution describing the Crimean referendum as illegal.



Japan seized Taiwan from China in 1895. At the conclusion of World War II, the island was given to the Nationalist faction in the Chinese Civil War. When the Communists defeated the Nationalists on Mainland China, the last Nationalists fled to Taiwan and successfully defended it against Communist invasion attempts. The international community slowly abandoned recognition of Nationalist China in favor of recognizing the PRC on the mainland. Taiwan was a dictatorship until it democratized in the 1990s.


Taiwan is de facto independent, but is not recognized as such by the international community.

The People's Republic of China (PRC), which occupies the mainland, continues to insist that it has sovereignty over Taiwan.

(Taiwan also claims sovereignty over mainland China but I won't worry about those claims here.)


A common Chinese aspiration is to integrate Taiwan with the mainland, possibly via military invasion.

Common Taiwanese aspirations are to maintain their de facto independence and get international recognition as a state.

International involvement

The international community generally does not recognize Taiwan. The United States does not recognize Taiwan as a state but supports it with a military alliance.


To decide when to recognize a right to self-determination, we must apply five principles: popular will, possession, governance, law, and territorial sanity.


As I argued here, self-determination fits a basic notion of democracy. If we believe that governments work better when people can select their leaders, we should also believe that they'll work better when people can select their government. The basic logic is not fundamentally different. If people in one region are dissatisfied with the national government, the state should work to build legitimacy and favor, not use legalism or violence to brush grievances aside.

In some cases, a proper referendum reliably reveals people's views. We should pay attention to a referendum even if the ruling power or international community does not recognize its legitimacy. Disputed legal legitimacy does not undermine a referendum's value as a tool for revealing popular sentiment. That said, referendums may be compromised by fraudulent or authoritarian activity.

Opinion polling by reputable organizations can also provide a measure of popular sentiment, although sometimes it's not clear how closely polls will correspond to actual voting preferences. In the US domestic political context, referendum outcomes on policy issues tend to be closer to 50/50 than would be suggested by polling, but secession seems like a matter of stronger political identity where poll results should be more meaningful. Opinion polling is also more resilient to fraud and boycotts.

At minimum, one may try to gather an impression of popular sentiment by looking at anecdotal sources like journalism, online comments, and interviews.

One complication here is that current opinion in a territory conveniently excludes the voices of people who were recently displaced or murdered, and this creates a perverse incentive. For instance, ethnic Abkhazians were previously a minority in their territory, but after committing massive ethnic cleansing of Georgians in the 1990s they became a majority that supported independence. In cases like this, we shouldn't solely judge on the basis of what the current population wants, for two reasons. First, because doing so would create a perverse precedent that the consequences of ethnic cleansing can be legitimized, which is especially bad when ethnic cleansing involves mass murder. Second, because surviving refugees usually have a substantial interest in being able to return to their homelands.

So an ideal referendum should include every refugee. This may not be done in practice, but we can at least speculate on how voting results would change given the presence of refugees. Additionally, in cases of recent democide, we should respect the views of the deceased and/or dismiss the views of the perpetrators for the sake of justice and precedent.

Conversely, people may try to exclude the views of recent immigrants as if they do not count as legitimate, indigenous citizens. This is wrong. One might argue that people's views should be excluded from our normative idea of the genuine popular will if they were brought by governments in a deliberate attempt at settler-colonialism within the last generation, but even then I would be reluctant to take a disenfranchising stance.

Note that the options for self-determination may be more complex than simple inclusion or secession. A compromise option like obtaining the status of a federation or autonomous region may be available. In this case more careful polling or judgment is required to figure out which option is best supported by popular sentiment; it may be a compromise option which is neither highly favored nor strongly rejected by many people. As with elections, plurality voting seems worse than other systems like STAR, approval or ranked choice voting.


The Armenians of Artsakh have traditionally been split on whether they prefer unification with Armenia or independence. Regardless, the main point is that either option is preferred over recapture by Azerbaijan. "There is near unanimous agreement that Nagorno-Karabakh should not return to Azerbaijan," and 88% of survey respondents see Azerbaijan as the main enemy of Armenia. (It's safe to presume that most of the remainder still regard Azerbaijan as a significant enemy of Armenia.) Sentiments against unification with Azerbaijan are likely to be even stronger now after Azerbaijan initiated the Second Nagorno-Karabakh War and Azerbaijani troops were implicated in a number of atrocities.

Although is statistically inevitable in a population of over 100,000, I have not heard of any residents of Artsakh preferring Azerbaijani rule over independence or Armenian rule, nor have I seen evidence of any residents of Artsakh fighting on the side of Azerbaijan. My impression from social media and published media is that opposition to integration with Azerbaijan is a universal assumption that is not even subject to internal debate. This is unsurprising considering that it's not clear whether there is even a single ethnic Azerbaijani still living in Artsakh; in 2005 the census recorded six Azerbaijanis and in 2015 the census did not even include the category.

Note that the main reason for this absence of Azerbaijanis is the history of ethnic cleansing in the territory. Armenians expelled roughly 30,000[2] Azerbaijanis from the current territory of Artsakh during the First Nagorno-Karabakh War, and the survivors - who may number 18,000 or so, just guessing - have still not been granted a right to return to their homes. But with an Armenian population of 120,000, this suggests a true level of 87% support for secession, which is still very high.

Support for secession has also been verified through formal expressions of the popular will. A petition in Nagorno-Karabakh during 1987-1988 (before any serious population displacement) gained signatures equaling 55% of the adult population - a very high figure for a petition. Then in 1991 a referendum showed 99.9% support for independence with 82% turnout (Azerbaijanis boycotted the referendum); I'm not sure how many Azerbaijanis had been displaced by this time, but with 108,000 "yes" votes, there cannot be enough Azerbaijani refugees to overcome the large majority in favor of secession.

Khojaly is an exception; while the current population of 1,000+ Armenians is presumably entirely in favor of secession, the original population was majority Azerbaijani, and hundreds of these civilians were killed by Armenians in the 1992 Khojaly Massacre.

Things get more complicated outside the current territory of Artsakh. First, there are regions which originally belonged to Nagorno-Karabakh but were taken by Azerbaijan in 2020. Today, these areas are solely populated by recent Azerbaijani migrants (some of whom may be refugees from the early 1990s) and are presumably unanimously in favor of Azerbaijani rule. However, the true popular will should encompass both Armenians who were displaced in 2020 and all the Azerbaijanis who were displaced in the early 1990s. By this standard, according to the demographics, petition and referendum results of the time, the mostly-Armenian population has similarly expressed a will for secession. An exception to this is the town of Shushi/a, which had a large Azerbaijani majority prior to forced displacement in the First Karabakh War.

Shahumyan Province outside of Nagorno-Karabakh is controlled by Azerbaijan, and today there are just some Azerbaijanis living there. However, it had a strong Armenian majority in the late 1980s, which was later forcibly displaced. Shahumyan declared itself part of the new Nagorno-Karabakh Republic in 1991, so public opinion there was presumably strongly in favor of seceding from Azerbaijan. I cannot find a source to say whether Shahumyan participated in the 1987-88 petition or the 1991 referendum.

In other regions of Azerbaijan which were occupied by Armenian forces in the early 1990s, the population today is entirely comprised of recent migrants from Azerbaijan (some of whom are presumably refugees from the early 1990s). This region has suffered ethnic cleansing of the relatively small number of Armenians living there in 2020 plus the quite large number of Azerbaijanis living there in the early 1990s. A fair understanding of popular will would include the voices of recently displaced or killed Armenians, but it should similarly include the far more numerous voices of Azerbaijanis who were displaced or killed in the first war. So the popular will from these territories is to remain with Azerbaijan. An exception to this is the town of Hadrut, which always had a large Armenian majority until it was ethnically cleansed in the Second Karabakh War.

