1 The Russian-speaking population as an emergent identity

In Identity in Formation (chap. 10), through the use of content analysis from the press in the four republics, I inferred the emergence of a conglomerate identity, the “Russian-speaking population.” The content analysis revealed a plethora of designations for the non-titular populations. The titular press typically referred to them as negations: the unrooted (nekorennye); people without a country (apatridy); foreigners (inorodtsy); foreign speakers (inoiazychnye); denationalized (denatsionalizirovannye); non-citizens (negrazhdane); not-from-Russia Russians (nerossiiskie russkie); illegals (nelegal’nye emigranty); and several other epithets as well, reaching down to damned Russians (prokliatye russkie), and Russian shit (russkoe govno). Also in the titular arsenal were colonist (kolonisty), occupiers (okkupanty), and migrants (migranty).

The non-titulars were, in the face of the double cataclysm, equally experimental in naming themselves. They were co-fatherlanders (sootechestvenniki) with those of the Russian Federation; Slavs; members of the Russian state (rossiiskie or rossiiane); ethnic Russians (russkie); Soviets; minorities; and Cossacks (which President Yeltsin expanded as a new cultural-ethnic social formation [kul’turno-etnicheskaia obshchnost’]). All of these designations hardly captured the deep common social and cultural challenge faced by these newly realized minorities. And the designation “Russian-speaking population,” a term basically invented in 1989, was widely adopted by groups of Russians, Ukrainians, Poles, Belarusians, Jews, and titulars who had become Russified in the Soviet era.

What to make of this content analysis? I was careful to backtrack on a claim I made in an earlier publication that this category of Russian-speaking was a nationality designation. My data only allowed me to portray it as a useful label to categorize a population (the difference between natsional’nost’ and naselenie). But I portrayed it as a powerful symbolic tool for ethnic entrepreneurs to unify the non-titulars (and Russified titulars) into a political project to assert a common response to the nationalizing projects of those who inherited the post-Soviet Union Republics.

My claims about this identity shift met with substantial disbelief among area researchers. Poppe and Hagendoorn (Reference Poppe and Hagendoorn2001, 58) argued that my categorization of a Russian-speaking population is not “accurate”: “Whereas the use of this label may serve as a simple means to generalize ‘the other’ from the titular perspective, the question remains whether different non-titular ethnic groups speaking the Russian language (e.g. Jews, Gypsies, Belarusians and Russians in Ukraine) consider themselves as one group. Moreover, the label ‘Russian-speaking population’ is ambiguous in those republics where a considerable number of titulars speak Russian as their first language.” This underestimates the “variety within Russian communities.” Equally critical, Fournier (Reference Fournier2002, 415, 430) argued that while there might be a fusion of the Russian-speaking Ukrainians and Ukrainian Russians, there is also a disjunction, one that my categorization missed. Russian and Ukrainian Russophones, she pointed out, typically do not share a common culture: “their identities [have] different configurations.”

Moreover, the evidence is mixed as to the staying power of this conglomerate identity in the past twenty years. Identity categories remain fluid. In 2017, Bhavna Dave reported (in a personal communication) that Kazakhstan and Kazakhstanis are still tinkering with the ethnic-civic debate: “They need to keep pedaling, reinventing the ‘national’ vs ‘soviet’ (kazakh vs kazakhstani) debate, inventing new terms, pseudo-ideologies to continue to mobilize support. They are officially talking about multilingualism – ‘trinity of languages’ with English added. It will turn into a quartet in a few years with Chinese added! ‘Eurasianism’, and not ‘Russian-speaking identity/nationality’ is now the dominant term.” Similarly, in a private communication, Volodymyr Kulyk insisted that “the error in Identity in Formation was in relying on elite media rather than self-identifications of ordinary people…if Laitin did that, he would not have found [at least in Ukraine] such an identity. The Ukrainians (even Russian-speaking) identified as national Ukrainians; the Russians with their localities (Crimea, Donetsk).”

However, as I pointed out in the book, the category of “Russian-speaking” has become such a cliché that it remains almost unnoticed in its ubiquity. Social commentators a quarter-century after the second cataclysm still have no other way to categorize this admittedly diverse population. Katja Koort, in World Affairs in Reference Koort2014, writes: “more than three hundred thousand Russian-speaking people (including Ukrainians, Belarusians, and others) are living in Estonia.” (Koort Reference Koort2014, 68). In a study of internet posts on the Maidan in three languages from Ukraine, Etling (Reference Etling2014) writes that his “early results point towards more support among Russian-speaking Ukrainians for the protests than the Russian government and mainstream media claim.” The all-Ukrainian non-governmental organization “Human Rights Public Movement ‘Russian-speaking Ukraine’” (Russian: Всеукраинская общественная организация «Правозащитное общественное движение «‘Русскоязычная Украина’”) was founded on March 1, 2008 in Severodonetsk, in the Luhansk region, during the second all-ukrainian congress of deputies of all-level councils, and registered on 10 August 2009 by the Ministry of Justice (Kyiv).Footnote ²

I invite regional experts to replicate the content analysis performed in Identity in Formation, but now using social media and modern forms of text scraping, to provide a quarter-century report on the cultural and political identities of the Russified populations in the former Soviet republics. It may be that my forecasts were incorrect and that the strategies of the titular elites to segment the minority populations have been successful, so that now local identities (Kulyk) or different legal statuses (Pettai and Hallik Reference Pettai and Hallik2002) predominate. But the attractiveness of the label suggests strongly that “Russian-speaking” has a latent power to mobilize political support (albeit with different coalitions in different republics) as a conglomerate identity.

2 Russian nationalism in Russia is likely to diverge from the nationalism of Russian-speakers in the Near Abroad

An implication of my claim of an emergent Russian-speaking identity (as I draw in chapter 11) is that we should observe different themes in the articulation of nationalism in Russia as compared to those in the titular republics. To be sure, there is no historically coherent expression of Russian nationalism as a political project.Footnote ³ But in the 19^th century, two forms of nationalism emerged under conditions of economic backwardness combined with an Orthodox authoritarian tradition. The first (and emphasized by Greenfeld (Reference Greenfeld1990) in her notion of ressentiment) was to emphasize that the Russian “soul” was deep and superior to the shallow values of (the more developed) West. Here religion becomes the glue that holds this nation together, as in Tsar Nicholas I’s “Official Nationality” proclamation declaring Orthodoxy, Autocracy, and Nationality as the three principles of his governing doctrine. As Dostoyevsky wrote, “he who is not Orthodox cannot be Russian” (quoted in Dinello, Reference Dinello and Shlapentokh1994, 198). The second strand emerged as a response to the incomplete rationalization of the Russian language to the imperial periphery, and the distinction between those who shared the core nationality (russkii) and those who were subjects of the empire (rossiiskie).Footnote ⁴ Russians who saw themselves as rossiiskie were internationalists, and this became the dominant theme of the Bolsheviks (at least until World War II) who wavered between discrediting nationality and heralding a “Soviet” national project (in which Russians were the elder brothers and Russian the language of internationality communication).

Examining these two strands, I speculated that in the wake of the Soviet collapse, there would be a resurgence of the Orthodox national tradition in the Russian Federation. Indeed, as I was researching my book, Orthodox imagery was becoming quite fashionable in Russian society. However, I inferred that in the near abroad, because Russians (at least in the Baltics and Kazakhstan) had been thrown into the same boat with non-Orthodox Russian-speakers, it would be inconvenient for political entrepreneurs representing them to rely on religious symbols to build political coalitions. Here the internationalist strain of Russian nationalism would become dominant. In practical terms, as I observed the scene post-collapse, Russian nationalism in Russia was consumed by the debate over Russia’s future role as a great power; meanwhile Russian nationalism in the near abroad was consumed by the debate as to whether the much-heralded “friendship of nations” (druzhba narodov) could be sustained.

While the discourse analysis presented in chapter 11 differentiated the two nationalist strands with their different emphases in the near abroad as compared to Russia, an examination of the survey data I collected with Jerry Hough provided no support for my inferences. There was no evidence of greater attachment to Orthodoxy in Russia in comparison with Russians living in the four republics of my study. Moreover, there was only weak support for Russians in Russia holding to a “Russian” identity with Russians in the near abroad holding a “Soviet” identity. I could not determine whether my inferences were incorrect, or whether it would take more time for these differences to appear.

Evidence from the World Values Study (World Values Study Wave 6, 2010-2014), however, provides new support for my inference. Comparing self-declared Russians in Russia and in Estonia on questions on the importance of religion, Russians living in the Russian Federation were 8.4% more likely to report that it is “very” or “rather” important, and 11.9% less likely to say “not very” or “not at all” important. On a measure of exclusiveness, the WVS asks if those who follow my traditions should be “very much” or “somewhat” like me. Russian respondents in the Russian Federation were 18.3% more likely to agree; and 33.4% less likely to want to share tradition with those who were “only a little” or “not” like me. And on the question of whether respondents trust folk from a different nationality, Russians from the Russian Federation were 11% less likely to agree, and 9.8% more likely to disagree. I interpret these results (albeit from only one titular republic) to reveal a more internationalist and secular version of these social identities in Estonia, and a more Orthodox and exclusive social identity in Russia. Further research is obviously needed to corroborate this trend.