In conclusion, Artsakh with its current territory is probably more unanimously in favor of secession than any comparable territory anywhere in the world. This is weakened a little by incorporating the preferences of the expelled Azerbaijani minority, but overall secessionist sentiment should still be considered well over 80%. Khojaly is a town where popular will is ambiguous depending on how one compares the current population versus the historically displaced/killed population.

Territories which were originally part of Nagorno-Karabakh or Shahumyan Province, but captured by Azerbaijan in 2020, have current sentiment in favor of Azerbaijani rule but originally had large majorities in favor of secession prior to the rounds of ethnic cleansing. Therefore the will of their populations should be understood as supporting secession. Shushi/a is an exception where things are simply ambiguous.

Finally, in other territories which were captured by Armenians in the early 1990s but regained by Azerbaijan in 2020, both the small current populations and the combined displaced populations have strongly opposed secession. An exception to this is the town of Hadrut whose people strongly prefer secession.


Crimea's 2014 status referendum provides one indicator of public opinion. Combining the results from the main referendum and the separate Sevastopol referendum shows that 97.3% of Crimeans voted for unification with Russia, with 83% turnout. However, the only alternative on the ballot was to restore the 1992 constitution, which may have been perceived as de facto independence for Crimea. Because of this problem, the many irregularities with the execution of the referendum suggesting the pro-Russian side could have been artificially exaggerated,[3] and the fact that many years have passed since 2014 and opinions may have changed, this is not a good indicator for Crimean public opinion.

A number of opinion polls offer more reliable evidence on Crimean public opinion. Unfortunately, all are flawed in some way.

Ideally, referenda and polls should offer multiple options with an approval or ranked-choice format. A 2008 Razumkov survey separately asked respondents to answer yes/no/unsure to each of a variety of options, so the "yes" totals may be interpreted as a score in approval voting; on this poll unification with Russia was the most popular option in both Crimea (64%) and Sevastopol (72%). This conclusion remains true if half the "unsure" votes are included as approvals. Additionally, a 2009 GfK poll asked Crimeans to pick their favorite and second-favorite out of four options; unification with Russia earned 54% of first-choice votes and 14% of second-choice votes, clearly beating the second-favored option of a fully independent Crimean state which had 19% of first-choice votes and 35% of second-choice votes. A 2009 Razumkov survey offered plurality voting over a range of options for Crimea's political status, and found that unification with Russia was most popular here with 32% support, with Russian national autonomy in Ukraine coming in second place with 20% support. These polls suggested that the status quo was one of the least popular options. But they are too old to carry much weight.

Other polls have asked respondents to simply choose between joining Russia or maintaining Crimean autonomy in Ukraine. A series of UNDP polls from 2009 to 2011 found that 65-70% favored unification with Russia and 10-15% opposed it. But a 2011 IRI survey using face-to-face home interviews found that 49% of Crimeans preferred autonomy in Ukraine compared to 33% preferring annexation by Russia, and in 2013 they found that the gap had widened to 53%-23%. Still, these polls are too old to rely on.

The dubious architecture of the 2014 referendum, asking people whether they wanted to join Russia or to "restore the 1992 constitution", was used by several polls. A 2014 Crimean Republican Institute of Political and Sociological Research poll found 77% of Crimeans and 85% of Sevastopolians in favor of unification, with only 8% of Crimeans (and I don't know how many Sevastopolians) favoring restoration. The 2014 GfK poll found 71% favoring reunification and 11% favoring restoration. Finally, polls by the Russian state-owned (hence, probably pro-Russian biased) entity VTsIOM asking people how they would vote in a repeat referendum found that support for reunification was 90% to 5% in 2015 and 89% to 3% in 2019.

Other polls since the 2014 referendum have asked Crimeans whether they endorse the 2014 referendum result. This is not very helpful because it is more of an approval rating where people give their opinion on the decisions of the Crimean people and government, potentially leading to excessively high favorability for unification. For whatever it's worth, a 2015 GfK poll found that 82% of Crimeans fully endorsed the results of the referendum and 11% mostly endorsed the results, and Levada Center polls found that approval for joining Russia was 86% in 2014 and 82% in 2019.

Some polls asked Crimeans whether they would like all of Ukraine to merge with Russia, but this is just a different question entirely.

To summarize, polls have generally suggested that unification with Russia is the most popular option, except for the IRI surveys in 2011 and 2013 which found that more preferred autonomy within Ukraine. I'm not sure how to account for those. IRI did face-to-face interviews, which seems very reliable, but with a 61% response rate maybe there was serious nonresponse bias. In any case, maybe opinions changed after 2013.

Overall, I think it is reasonably apparent that a large majority of Crimeans prefer unification with Russia over any other option.


87% of Taiwanese surveyed in 1999 opposed reunification under the "One Country, Two Systems" model, and 73% of Taiwanese surveyed in 2000 would oppose reunification even if China made a more generous offer than was given to Hong Kong and Macau. Wikipedia cites a broken page which apparently contained information that this was still true in 2017. The recent suppression of political freedom in Hong Kong and the concentration camps in Xinjiang have probably made Taiwanese people even more opposed to reunification with China.

Taiwan is a full democracy, meaning that election results can be used as a good gauge of the popular will. And the parties which support unification under the PRC have very little electoral success, with the Labor Party and the For Public Good Party controlling none of the 108 national seats, none of the 22 local leadership seats and just 3 out of 675 local councillor seats.

Even those Taiwanese who support reunification may do so for pragmatic reasons - that is, they may believe it would be better to give up rather than risk a war. While their opinion may be correct, it still backs up the idea that reunification under the PRC is not ultimately desired by the people.

Overall, Taiwanese public opinion overwhelmingly opposes reunification with PRC.



Generally speaking, it's better to maintain the de facto status quo rather than introduce a new arrangement. In the best case, changing territorial arrangements involves some disruption of organizations and livelihoods (even Brexit, which involved separation from the EU not a change in actual sovereignty, was rather disruptive); in the worst case, it involves ferocious conflict and war crimes.

Sometimes the territorial situation is unlikely to change at all. In these cases, involved parties can benefit by accepting the status quo and allowing the territory in question to have the ordinary rights and privileges of trade, travel, and international institutions.

This is basically just the uti possidetis principle of international law, but applied as a pro tanto principle to protect unrecognized states and unrecognized borders.


Artsakh has de facto independence from Azerbaijan, and Armenians in both Artsakh and Armenia proper are very strongly opposed to Azerbaijani annexation. Artsakh is also currently protected by Russian peacekeepers. Therefore, by far the most obvious and likely way for the territorial situation to change would be another successful Azerbaijani invasion following a withdrawal of Russian peacekeepers. Despite having lost in 2020, it is hard to imagine that Armenians would simply give up, so we would see another bloody war causing thousands or tens of thousands of deaths and possibly tens of thousands permanently displaced.[4] Such a war would inevitably involve miserable atrocities and would provoke further hatred in the region.

It is possible that Azerbaijan will capture Artsakh without a war at all. If the Armenian military is very weak (as it is right now), and Russia withdraws its peacekeepers, then Azerbaijan (possibly with the help of Turkey or even Russia) might just coerce Armenians into giving up. But even in this 'peaceful' scenario, many Armenians would still flee out of fear of racist actions by the Azerbaijani state and society, and those who remained could suffer mass atrocities by the Azerbaijani military.

Side effects of Azerbaijani advances would be mild strengthening of the Azerbaijani government's domestic position (possibly a bad thing, considering that they are a dictatorship) and domestic political turmoil in Armenia (like what happened at the end of 2020, but worse; probably a bad thing).

Unless there is some serious change in Armenian and Azerbaijani military capabilities, Armenian conquest of Azerbaijani territory is implausible even if Russian peacekeepers withdraw. Theoretically, Russia might act on the side of Armenia and force Azerbaijan to give up some territory, but it seems exceedingly unlikely that Russia will want to do that.