3 Russian-speakers in the Baltics would face incentives to assimilate into the titular culture

The imperial and Soviet states permitted, even encouraged, professional mobility of Baltic elites within their republics; and given the economic dynamism of the Baltic states, with highly educated populations, those titular elites that accommodated to Russian hegemony could garner jobs in the arts, the sciences, and in management in their own republics, with little need (and no encouragement) to explore career success in St. Petersburg (Leningrad) or Moscow. I called this the integralist pattern in Russian/Soviet expansion. Those Balts who were active Bolsheviks, once they took responsible positions in the homeland of their ancestors, acquired facility in Russian, but operated politically and professionally in the titular languages. Thus at the time of republican restoration and the passage of the new language laws, virtually all Balts accepted (and rejoiced in) the nationalizing program. And given the claim of illegal Soviet incorporation, the Russian-speaking population that moved to the Baltics after World War II (almost none of whom spoke Estonian or Latvian) were denied citizenship. In what I pictured as Kafkaesque treatment (as in Joseph K. in The Trial), the Russian-speaking populations were left in legal limbo, insecure with their pensions and passports with their dignity challenged.

What could they do? Exit to Russia was not attractive. I overheard many Russian-speakers in the telephone center in Narva, where I did the fieldwork for this study, begging (distant) relatives to take them in, but to little avail. Across the Narva River, Estonian Russians could see the dismal poverty in the neighboring town of Ivangorod, hardly a place to resettle. And Baltic Russians had implicit property rights to their Soviet-era apartments. A federal solution with Russian and Baltic zones was demographically impossible given the overlapping populations in Latvia; but possible with an autonomous region in northeast Estonia. Yet especially in Estonia, the titular elites were resolute in dropping Russian from their (children’s) language repertoires, and made sure that for jobs and university training, the titular language was a sine qua non. Federalism was not an option for the Russian-speakers.

An attractive alternative for the Russian-speakers was assimilation. For one, in an experimental test, they demonstrated high respect for co-ethnics who had facility in the titular language (Laitin Reference Laitin1998, 235).Footnote ⁵ Also, there were excellent job opportunities for Russians in the expanding Baltic economies. These factors help explain why Russian-speaking respondents from the two Baltic states in my high-n surveys had the highest scores on the index for openness to assimilation (Laitin Reference Laitin1998, 252). With these opportunities, the third option (exit through emigration) was less attractive.

Regional experts were aghast (and this before my fieldwork) when I proposed from a game theoretic model as to why we should expect Russians to become Estonians and Latvians more or less as Andalusians became Catalans in Catalonia and Russians became New Yorkers on the Upper East Side of Manhattan. Indeed, the game model predicted a slow trend for perhaps a generation, and then an assimilationist cascade.Footnote ⁶ However, and even after my fieldwork, I had only attitudes towards assimilation as evidence, but hardly any examples of actual assimilation in real life.

After twenty years, what does the evidence reveal? Preliminary evidence analyzed by Koort (Reference Koort2014) shows that twenty-one percent of non-ethnic Estonians have been successfully integrated. This group comprises the younger people who were born and educated in Estonia. They have a good command of Estonian and they can cope economically. Most of them are Estonian citizens and consider themselves part of the Estonian community. They trust the Estonian state and authorities, live and work in an Estonian-speaking environment, and follow Estonian media and prefer it to Russian channels. They live in Tallinn and other larger cities, and are underrepresented in Ida-Virumaa County in the northeast, where Russian-speakers predominate.

Data from the Estonian Ministry of Education show similarly a rather smooth transition to Estonian for non-titular youth and no resistance from their parents. At the moment of restored independence, virtually no Russian-speakers were being educated in Estonian as a medium of instruction. The recent data show that in grades 1-3, 28.9% of Russian mother-tongue students in Estonia (and 23.9% in Ida-Virumaa, where about 95% of the population is Russian mother-tongue) are learning in Estonian. At the higher levels (10-12) virtually all Russian speakers, whether from Ida-Virumaa or elsewhere, are exposed to Estonian medium of instruction. This is certainly a strong indicator of linguistic assimilation, as the model predicted.

Note for publication in 2025: the url with these data is unavailable. Very similar data for the year 2016 are archived at https://andmed.stat.ee/en/stat/Lepetatud_tabelid__Eri-valdkonnad.%20Arhiiv__loimumine/HT140/table/tableViewLayout2

In interpreting these data, Aksel Kirch, Marika Kirch, and Tarmo Tuisk (personal communication) write that since 2011, all state and municipal gymnasiums which were teaching solely in the Russian medium now must teach 60% of subjects in Estonian and 40% in Russian. They infer that Statistics Estonia (supplier of these data) considers these Estonian medium schools. In addition, there is a strong step-by-step immersion in many Russian language schools during grades 4-9, and the idea of it is to prepare Russian-speakers at minimum for these 60/40 gymnasiums. If their Estonian language skills are not sufficient for gymnasium, the students are typically tracked into vocational schools or directly into the work force, as there are no state schools at that level operating solely through the Russian medium. Importantly, this educational rationalization in Estonian has not met with resistance. The Integration Monitoring Study of 2011 showed that 32% of Russian mother-tongue respondents prefer their children to have their education in the Estonian language and another 25% prefer schools with (Estonian) language immersion programs. That leaves some 37% of Russian-speaking parents who prefer their children learning in Russian schools. Intergenerationally, this is approaching a language cascade.

Data from Latvia tell a similar story. Bonnard (Reference Bonnard2013) describes a dual dynamic: non-Russian minorities adopting Russian as a mother tongue (this the emergence of a trans-ethnic Russian-speaking minority); but a slow progression of all minorities moving toward Latvian as their mother tongue. Bonnard (Reference Bonnard2013, 286) calls this a “double melting pot”. Indeed, in the 2000 census, there were 1.8% more Latvian nationals by declaration than in the registry data, and -0.6% fewer Russians. Especially among the young, Bonnard writes (2013, 280-81), “The slight hold of origin on the feeling of ethnicity facilitates trajectories of assimilation.” In the choice of ethnicity of children of mixed Russian-Latvian marriages from 1989 to 1998, there was a 27.1% increase in declaring the child as Latvian (from 61.3% to 88.4%). In 1960, 57% of the children in Russo-Latvian marriages classified themselves as Latvian on their passports. In 1989, this reached 69.8%. Even for Russian and non-Russian/non-Latvian marriages, where children were typically registered as Russian, there is now a trend for classifying the kids as Latvian (2013, 284). Bonnard (despite his vigorous objections to my forecasts) identifies both the emergence of a latent Russian-speaking identity and the creeping success of a Latvian assimilationist project.

In the final analysis, I did not discount the possibility of the formation of a Russian-speaking nationality as a possible counter-hegemony to the Baltic nationalizing projects. But I thought the powerful incentives for the fulfillment, through the cumulative assimilation of Russian-speakers into Balts, of both an Estonian and a Latvian nation-state in the model of 3^rd Republic France (Weber Reference Weber1976), would predominate.

Missing, however, from the tipping model and my political analysis was the implication of future EU membership for the Baltic States. This gave young Russian-speaking Balts incentive to learn enough of the titular language to get citizenship and then emigrate to the EU job market (learning English or German, with their rudimentary Estonian or Latvian falling into desuetude). To some degree, and this needs to be examined empirically, this would amplify the assimilationist logic. Those who remain in the Baltics were likely to be those most comfortable in Estonian and Latvian. The resulting selection dynamic would exaggerate the success of my assimilationist prediction, but with a different mechanism.