If Russian peacekeepers remain in the region, any real attempt at territorial conquest would provoke conflict with Russia. If Azerbaijan or Armenia does this, they will obviously lose. If a large foreign power like the United States provokes this, they will risk nuclear escalation for no good strategic reason. Either way it is out of the question.

Overall, the solution which would best suit the principle of possession would be for Artsakh with its current territory to become recognized as independent.


Crimea is integrated with Russia for all practical intents and purposes. Returning it to Ukrainian control would require a full-scale war against Russia, with accompanying risks of nuclear escalation, against strong local resistance by the Crimeans. For this reason it is simply out of the question, and the status quo will continue whether we like it or not. So clearly, the solution which would best suit the principle of possession would be to recognize Russian ownership of Crimea.


Taiwan is de facto independent, and unlikely to agree to unification under the PRC. Strong Chinese pressure may force their hand, leading to a peaceful but potentially rough process of annexation.

Alternatively, a Chinese military invasion is a distinct possibility, despite the accompanying risk of nuclear escalation involving the United States.

Clearly, the solution which would best suit the principle of possession would be to recognize Taiwan's sovereignty over its island.



Some governments are better than others. In extreme but not uncommon cases, governments threaten massive atrocities against ethnic or religious minorities in their territory. We can look directly at the merits and demerits of competing national governments and see which would do a better job.

To some extent, we may wish to generally reward good governments with more territory and punish bad governments with less territory. But the main criterion here is assuring good arrangements for the population that is seeking self-determination. So we should compare the ways that the different governments treat this particular population.

Of course, we should not dig up historical issues which are no longer meaningful. We should make predictions for the future while being aware of local conditions and needs in the specific territory which aspires to self-determination.


Freedom House gives Artsakh (labelled Nagorno-Karabakh) a global freedom score of 35 compared to 10 for Azerbaijan. Artsakh is not evaluated by democracy indices, but its government does function partly democratically (journalist Neil Hauer thinks that actually it's the only reasonably functional democracy in the entire Caucasus), whereas Azerbaijan is strongly authoritarian.

Because a common aspiration in Artsakh is to unify with Armenia, we should also compare Armenian to Azerbaijani governance. Freedom House gives Armenia a global freedom score of 55 compared to 10 for Azerbaijan, an internet freedom score of 75 compared to 38 for Azerbaijan, and a democracy score of 33 compared to 1 for Azerbaijan. The Democracy Index rates Armenia 5.35 compared to 2.68 for Azerbaijan. The Heritage Index of Economic Freedom rates Armenia 71.9 compared to 70.1 for Azerbaijan. The World Press Freedom Ranking rates Armenia 28.8 compared to 58.7 for Azerbaijan (lower score is better here). Finally, the Good Country Index - which I would recommend taking with a very large grain of salt - gives Armenia a rank of 50th in the world, and Azerbaijan an inferior rank of 113th.

Note that the Azerbaijani regime has substantial oil wealth which can be funneled into economic development, and thus may seem like a better regime for improving local economic conditions, but this doesn't count because it simply involves a transfer of wealth which would otherwise be used for the benefit of people in Azerbaijan proper.

Overall, on every available metric except economic freedom (where they are about equal), Armenian governance appears generally superior to Azerbaijani governance.

When looking at the specific treatment of the people in question, this disparity grows even stronger. Specifically, the Armenians in Nagorno-Karabakh would suffer dramatically lower social, cultural and political status compared to other residents of Azerbaijan due to severe racism which pervades all levels of Azerbaijani society and creates a consistent pattern of wartime atrocities and culture eradication. In fact, the Azerbaijani state has acted so criminally towards Armenians that Azerbaijan is essentially disqualified from having a morally legitimate claim to sovereignty over them.[5]

Azerbaijani rule in Nagorno-Karabakh might be less malign if they grant autonomous status to the territory. But Aliyev has declared that he will never allow this. Even if Azerbaijan does make such a promise, meaningful autonomy seems unlikely considering the long track record of broken promises and vicious behavior by the regime. The track records of autonomy in Xinjiang and Hong Kong give poignant examples of how autonomy promised on paper can give little protection against an authoritarian regime. And even true autonomous status would still be clearly inferior to independence or unification with Armenia.

To summarize, the residents of Artsakh would have a grim future under Azerbaijani rule. Assuming that the population is not ethnically cleansed, they will face the double scourge of authoritarianism and severe anti-Armenian racism in Azerbaijani society.

Similarly, while governance by Armenia and Artsakh is generally superior to Azerbaijani governance, the benefits may not extend to Azerbaijani residents. Armenian troops have repeatedly committed wartime atrocities against Azerbaijanis,[6] which Armenians often deny or deflect with whataboutism, and anti-Azerbaijani prejudice seems commonplace among Armenians.


Freedom House gives Ukraine a global freedom score of 60 compared to 20 for Russia, an internet freedom score of 61 compared to 30 for Russia, and a global freedom score of 39 compared to 7 for Russia. The Democracy Index rates Ukraine 5.81 compared to 3.31 for Russia. The Heritage Index of Economic Freedom rates Ukraine 56.2 compared to 61.5 for Russia. The World Press Freedom Ranking rates Ukraine 33.0 compared to 48.7 (lower score is better here). So Ukraine is much better on democracy and civic freedom, whereas Russia is slightly better on economic freedom. That being said, the Good Country Index - which I would recommend taking with a very large grain of salt - gives Russia a rank of 47th in the world, and Ukraine an inferior rank of 72nd.

Ukrainian governance may be worse in Crimea, where ethnic Russians have long felt marginalized. Residents of Crimea tend to feel that their lives are better under Russian rule, and while this may be an irrational outgrowth of political ideology, it may simply reflect genuine experience and judgment. Russia could be more accommodating to local culture or a better guarantor of economic security.

One cause of Crimean preference for Russian governance seems to have been that Russia threatened to withdraw trade from Ukraine as punishment for associating with the EU. This is a morally thorny issue - it's unfair and sets a bad precedent to reward Russia for this kind of harmful behavior, but at the same time it's unfair to force Crimeans to accept the costs of a strategy to punish Russia for such a move.

Overall, it's really not clear whether governance in Crimea would be better under Ukrainian or Russian rule.


Freedom House gives Taiwan a global freedom score of 94 compared to 9 for China. The Democracy Index rates Taiwan 8.54 compared to 2.27 for China. The Heritage Index of Economic Freedom rates Taiwan 78.6 compared to 58.4 for China. The World Press Freedom Ranking rates Taiwan 23.9 compared to 78.7 for China (lower score is better here). Note that Chinese authoritarianism includes a concentration camp system in Xinjiang replete with abuse and forced sterilization for the Uyghur minority.

In practice, a "One Country, Two Systems" type arrangement would allow Taiwan to keep some of its superior governance. However, as Hong Kong has demonstrated, such an arrangement would face a high risk of degradation towards the kind of authoritarian regime found on the mainland.

Overall, Taiwanese governance is far superior to Chinese governance.



It is bizarrely common for people to judge the legal status of a self-determination effort by simply looking at whether the international community recognizes disputed states and borders. This is rather silly because the very thing that we are debating here is whether the international community should recognize certain disputed states and borders. The mere fact that the international community treats things a certain way does not tell us whether they are legally correct in doing so. Moreover, we can expect the international community - being made up entirely of established, recognized states - to be biased against aspirations for self-determination.

A slightly better step is to look at whether self-determination is consistent with the law. This involves both the laws of the parent country and international law.

Neither standard can be given much normative weight. We can expect states to be biased against self-determination movements, especially those within their own countries, just as we would expect men in the early 1900s to be biased against female suffrage, landowners in the 1700s to be biased against universal male suffrage, and so on. Moreover, if new governments could only exist with the permission of previous governments, we'd all still be living in monarchies. Some process for radical change against the will of incumbent power structures must be allowed; that is both common sense and a lesson that we can learn from history (see the American Revolution or the World War II liberation of Europe for famous cases where Americans established new governments without the consent of old governments). Finally, conditionalizing legitimacy on permission from the incumbent government means that self-determination movements will have an easier time obtaining legitimacy in fair, liberal countries than in spiteful authoritarian ones, which is perverse.

International law should arguably be obeyed as a matter of precedent. If countries start flouting international law, then future laws won't be respected, and the whole international system could devolve into an even more chaotic form of anarchy. The first problem with this argument is that it's just not very strong, by analogy with the generally-agreed virtue of illegal efforts to secure democracy and liberal rights; some things are just more important than law and precedent. Additionally, the argument only really applies to signed treaties. If states make an agreement to crush movements for self-determination, then they'll have to follow it or else their other promises will no longer be as trustworthy in the future. But when we are talking about customary international law instead of treaties, countries haven't explicitly agreed to it, and it is broadly accepted that the customs can change. So there is no general penalty to trustworthiness if some states start acting in ways which violate existing customary international law.

Overall, laws should not be considered a very strong principle here. But just for the sake of not being too radical, and encouraging the use of legal channels whenever they happen to not be hideously authoritarian, we should still consider existing law as a pro tanto argument for the legitimacy of a self-determinative move.


The legal status of Nagorno-Karabakh can be evaluated through two lenses, that of secession law in the Soviet Union, or international law on self-determination.

The question of whether Nagorno-Karabakh lawfully seceded from Azerbaijan according to Soviet law is a theoretically straightforward legal question and is addressed quite well by Krüger (2010), chapter A section III, where he concludes that Nagorno-Karabakh remained part of Azerbaijan when it achieved independence. He addresses all the arguments that I have seen from the Armenian side. However there are still some flaws in his analysis and I have sketched a theory which suggests that Nagorno-Karabakh did legally secede under Soviet law.[7]

The question of whether Nagorno-Karabakh lawfully seceded according to international laws on the principle of self-determination seems more philosophical and subjective. Krüger (2010), chapter A section IV, determines that Nagorno-Karabakh did not and does not have the right to secede under international law, but his argument in this section is both more flawed and more dependent on subjective assumptions.[8]

So under both frameworks, it appears ambiguous whether Nagorno-Karabakh has legally exercised a right of secession from Azerbaijan, although at the end of the day, especially considering what seems to be the majority opinion in scholarly literature and the findings of (admittedly biased) international institutions, it seems there is a preponderance of evidence in favor of the Azerbaijani position.

For Shahumyan Province, it clearly had no legal right to secede under Soviet law, but one might argue that it belongs to Artsakh according to international law on self-determination.

As for the other areas outside of Nagorno-Karabakh, there is zero legal ambiguity that they legally belong to Azerbaijan.


Podolian (2018) argues that the decision to hold a referendum in Crimea violated Ukrainian law, that the Crimean parliamentary session which made this decision was compromised and lacking in legal legitimacy, that the referendum was compromised and lacking in legal legitimacy, and that Russian annexation of Crimea violated Russian-Ukrainian treaties as well as broader international law. Merezhko (2015) and Zadorozhnii (2016) more thoroughly dismantle the legal theories advanced by Russian scholars in favor of Crimean independence and Russian annexation. Legal scholars outside of Russia seem to have a consensus that the actions were illegal.


Wikipedia has a list of arguments on the legal status of Taiwan. The case for PRC sovereignty over Taiwan requires both that Taiwan belonged to the ROC after World War II and that international recognition of the PRC as the sole government of China entails that the PRC is legally sovereign over all Chinese territory. I'm not sure whether Taiwan is legally independent, but even if it isn't, the PRC's case is clearly inadequate. As I described previously, merely looking up whether the international community recognizes something provides no genuine legal guidance. The international community could be disregarding the law. And given the absence of substantive legal arguments for PRC sovereignty over Taiwan, that indeed seems to be the case.

Territorial sanity


International law in practice has traditionally given top priority to the principle of territorial integrity - the idea that it is illegitimate to use force to change state borders.

Note that territorial integrity is not a positive affirmation of states' legal right to violently enforce their existing borders. Therefore, it doesn't mean that self-determination should necessarily be denied. That said, in practice, it has generally served as a legal foundation to ignore or crush even the most sensible self-determination efforts. It also inherently conflicts with the possibility of forcefully protecting self-determination.

Territorial integrity does not make a lot of sense in terms of philosophical principles. Merely because borders were established one way does not mean they should stay that way.[9] One argument for territorial integrity is that it gives a legal barrier to warfare, but this is not compelling; war for any reason was outlawed by the Pact of Paris and eliminating the principle of territorial integrity will not change that. Additionally, territorial integrity can serve as a justification for waging new wars for revanchism and for conquering so-called "rogue provinces". Getting rid of territorial integrity without putting anything in its place could reduce legal pressure against warfare, but if we replaced it with stronger ideas of self-determination then we would have similarly strong legal norms regulating the use of force. We would just be changing from saying that it's illegitimate to use force to conquer foreign territory, to saying that it's illegitimate to use force to conquer territory that doesn't want to be conquered and illegitimate to use force to defend territory that doesn't want to be defended.

A common argument on this topic is that weakening the principle of territorial integrity would give fuel to new violent conflcts. Critics point to the many cases of wars of self-determination, such as those which followed the collapse of the Soviet Union, as if they show that granting a right to self-determination leads to interminable conflict. This however is nonsensical because none of these self-determination movements actually received international recognition. The wars and atrocities that we have seen in places like the Balkans and the Caucasus are not the potential consequences of a hypothetical international system founded upon self-determination, they are the actual consequences of the current international system which mostly denies self-determination. Defenders of the existing international system with its fetish for territorial integrity must own up to the observable reality that multitudinous cases of mass murder and genocide have occurred under their preferred system, rather than shifting blame for these events onto aspirational reformers who wield no genuine power.

Of course, the counterfactual is uncertain. Defenders of territorial integrity can say that atrocities are inevitable but that opening up self-determination would make things even worse. They could even argue that we just aren't going far enough; they could argue that the idea of self-determination should be scrubbed from international law so as to minimize ambiguity. Perhaps if we did that, then the subjects of every oppressive regime would clearly understand that they should not even think about separatism, and thus internal conflicts would be averted. But it is highly questionable that things would actually work out this way.

The common argument by defenders of territorial integrity is that greater allowance for self-determination would lead to more violent separatism, and as evidence for this claim they point out that self-determination movements frequently cite international documents promoting self-determination and the precedent of Kosovo to bolster their claims. But just because they do this doesn't mean it has any causal impact on their decision to engage in armed separatism. In reality, people fight these wars for immediate political and cultural reasons, then make up whatever arguments will best help them gain the sympathy that they want on the international stage. Self-determination movements lobby for international recognition (because of course why wouldn't they), but their strategizing does not assume that they will get it. In most cases their main strategy is to enforce realities through military force on the ground, presumably because they accurately recognize that it is their only viable strategy. And in cases where the main strategy for self-determination is to gain political recognition, such as in Catalonia or Scotland or West Papua, the dispute does not spiral into vicious warfare. In places like Nagorno-Karabakh, demands for self-determination start through peaceful political channels, and choices by authorities to reject such demands precipitate vicious conflict. So it seems likely that recognizing self-determination more broadly would lead to less conflict.

Additionally, even if recognizing self-determination does lead to an increase in violent separatism, it is possible that the benefits of actually achieving self-determination tend to outweigh the harms.

All that being said, abandoning the notion of territorial integrity without introducing any serious replacement could lead to certain odd problems.