4 Russian-speakers in Kazakhstan would exit and establish themselves as citizens of the Russian Federation

Kazakhstan, as with the other Central Asian Republics, was incorporated as a colonial territory, more or less like Algeria into France. In the wake of the double cataclysm, non-Kazakh Russian-speakers, as confirmed by the contempt for Russians performing as Kazakh speakers in the matched guise experiment (comparing respect scores for the same voice of a Russian speaking in Russian compared to speaking in Kazakh in Laitin Reference Laitin1998, 235), could not even dream of assimilating into a Kazakh nationality. Meanwhile, since nearly all the elite Kazakhstanis were Russian-speakers (many of whom had intergenerationally lost facility in Kazakh), Kazakhstani elites were divided among themselves whether to promote a culturally nationalizing program. Thus, as a prudent nationalist, President Nazarbaev pressed for a civic agenda, where both languages would have their official roles. He maintained publication of the two state newspapers, one in Russian and the other in Kazakh, whose origins went back to the early 20^th century (Abishev et al. Reference Abishev, Kurmanov and Sabitov2024). The crucial question for the Russian-speaking technical experts (of both nationalities) in the oil and missile industries was whether Nazarbaev (and his successors) could stave off a second-generation of Kazakh leaders with nationalizing goals. Already in the early 1990s, enormous investments in new mosques gave the impression of a cultural revival. Epithets for Russian-speaking Kazakhs – such as being called mankurty (zombies) – were also evident. And Russians watched carefully, fearing the fate of the pieds noirs in Algeria, many with their suitcases packed, ready to find a new home in their diasporic homeland. I therefore predicted a reasonably stable Russian and Russian-speaking Kazakh cosmopolitan elite living tentatively in Kazakhstan, but ever ready to emigrate.

My prediction that the Russians would begin to “pack their suitcases” when they foresaw a nationalizing project is only partially right. From a height of 42.7% of Russians in Kazakhstan in the 1959 census, it dropped to 29.9% in 1999, reflecting the results of a near decade of republican independence. In the latest ethnic demographic data of 2014 (as was reported under “Ethnic Demography of Kazakhstan” in Wikipedia on April 8, 2025), this figure of Russians (as an ethnic group) in Kazakhstan is down to 21.5%. To be sure, there is no evidence of political marginalization of Russified Kazakh intellectuals. Nor has emigration of Russians been of the cascade quality. But as I wrote, this may well be because of the technical expertise required for the oil boom, and the prosperity that has followed. Indeed, in 1990 Kazakhstan’s GDP was $26.93 billion, behind Ukraine and about equal with Uzbekistan. But in 2013 it reached US$236.6 billion, well richer than Ukraine and Uzbekistan. Although there has been a retraction with a drop to US$133.7 billion in 2015, the World Bank classifies it as an upper-middle-income country with GDP per capita of nearly US$10.5 thousand. Wealth and autocratic rule have permitted Nazarbaev to successfully resist a radical Kazakh nationalist project and Salafi fanaticism, factors that would have accelerated the slow demographic disappearance of ethnic Russians into a cascade.

5 Ukraine – voice dominates exit and language maintenance across generations

Ukrainians in both imperial and Soviet times had the privilege of what I dubbed “most favored lord” status. Ambitious titular elites were able, with minimal transition costs, to cultivate career ambitions anywhere in the Soviet Union. In the imperial era many Ukrainian parvenus were able to procure noble status from the tsar. By the later Stalinist era (1939-1953), there was mass famine and repression, with Ukrainians absent from the Politburo. But by the mid-1950s, Ukrainians became “junior elder brothers” compared to the lesser nationalities and got more top positions in all republics than any other non-Russian nationality. The Kuban region of Russia and the south and east oblasts of Ukraine were populated by mixed populations based on language (rodnoi yazyk) and nationality. While Besançon (Reference Besançon1978) viewed Soviet Ukraine in terms of a province analogous to those in the Austrian empire, I characterized it as closer to the Languedoc region in France with the possibility of a common Slavic-based cosmopolitan identity (Weber Reference Weber1976).

Unlike Languedoc, however, Ukrainian peasants were not transformed into Russians. Given the policies of what Martin (Reference Martin2001) dubbed “the affirmative action empire,” many expected a powerful and successful rationalization of Ukrainian culture in Ukraine. In Martin’s analysis of Soviet nationality policy, Galicia (in Western Ukraine) would be the “Piedmont” for the western borderlands that would bring cultural unity to the region. As a Union Republic, Ukraine’s national narrative persisted through the entire Soviet period. And in line with prediction (and supported by field notes collected by Dominique Arel), the post-independence Kyiv government pushed the Galician historical narrative celebrating the “heroism” of the ultra-nationalist and anti-Soviet Ukrainian insurgency in World War II. In terms of language, Ukrainian school authorities worked to rid the national language of Russian influences that were imparted in the 1930s. With a strong national narrative and only minimal costs for Russians to learn Ukrainian, assimilation of Russian-speakers into a Ukrainian-speaking nationality was therefore in the cards.