For one thing, if just any group can declare secession, then we could see the emergence of excessively gerrymandered polities. Of course, to some extent it makes sense for borders to be drawn around ethnic or ideological groups, because that maximizes the extent to which people get to live in their preferred country. But doing this too much could introduce serious practical complications. For instance, what if all the Muslims of Paris, comprising 10% of the city population, draw a squiggly map around their communities in a manner resembling Baarle-Hertog, then vote for secession? This would substantially complicate Parisian life and city services.

People also might form states in order to shrug off obligations to outsiders. For instance, wealthy regions might secede in order to avoid the tax burden of a redistributive state, or people with nefarious intentions might form a new state to circumvent international treaties.

Finally, many people would certainly like to form tiny states because of hobbyism or silly ideological reasons, as we see with various micronations.

Note that the more that states fragment and the more that their borders get gerrymandered, the more the total length of national borders will increase. This would unfortunately raise more obstacles against travel, trade and immigration around the world. However, countries could adapt to this by relying more on supranational organizations to fulfill these functions. We already see this with groups like the EU, and it is a straightforward component of humanity's trend towards neomedieval civilization where we stop assuming that all governance functions must be controlled at the level of the sovereign state. Likewise, the problem of wealthy areas seceding from redistributive states could be ameliorated by global minimum taxation, an idea which is gaining political notability. Still, we cannot rely on these options because they are yet to be properly realized.

Ultimately, we should throw out territorial integrity, but introduce a new notion of territorial sanity to address these issues. Territorial sanity just means that we should not draw borders in really problematic ways. Enclaves, exclaves, and very squiggly borders should be minimized. Economic gerrymandering should be avoided. And states should be large enough to support the basic functions befitting their political status, including reliably interfacing with the international community. But territorial sanity is not inviolable if local context requires a really strange arrangement.


Artsakh is an enclave within Azerbaijan, so granting its secession does complicate borders.

Artsakh is a relatively small and poor country, but is still large enough to fulfill the basic functions of a state. A number of island nations have both smaller populations and smaller economies compared to Artsakh. Artsakh's reliance on Armenia is a product of the blockade and military encroachment by Azerbaijan rather than an intrinsic feature. Regardless, if Artsakh unifies with Armenia then it does not need to stand as an independent country anyway.

An arrangement for guaranteeing traffic along the Lachin Corridor will have to be maintained. Currently, the corridor belongs to Azerbaijan and Russian peacekeepers protect traffic according to the 2020 settlement. In the long run, sovereignty over the corridor could be transferred to Artsakh for the sake of convenience and security, or a joint arrangement could be maintained.


Crimea has very little land border with Ukraine, and its transfer from Ukraine to Russia doesn't complicate borders in any major way.


Taiwan is a large island nation and its independence poses no border problems whatsoever.

Overall judgment


After considering the various criteria of respecting the desires of locals, respecting possession, promoting good governance, following existing laws, and maintaining territorial sanity, we must make an overall judgment for resolving cases of disputed sovereignty.

Popular will should generally be considered a necessary condition for changes in sovereignty. If egregious human rights violations are being perpetrated by the state against a local minority, then secession might be justified under the notion of governance even if a majority of locals oppose it, but there are probably better remedies here like international intervention to protect minority rights.

Popular will is not sufficient to justify a change in sovereignty; it may be defeated by other factors. Law should count for a few percentage points of popular will, so if 51% of the people support an illegal secession then we should not recognize it. Other factors may be important or negligible or anything in between depending on the specific circumstances. It is impossible to specify an adequate theory to guide these tradeoffs; instead we need to go through case studies as a process to get a handle on which solutions tend to work best.

Note that there are compromise solutions for tough dilemmas on self-determination. Minority regions are often granted autonomous status. The more radical status of the Aland Islands involves far-reaching autonomy and demilitarization and has been suggested as a precedent for Nagorno-Karabakh. One problem with these solutions, as evidenced by the degradation of China's "One Country, Two Systems" policy of autonomy for Hong Kong, is that the state may crack down on the territory anyway and the international community will typically be unwilling or unable to stop it.

Assuming that the state respects autonomous status, it can be a good compromise, but it will remain only a compromise. That is, it can achieve many of the benefits of self-determination, but also sacrifices many of the (alleged) benefits of upholding territorial integrity. For instance, if the problem with recognizing secession is that it creates too much incentive for armed separatism, then recognition of radical autonomy could similarly create too much incentive for armed insurrection, with locals threatening war if they do not get the radically autonomous status that they want.

Therefore, compromise policies should be used prudently for situations where both sides have good cases and a compromise is really merited, not used as the default for every case where self-determination conflicts with territorial integrity.


The population of Artsakh is nearly unanimously opposed to unification with Azerbaijan. And they have good reason to do so, for not only is Azerbaijan a strictly authoritarian state, but it is deeply racist against the Armenian population of Artsakh. Azerbaijani capture of Artsakh would probably involve intense warfare, widespread atrocities, ethnic cleansing, and destruction of cultural heritage. These points strongly outweigh the minor issues of the complexity of drawing a border around the Artsakh and the arguable illegality of Artsakh's secession. Therefore, Artsakh must be recognized by the international community as a sovereign state (at which point its residents may freely and legally join Armenia if they so choose). A referendum is unnecessary because the original referendum in 1991 should be accepted as legitimate and because there are no meaningful doubts about the current popular will in Artsakh. As an exception, the town of Khojaly maybe ought to belong to Azerbaijan considering its historical Azerbaijani majority and its symbolic importance as the site of a massacre of Azerbaijani civilians, but a handover to Azerbaijan cannot be accepted considering the likely implications for the current Armenian residents. Some kind of international city or split sovereignty status seems appropriate.

Artsakh also claims a much larger amount of territory which Azerbaijan captured in the 2020 war. In parts of this territory, including the town of Hadrut and most of Nagorno-Karabakh, the population has been mostly Armenian and presumably mostly in favor of secession since the late Soviet era, but the population was ethnically cleansed by Azerbaijan during the 2020 war. Ideally, these territories should be handed back to Artsakh in spite of the views of the relatively small number of recent Azerbaijani settlers and returning refugees. The transfer might be formally legitimated with a kind of referendum of all displaced persons. But given that no state is willing and able to coerce Azerbaijani here, this is not a realistic possibility.

Other former areas of Artsakh including Shushi/a are probably settled and sought by more Azerbaijani refugees than Armenian ones, considering that they were majority Azerbaijani during the late Soviet era. Ideally, a referendum including all displaced persons might determine the appropriate status for these territories. But a return of territory to Artsakh is fully unrealistic due to Azerbaijani political resolve and military superiority, absent a serious change of pace by the international community. Shushi/a warrants special "international city" status considering its fairly high number of Armenian refugees and its strong cultural heritage and meaning to both sides, which might be achievable through serious cajoling or coercion of Azerbaijan (perhaps in a deal also involving Khojaly).

The only area where we can seriously entertain the transfer of Azerbaijani territory to Artsakh is the Lachin Corridor. Transferring the five-kilometer roadway to Artsakh would provide a political guarantee of access between Artsakh and Armenia proper, meaning that Azerbaijan could not blockade Artsakh without being guilty of interstate aggression and thereby incurring nontrivial international attention. Still, achieving this transfer of territory would be very difficult. Guarantees for Armenian access on the road, similar to the guarantee for Azerbaijani access on the road through Syunik to Nakhchivan, might be adequate (although the requirement for access to Artsakh is much more important than the requirement for access to Nakhchivan, given that the latter borders Iran and Turkey and has a safe airport).

To minimize the risk of another Azerbaijani invasion of Artsakh, Russia must continue its peacekeeping mission indefinitely until there is no longer a risk of another war. To punish the Azerbaijani invasion, minimize the risk of another one, and increase pressure towards normalization, countries such as the United States should cease military assistance to Azerbaijan.

In addition to all this, parties should take measures of decency such as recognizing the right to return for all refugees, compensating victims of atrocities, prosecuting their own war criminals, releasing prisoners of war, turning over minefield maps, and so on.


Crimean public opinion seems fairly strongly in favor of remaining part of Russia. While Russian governance is generally inferior to Ukrainian governance, they may be competitive or even superior at managing Crimea in particular. Russian possession of Crimea is firmly entrenched and does not complicate international borders. The problem with Crimea's status is that it clearly violates both international and Ukrainian law, but this alone is not a good reason to overrule self-determination.

Ultimately, Russian sovereignty over Crimea should be recognized, but for the sake of precedent and certainty the international community should require a proper referendum first. This should involve good ballot options, reputable observers and no interference by local militias or the Russian military. Of course, if this referendum shows a majority preference against Russian unification then the international community should continue to call Russian presence an illegal occupation, but I don't think anyone would really expect such a voting outcome. For that reason, Russia and Crimea should agree to the second referendum, knowing they have nothing to lose except for the minor political problem of implicitly acknowledging that the first referendum was illegitimate.

Recognition of Crimea's new status would entail (partial, at least) relaxation of sanctions against Russia, thus leading to economic growth. It would also help alleviate tensions between Russia and the US/Europe.


The population of Taiwan is firmly against unification under the People's Republic of China. This is for good reason, for they have a clearly superior government and have strongly established possession of their territory. Taiwan's borders are very simple because it is an island, and its legal claim in this dispute is much stronger than that of the PRC. There is no principled reason to oppose Taiwanese self-determination.

That said, it may or may not be wise for the international community to recognize Taiwan's sovereignty over the island, depending on the extent to which it could provoke China into hostile behavior.

Unfortunately, the PRC may invade Taiwan and forcefully incorporate it. Deterring such a war would clearly be a good move, and if the PRC does invade then it would clearly be preferable for Taiwan to win. It may or may not be a good idea for the international community to offer military support to the defense of Taiwan, depending on other issues like the risk of nuclear escalation.

Self-determination more broadly

I chose Artsakh, Crimea and Taiwan as examples because these cases are relatively straightforward cases for self-determination and are highly relevant to current policy. Furthermore, by showing that the international community has wrongfully neglected self-determination in all three of these case studies, and that existing international law wrongfully prescribes territorial integrity in Crimea and possibly in Artsakh, I have reinforced my original point that we should change things to better accommodate self-determination.

Artsakh, Crimea and Taiwan are not isolated examples. They belong to a class of cases where the population has strong sentiment against internationally recognized borders and has already effectively achieved their preferred arrangement on the ground. Other cases like this include Kosovo, Western Sahara, Rojava, and the Turkish Republic of Northern Cyprus. These are straightforward cases where self-determination must be recognized.

Abkhazia, South Ossetia, Donetsk, and Luhansk are cases where breakaway republics have effectively seceded but the appropriate judgment is more nuanced. In Abkhazia and South Ossetia, the current populations are heavily in favor of secession but this is at least partly due to the recent history of ethnic cleansing of Georgians who originally constituted one-half and one-fourth of the respective populations. In Donetsk, polling suggests that the majority of locals would like to return to Ukraine, and in Luhansk, there doesn't seem to be any reliable information on popular opinion, and of course neither of their referenda are trustworthy. Neither Ukraine nor Georgia have acted particularly atrociously towards people in these regions, and all four of these breakaway republics seem less free and democratic than their parent states. The secessions of Donetsk and Luhansk were blatantly illegal just like that of Crimea, whereas the secessions of Abkhazia and South Ossetia might be justified under the 1990 Soviet secession law referenced for Artsakh but are probably illegal.

Overall, the breakaway Georgian and Ukrainian republics are nuanced cases where we should use fair referenda and diplomacy to determine the appropriate status. The Minsk II agreement or some similar successor must be implemented in Ukraine. The right to return for refugees is of course very important in Abkhazia and South Ossetia. There are also a variety of local injustices that need to be halted such as kidnappings and borderization. Still, it's hard to see how to make progress on any of this in the face of both the Russian military and local separatists.

Then there are cases where the population may wish to secede against the de facto status quo. These include Tibet, Xinjiang/East Turkestan, West Papua, Scotland and Catalonia. In these cases, the important first step is simply that the people's right to a fair referendum be recognized, even though it is often dubious whether voting for secession would actually be wise and it is often unclear how to actually convince countries to allow such referenda. If it is clear that majority opinion favors secession or at least a move towards greater autonomy, further steps must balance the need for political change with pragmatic questions of how to achieve it without causing too much damage.

Finally, there is the dilemma of Israel and Palestine. This has very high relevance to current politics, but it is more complicated with the demands for self-determination by the intertwined Israeli and Palestinian communities, including the Zionist conception of self-determination as implying a state which is inherently or mostly ethnically Jewish. Moreover, many Palestinians wish not for independence from Israel, but for greater integration and rights within Israel. Certainly the status quo does not adequately protect Palestinians' right to self-determination, but whether the appropriate solution involves two separate states, a binational confederative state, or a unified democratic state requires a much more extensive analysis beyond the basic principles of self-determination.

My approach here contrasts with the dominant paradigm by which the international community and liberal think tanks aim to resolve sectarian conflicts. That is, I do not believe that bilateral or broader negotiations are very effective, and I do not believe that the best way to approach conflict resolution is to cajole both sides into admitting their faults and learning to recognize the validity of each other's grievances. These things are certainly nice, but I believe they're not very effective. We should focus on the underlying political problems - the lack of rights, lack of security, lack of agreement on political status, the presence of malign regimes - and push for robust institutional fixes to these problems, even though at least one side will invariably be dissatisfied in the short run. After we do that, people of different cultures and nationalities will learn to get along with each other without our patronization.

If my view is so sensible, why do relatively few people follow it? I would speculate: first, aside from elites, hardly anyone thinks at all about international law. As for elite opinion, they disproportionately represent the interests of existing power structures (i.e. recognized states), they are affected by status quo bias, and they are reticent to take sides in complicated ethnic conflicts because (a) doing so is seen as unfashionable and low status in elite circles and (b) doing so inevitably provokes anger and offense by the other side.

How to change the world

I've outlined a philosophical framework for how we as global citizens should ideally think about self-determination. However, it's clearly not suitable to work as a legal framework for the existing actors of the international community.

Lawyers, politicians and legal scholars could try to restructure international law in a way that appropriately recognizes self-determination, similar to how reformers in the early 1900s worked against interstate warfare. But at a more basic level, we don't need to present a full legal architecture to replace the current one. Self-determination is already a principle of international law. All we have to do is reinforce it at the cost of weakening the principle of territorial integrity. This is a topic of active legal and philosophical debate, and existing international law leaves a good deal of gray area for states to act marginally differently. Merely raising the importance of self-determination is not as philosophically sound as the full framework that I've described here, but it's good enough to get the job done and simple enough to easily communicate, and the international community can figure out the details.

One way to strengthen the idea of self-determination is to lobby for government actions which set precedents in its favor, as happened with Kosovo and Crimea. Neither of these have been very effective precedents, first because states (dubiously) insisted that their recognition of Kosovo was not a precedent for anywhere else, and secondly because the annexation of Crimea was flagrantly illegal. Recognizing self-determination in cases with better legal footing, and without denying that we are setting a precedent, could go much further in changing norms.

On a rhetorical level, it may be helpful in some contexts to emphasize that prioritizing territorial integrity is inherently a conservative idea (valuing stability and legal formality, or "law-and-order" over the rights and aspirations of peoples) whereas prioritizing self-determination is inherently a liberal idea. In communities where leftism or liberalism predominate, like Western academia and mainstream media, emphasizing the right-wing nature of territorial integrity may invite greater scrutiny and skepticism of the concept. Additionally, we can illustrate the illegitimacy of how states and the international community forbid self-determination by likening it to monarchs forbidding elections or white male governments forbidding women and other races from voting.