Nonetheless I did not foresee success of the Piedmont strategy. In the Hirschman trichotomy, for Russian-speaking Ukrainians, exit was unattractive to a crumbling Russia and complex to an attractive Europe; and loyalty to the nationalizing project was seen as useless. In the matched guise experiment, on average, Russians downgraded fellow Russians somewhat for speaking in Ukrainian compared to speaking in Russian. Why, one Russian-speaker from Kyiv asked me, should our children read Pushkin in Ukrainian? As the remaining alternative, voice was the preferred option. I therefore foresaw a political backlash in the eastern and southern regions of Ukraine whose Russian-speakers (both Russian and Ukrainian in declared nationality) were pressing for a political bargain (and this is the voice) in which they would commit to loyalty to the Ukrainian state in return for cultural (and linguistic) autonomy in their regions. The west would be Ukrainian in language; the east and southeast Russian; and Kyiv would be (as with Brussels in the Belgian analogy I offered) a bi-lingual city. Consociation rather than rationalization in Ukrainian was my predicted outcome. The Party of Regions, strongest in the east and south, emphasized the rights of Russian-speakers in Ukraine, and supported a form of federalism. Its government passed the 2012 language law granting Russian regional status, and this was a step in favor of the Belgian model. Strong popular rejection of Ukraine as a unitary state in polls conducted in the Donbas in 2014 (Giuliano Reference Giuliano2018, 167) reflected a mass base for a federalized, or at least decentralized state.

Although the Brussels model did not transpire, nor did the evolutionary progress of the Ukrainian national project, at least as foreseen by Ponarin (Reference Ponarin2000). In part, this was due to President Victor Yushchenko’s contempt (along with many of the Kyiv protesters in the Orange Revolution and later in Maidan Square) for the eastern voters and the subsequent parliamentary vote against the 2012 language compromise. This played a role in activating an insurgency in the Donbas leading to declarations of sovereignty by the Peoples’ Republic of Donetsk and the Peoples’ Republic of Luhansk, in accord with the views of a significant minority of the region’s population (Giuliano, Reference Giuliano2018). But the Russian invasion of the Crimea, its inhuman role in the shooting down of Malaysia Airlines Flight 17, and its duplicitous role in fomenting the violence in the Donbas, militarized an already tense relationship between Ukraine’s regions. In view of the substantial constituency against the Ukrainian nationalizing project in Donbas, my prediction that that Russian-speakers in the eastern region would not assimilate into Ukrainian-speakers was on the mark.

However, upsetting my prediction, a non-peaceful merging toward a common Ukrainian identity may now be taking place, as suggested by a set of surveys conducted by the Kyiv International Institute of Sociology spanning 2012-2017. In the latest survey, well after the military occupation of Crimea and the onset of war in the Donbas (though lacking data from the war zones), the data reveal a two-step process from respondents seeing themselves as Russian/Ukrainian hybrids to unequivocal Ukrainians. The resolute abandonment of Russian ethnic and linguistic identities, even in the east, “can be viewed as the more or less conscious rejection of Moscow’s claim to ‘protect’ them, in stronger words, the refusal to identify with the entity they came to perceive as the enemy.” In consequence, support for a Ukrainian nationalizing project is growing (Kulyk Reference Kulyk2018, 134-5).Footnote ⁷

The war in the Donbas speaks to my final prediction.

6 As long as Russia could neither commit nor ignore the rights of its “co-fatherlanders” in the Near Abroad, the likelihood of ethnic warfare between titulars and Russians would be minimal

Extending my analysis of national identities, and relying on Brubaker’s (Reference Brubaker1995) framework, I sought to reckon the possibility of ethnic violence in the newly independent republics. Brubaker portrayed the post-Soviet world analogous to the structure of ethnic conflict in the interwar years in Eastern Europe. This structure was a “triadic configuration” among “national minorities” (e.g. Germans in the Sudetenland), nationalizing states (e.g. Czechoslovakia), and “external national homelands” (e.g. Germany). To the extent that the government of the national homeland seeks to protect its ethnic kin from the nationalizing impulses of the state housing the national minority, the possibility of civil war (induced by the national minority against the nationalizing state) and interstate war (between the national homeland and the nationalizing state) is ever present.

Van Houten (Reference Van Houten1998) formalized Brubaker’s framework and specified the parameter space in which ethnic violence was likely. His strategic analysis pointed to the degree of capacity and interest in the fate of the national minority by its national homeland. On the one hand, the higher the interest, the more emboldened the minority will be to initiate violence, create martyrs, and beg for military support from their putative homeland. On the other hand, the lower the interest, the more emboldened the nationalizing state will be to repress the national minority, inducing them to insurgency. Consequently, according to van Houten’s model, if the government of the national homeland takes a middle path between solid support and complete indifference to their ethnic brethren now living in nationalizing states, it has, according to this logic, done its best to guarantee peace.