  1. Note that I don't bother with detailed histories of political grievances or cultural differences. I disagree with the common assumption that deciding a solution to a sectarian conflict requires deep knowledge of the histories and cultures of the involved parties. Although I am personally familiar with many more details of these conflicts, I don't include it because I think it is unnecessary. Just like we don't need to know very much about a culture in order to talk about whether they ought to change their approach to democratic elections or human rights, the same goes for talking about their approach to self-determination. This disregard for details may seem crazy, but if you disagree I simply challenge you to try to name some historical or cultural facts which should change my judgments. ↩︎

  2. About 40,000 Azerbaijanis were expelled from Nagorno-Karabakh, I am guessing 30,000 for the diminished current territory of Artsakh. ↩︎

  3. There are many allegations that this referendum was rigged. The prior likelihood of election-rigging in favor of Russia is high considering that the referendum occurred in the presence of soldiers from Russia (an autocratic state with no pattern of respect for democratic principles) and separatist militias who forcibly turned away OSCE election monitors and had very recently exerted extralegal armed control of the parliament building as the parliament voted to schedule the referendum. Russian military interference seems particularly likely because at this point in time the Russian government was denying that Russian soldiers were even present in the country (if you can deny that you were present, you don't have to worry much about consequences for misbehavior).

    No OSCE or UN observers were present to validate or invalidate claims of electoral fraud. A variety of alternative observers were able to monitor the election, and as far as I know all affirmed that the election was fair, but at least some of these observers can be disregarded due to being overtly politically biased, e.g. ranting against US hegemony or being members of far-right European political parties. I haven't seen evidence that all (or even most) monitors were like this, but considering the roadblocks stopping OSCE monitors, we can at least presume every election monitor who did attend was only there at the allowance of Russia and separatist militias, and is therefore not particularly credible.

    So monitors don't provide good evidence for or against the integrity of the referendum. There have been specific allegations of fraud, but these can easily be exaggerated or false (Americans will recall the 2020 presidential election drama for an analogy), and I haven't seen serious verification.

    Perhaps the best way we can judge this without having to delve deep into primary sources is to ask the Crimean people. In 2014 surveys, 91% of Crimeans said the referendum was free and fair and 83% said the results likely reflected the views of most people there. This is good enough to reject the possibility that fraud changed the outcome of the referendum. If 51% of Crimeans voted against unification with Russia, you wouldn't see 83% of them saying that the results reflected the views of most people - it's just not plausible for people to be so badly mistaken about the views of their neighbors. And if there was widespread fraud in a direction that disenfranchised the majority, enough reports about it would spread that fewer than 91% of people would say the election was free and fair.

    However, it could still be the case that fraud increased the vote margin. While the referendum result was consistent with some of the polls (given that opponents of unification tended to boycott the referendum), there was certainly the potential for widespread electoral fraud even if there is not strong evidence for it. So overall, it seems likely that electoral fraud increased the victory margin while not changing the outcome. ↩︎

  4. Another war could be similar to the last one, which caused the deaths of 7,000 soldiers and 200 civilians and involved numerous war crimes. But war this time could be much worse because Azerbaijan could easily cut the Lachin Corridor, thus trapping the entire population of Artsakh (120,000 people), which would be the obvious thing to do from a military perspective, and was a major operational aim of the Azerbaijani military in 2020 (they succeeded just prior to the end of hostilities). Cutting the Lachin Corridor at the outbreak of war would threaten death and atrocity on a much larger scale. Armenian civilian deaths in 2020 numbered only about 100 partly because most civilians in Artsakh were able to flee to Armenia.

    Azerbaijan may want to avoid the political consequences of creating such a humanitarian catastrophe; this may have been part of their motivation for agreeing to a peace settlement in 2020. Therefore, it is possible that most Armenian civilians would be able to flee from Artsakh. If Azerbaijan then captures Artsakh, the refugees will probably be too afraid of Azerbaijani retribution and too opposed to life under such a regime to try returning to their homes after the war. I have not seen evidence of any Armenians trying to return to homes in territories lost in the 2020 war, despite the Azerbaijani government claiming to allow it. ↩︎

  5. During the Soviet era, Nagorno-Karabakh may have suffered discrimination within the Azerbaijan SSR. Zverev (1994) writes, "For 65 years of the NKAO's existence, the Karabakh Armenians felt they were the object of various restrictions on the part of Azerbaijan. The essence of Armenian discontent lay in the fact that the Azerbaijani authorities deliberately severed the ties between the oblast and Armenia and pursued a policy of cultural de-Armenization in the region, of planned Azerbaijani settlement, squeezing the Armenian population out of the NKAO and neglecting its economic needs." Krüger (2010) writes, citing a German book, that economic shortcomings "did not necessarily appear to be a consequence of a discriminatory economic policy from Baku, but rather of the overall economic situation in the USSR".

    During the 1988-1994 war, Azerbaijanis committed mass murders of Armenians in Sumgait, Kirovabad, Baku and Maraga; the Azerbaijani government and public opinion have variously ignored these events or spread false conspiracy theories accusing Armenians of staging them as false flag operations. Azerbaijanis even regarded some of the perpetrators of the Sumgait pogrom as heroes. Then in 2004, an Azerbaijani military officer murdered an Armenian officer in Hungary, for no apparent reason other than racism; upon being extradited to Azerbaijan, he received a pardon and a hero’s welcome by Azerbaijani officials.

    Azerbaijani war crimes against Armenians in 2020 included the use of cluster bombs and other munitions indiscriminately against civilians, the execution of civilians and prisoners of war, torture of civilians and prisoners of war, and desecration of Armenian corpses. There is scant evidence of any accountability for these war crimes, and in one analyst’s view, “it's impossible to not assume that the Azerbaijani state has given its tacit approval to these war crimes.” Azerbaijani comments on social media generally denied these war crimes or excused them with whataboutism. Azerbaijan has also kept many Armenian prisoners of war for months after the conclusion of hostilities.

    Hate speech against Armenians is rampant in Azerbaijan and is primarily promulgated by Azerbaijani politicians and civil servants. In 2020, an Azerbaijani football official explicitly, publicly called for the genocide of Armenians; he was apparently punished by the government, but “probably not” fired. A victory theme park opened in Baku in 2021 displayed a wall of helmets of dead Armenian soldiers and featured mannequins formed as racist caricatures of Armenian soldiers.

    Given this environment, discriminatory actions against Armenians living in Azerbaijan are almost inevitable, although we don't know much because the number of Armenians living in Azerbaijan proper (outside Artsakh) is very small and because they "need to hide their ethnic affiliation and there is no organisation of the Armenian minority in the country".

    Azerbaijan has also strongly attacked Armenian culture. Azerbaijan has destroyed numerous ancient Armenian monuments, including a medieval cemetery in Julfa. A 130-year-old cathedral in Shushi was shelled twice by Azerbaijan during the recent war (Azerbaijan denied responsibility for this, suggesting that the Armenians staged another false flag), and later vandalized with graffiti. Another church in the city was also damaged at some point. Going along with this, the Azerbaijani government and academia are propagating a pseudohistorical narrative which denies the history of Armenian heritage in the region. The historical buildings and monuments which remain in Artsakh would have a dubious future under Azerbaijani rule. Meanwhile, as Azerbaijani soldiers have retaken parts of Artsakh, they have been filmed desecrating recent Armenian graves. Azerbaijan also does not recognize the Armenian genocide. ↩︎

  6. See events from the First and Second Karabkh wars. ↩︎

  7. He argues that while Article 72 of the Soviet constitution did not invalidate the 1990 secession law, it nonetheless provided an alternative procedure for republics to freely secede without going through the cumbersome process defined in that law, and that Azerbaijan availed itself of the constitutional right. So far so good, but he is wrong to infer that Nagorno-Karabakh therefore did not have the right to self-determination that was specified by Article 3 of the 1990 law. That is, the Republic of Azerbaijan may have had the constitutional right to freely secede from the USSR according to Article 72 of the Soviet Constitution, but that does not mean it did not violate the 1990 law when it chose to keep Nagorno-Karabakh with it.

    The reason the 1990 law is considered unconstitutional is that it introduced a very complicated procedure contrary to Article 72's statement of a free right to secede. In fact the 1990 law was purposefully designed that way to inhibit the dissolution of the USSR. The envisioned process was so cumbersome that it was never actually followed by any Soviet republics, and the central government ended up recognizing secessions anyway in spite of their disregard for the 1990 law. Yet this does not mean that republics had no obligation to abide by the law at all. The 1990 law provided for more than just an optional process by which republics could secede; it also laid down certain rights and obligations which were not obviously limited to cases where republics followed the secession process laid out in that law. Consider that Articles 15 and 16 established certain rights and privileges that seceding republics had to guarantee for their citizens; these were presented not as procedural requirements for secession, but simply as legal obligations for the seceding republic. Even though seceding republics could cite Article 72 of the Constitution to bypass most of the 1990 law, I believe they were nonetheless legally obligated to follow Articles 15 and 16.

    Crucially, a similar argument applies to Article 3, paragraph 1, sentence 2. While sentence 1 describes the referendum for an autonomous region as being just a part of the overall procedure for secession by a republic, sentence 2 provides a standalone right of self-determination which does not assume that the republic is seceding through the cumbersome process outlined elsewhere in the law.

    Krüger argues that even if Nagorno-Karabakh had this right, it never followed the procedures that the 1990 law laid out for secession. This is true, but it does not follow that Nagorno-Karabakh had to remain part of Azerbaijan, for that would abrogate its right to self-determination as defined by Article 3 para. 1 sentence 2. Nagorno-Karabakh could theoretically still have had this right while being part of Azerbaijan because it could appeal to the Azerbaijani government for secession, but this is obviously not tenable considering Azerbaijan's attempt to annul Nagorno-Karabakh's autonomous status as well as its more general racist and authoritarian characteristics. Basically, Nagorno-Karabakh had a remedial right to secede from Azerbaijan because that was the only way to protect its right to self-determination as guaranteed by Article 3 para. 1 sentence 2.

    Krüger claims that Article 3 para. 1 sentence 2 "clearly presented a hurdle and a serious consequence" for states looking to secede, thus making it "an extraordinary condition and substantial restriction" of Article 72. However, his arguments here fail. Why is it so bad for a republic to suffer the loss of an autonomous region that we should consider it a substantial barrier to secession? He writes it "would thus have led to the splintering and downsizing of assured union territories", yet the autonomous regions were generally very small relative to their parent republics, were not particularly numerous, and some of them preferred to stay united anyway. For Azerbaijan, losing Nagorno-Karabakh would have meant losing just 5% of its land and just 3% of its population. He says that secession would have caused "serious economic and geo-strategic damage", but in what respect? Of course a slightly smaller republic will have a slightly smaller economy, but this does not mean that wealth and standards of living per capita would fall. The inhabitants of the rest of Azerbaijan would be just as well-off as before. And losing territory could indeed cause some geo-strategic damage, as smaller countries are generally weaker, but considering the relatively small size of most autonomous regions (including Nagorno-Karabakh) it would not be appropriate to call this serious damage. And in practice, it is the attempts to quash secessionist movements which have done more to damage the geopolitical positions of states like Azerbaijan and Georgia.

    Krüger goes on to say that self-determination for autonomous regions "would have posed the risk of ethnic, political, and military conflicts, which was indeed tragically confirmed". This is an incredibly stupid comment. Such tragic confirmation was impossible in our timeline of history because neither Nagorno-Karabakh nor any other autonomous region was actually granted recognition of a right to self-determination; in reality, parent republics such as Azerbaijan waged wars to reinforce their territorial integrity with tacit permission from the international community. The fact that many peoples' unrecognized efforts to secure self-determination led to tragedy (and only in concert with parent states' attempts to quash such aspirations) does not show that actually recognizing self-determination would have led to tragedy. On the contrary, it is probable that a more liberal and sane approach to self-determination in these territories would have led to less war and tragedy, at the "cost" of diminished territorial grandeur for the union republics which were characterized by racist and authoritarian policies towards their subjects.

    Of course, states have clearly acted as if the loss of territories such as Nagorno-Karabakh is inherently extremely costly. Yet just because they have acted this way does not mean it is rational or reasonable to do so. It is driven by ideology divorced from common sense and honest concern for the welfare of one's people, similar to the Scramble for Africa and other nationalistic fads. In hindsight at least, it is beyond apparent that Azerbaijanis would be better off had the country simply granted Nagorno-Karabakh's independence in 1991, and similar things can be said for other countries wracked by these kinds of conflicts.

    So Article 3 para. 1 sentence 2 of the 1990 law did not contradict Article 72 of the constitution. Now, this does not mean that it was constitutional. Krüger states that it also violated Articles 78 and 86. But if a Union Republic secedes then it is no longer a Union Republic and therefore not subject to those articles. Additionally, if a republic secedes then it implicitly consents to self-determination for its autonomous regions according to the 1990 law, thus undermining the claim about Article 78.

    To summarize, my theory is that seceding states had to grant certain reasonable rights and privileges to their constituents that were outlined in the 1990 law even though they didn't necessarily have to follow the cumbersome procedure that the 1990 law outlined for how secession was supposed to work, that self-determination for autonomous regions was one of those rights, that unilateral secession was an appropriate remedy given the unfeasibility of the secession process outlined in the 1990 law, that this did not violate Article 72 of the Soviet Constitution because the mere act of secession by autonomous regions did not unduly harm Soviet Republics, and that this did not violate Articles 78 and 86 of the Soviet Constitution because those tenets were affected by Azerbaijan's decision to secede. I am not highly confident in this theory so I regard the overall question of the legality of Nagorno-Karabakh secession under Soviet law as ambiguous. ↩︎

  8. His interpretation of international law depends too heavily on the practices of states as opposed to the texts of international legal documents. His rejection of the idea that Armenians of Nagorno-Karabakh face severe human rights violations under Azerbaijani rule makes empirical and conceptual errors which are bad enough as to be racially offensive, with the main problems being his failure to mention multiple larger anti-Armenian pogroms besides the one in Sumgait, blaming the Armenian people for provoking violence through their secession movement (which seems factually accurate enough but is nonetheless invalid victim blaming when it is used to reject a remedy for human rights violations), and relying too heavily on the history of Armenians under Soviet Azerbaijani rule as opposed to recognizing the racist character of the current regime. He also assumes that "the employment of the human rights system and the humanitarian intervention with a mandate of the UN Security Council" would probably be sufficient to guarantee human rights in an Azerbaijan-controlled Nagorno-Karabakh, which is dismissible nonsense.

    Note that there is a separate question of whether Nagorno-Karabakh might have the right to secede under international law but without recourse to self-determination, but Krüger effectively refutes this in chapter A, section IV part 1. ↩︎

  9. Here it is worth addressing the trope of Chesterton's Fence which is popular in certain amateur intellectual circles and is particularly poignant in the case of national borders. The idea is that we should not reform institutions without first understanding why they were made that way in the first place. But in the case of national borders, this is easy to dismiss - the history of wars and treaties leading to current borders can be read quickly from Wikipedia or a history textbook, and generally reflects historical power balances and battlefield successes which lack normative meaning for the modern day. ↩︎

No comments on this post yet.
Be the first to respond.
More from kbog
Curated and popular this week
Relevant opportunities