Relying on this model, and noting that (as of my writing in 1998) Russia remained too weak to credibly commit to support its “co-fatherlanders” in the near abroad should they seek to fight for autonomy, and the nationalizing states were too constrained against minority repression by pressures from European institutions that they hoped to join, I foresaw no uprising by the Russian-speaking populations.

Indeed, the Russian Federation sent mixed signals to their “co-fatherlanders” in the near abroad. In the summer of 1992, Andrei Kozyrev, the Minister of Foreign Affairs, emphasized diplomatic resolution to all issues involving Russian minority populations in the former Union Republics. But at the same time, the Ministry of Defense was drafting a new military doctrine with protection of Russians living abroad as a principal justification for military action. In light of these contradictory signals, neither the titulars nor the Russian minorities had sufficient surety about Russia’s likely response to aggress upon the other. In light of this analysis, and because the parameter space specified by van Houten was peaceful (ambiguous signals from Moscow; low repression by the nationalizing states), I foresaw only a low probability of Russian troops moving into the republics of the near abroad to protect or incorporate their co-fatherlanders in a way similar to the German march into the Sudetenland in 1938.

Subsequent to publication of Identity in Formation, two factors pushed the variables in van Houten’s model in opposite directions. In Estonia and Latvia, NATO membership, while it increased the costs to Russia of attack, by its security guarantee decreased the incentives of the Baltic States to accommodate Russian-speakers. However, the prospect of (and eventual success in attaining) European Union membership put the Baltic states in a regulatory regime that demanded minority protections.Footnote ⁸ Moreover, as mentioned earlier, EU membership (after a while) permitted all Estonian and Latvian citizens (with increasing number of Russian-speakers qualifying) to live and work in Western Europe. This opened up vast opportunities for the Russian-speaking minorities. And so, greater protection from an attack by Russia, greater minority protection in the Baltic States, and a vastly expanded labor market for all Baltic citizens have kept the Baltic States in the safe zone of the parameter space.

Ukraine tells us a different story. In the hope of following the Baltic States on the route to Western Europe, and pushed by popular mobilization in Kyiv, the interests of eastern Ukrainian citizens were discredited. Russian media (fully available in eastern Ukraine) called this everything from minority repression to fascism. The latent Russian nationalism in this region (in part articulated by the Party of the Regions) was aroused. Meanwhile, after Russia’s bungled invasion of Georgia, President Putin put enormous effort in rebuilding the Russian Federal army into a modern force. This gave him the interest and capability – after showing his cards in Crimea – to incite his co-fatherlanders in the Donbas to rebel.Footnote ⁹ And because I did not predict a civil war in Ukraine, I did not model how Russian duplicity would help to unite Ukrainians from east and west.

In sum, although I did not foresee war in Ukraine, the van Houten model, with the parameter shifts, correctly pointed to the conditions that would threaten inter-nationality peace.

On Assimilation Cascades

Survey data in Estonia (Kalmus et al. Reference Kalmus, Lairistin, Signe and Vihalemm2020) challenged my proposed mechanism of assimilation, which I called “competitive assimilation.” This theoretical model motivated my prediction for assimilation cascades in the Baltics. This game theoretic model assumed, as was the case in the Baltics, a vibrant labor market for white-collar workers that advantaged Russian-speakers who had full facility in the titular language. Russian-speakers would have preferred a linguistically segmented labor market that provided them a full range of opportunities, but as each family defected from this vision by enrolling their kids in the few open slots in Baltic medium schools, there would be a race for early success in the titular-led labor market, leading to an assimilation cascade.

In a critique of my work, the research team led by Kalmus argued that a single model was inadequate. Instead, there have been multiple paths of accommodation to the Soviet collapse by the Russian-speaking minority. Some 21% of them are apathetic, nostalgic for Soviet culture, and lack ambition to crack the white-collar labor market. Meanwhile 26% of respondents are proud of Russian culture with its ties to the Orthodox Church. They have technical skills, good jobs, and are loyal to Estonian institutions, even though 90% of them have no facility in its language. A related group of 35% balances loyalty to the Estonian state with political engagement to protect their minority rights, though a third of them retain their Soviet-era passports. According to Kalmus et al.’s data, it is only a final group of around 19% who have cascaded toward a full Estonian identity. There is (as of writing) no evidence of this final group influencing the choices of the other three groups to assimilate before their children lose future opportunities. As noted earlier, this is in part due to an option – emigrating to other EU countries – that I did not foresee. In any event, undermining my theoretical expectations, cultural shift in Estonia is more of a slow linear process rather than an assimilation cascade. In this regard, Pål Kolstø’s judgment at the 2018 Association for the Study of Nationalities panel was profound: Identity in Formation was prescient, but the predictions it offered were not derived from my abstract model.Footnote ¹⁰

On the predicted exit cascade from Kazakhstan

As reported in my 2018 paper, there were Kazakhstan was 42.7% Russian in the 1959 census, 29.9% after a decade of independence in 1999, and 21.5% in 2014. This went down to 14.9% in 2024 (as was reported under “Ethnic Demography of Kazakhstan” in Wikipedia on April 8, 2025). Several factors have sustained the exit of Russians from Kazakhstan. First, since the early 2010s, Kazakhstan has been spearheading greater inter-regional cooperation in Central Asia that does not include Russia – partly in reaction to Russia’s 2014 invasion of Crimea. Second, the Russian-led intervention of Collective Security Treaty Organization troops to protect President Kassym-Jomart Tokayev in mass protests in Kazakhstan in January 2022 led to a backlash. Although the protests were over economic issues, it raised red flags concerning Russia’s interest in dominating the country. To avoid appearing as if his country were in Moscow’s orbit, Tokayev refused to back Russia’s war against Ukraine. Kazakhstan abstained in UN votes to condemn Russia’s invasion; yet no troops were committed despite President Putin’s demands. Tokayev refused to help Russia evade sanctions, and blocked weapons sales to Russia (Jones and Smyth, Reference Jones and Smyth2022). Kazakhstan has also been spearheading greater inter-regional cooperation in Central Asia that does not include Russia, seeking closer military alliances with Uzbekistan as well as other non-Russian partners (Jones and Smyth Reference Jones and Smyth2022). Furthermore, the Kazakhstani government has increasingly been giving new Kazakh names to cities and villages and inducing ethnic Kazakhs from other regions and abroad (from China, Russia, and Uzbekistan) to move to where ethnic Russians live (Daminov Reference Daminov2021). Russian-speakers in Ukraine were largely unnerved by this volte face in Kazakh policy. In addition, the influx of anti-war Russians escaping the draft for the Ukrainian war has exacerbated ethnic tensions on the border with Russia. Jones and Smyth (Reference Jones and Smyth2022) report that during an anti-war demonstration in March 2022, showing fear of a Russian invasion, Kazakhstanis held signs that read, “Yesterday Georgia, today Ukraine, tomorrow Kazakhstan?” Finally, the delicate interethnic balance fostered by decades of rule by Nursultan Nazarbayev has been eroding, marked by several interethnic riots among minority populations in the country’s south (Daminov Reference Daminov2021; Lim Reference Lim2021). These factors are reflected in public opinion. Survey evidence (relying on a clever list experiment) showed that while 64% of ethnic Russians supported Kazakhstan’s membership in the Russian-dominated Collective Security Treaty Organization, this compared to only 42% of ethnic Kazakhs. The unstable ethnic balance in the country sends a signal to the Russians that they cannot rely on future state support for their livelihoods, and these factors help explain what I called their “contingency plans for exit” (Laitin Reference Laitin1998, 361). This is not quite the cascade of the pieds noirs from Algeria, but with the same threat of a nationalizing program by the indigenous government. In this case, contra Pål Kolstø, the model did offer insight on the increasing incentives for Russian-speakers in Kazakhstan to exit.

On internationality violence

My prediction of a Belgian-like consociation in Ukraine appears today egregiously naïve. The portrayal of a Russian-speaking east, a Ukrainian-speaking west, and a bilingual Kyiv turned out to be a fantasy. What did I get wrong? First, I did not appreciate the unwillingness of Ukraine’s western elites to consider a federal bargain. Federalism was resolutely opposed by President Yushchenko and virtually by the entire western population. Although agreeing on the boundaries of the federal units would have engendered conflict, these could have been reconciled with plebiscites. The vote to repeal the 2012 language law in 2014 ended any chance for federalism and was a spark to the subsequent uprising in the Donbas. Second, and this is shown in Arel and Driscoll (Reference Arel and Driscoll2023), I did not factor into my analysis the degree of miscalculation by both sides of the conflict despite centuries of cultural and political interaction. (They knew each other well.) Putin surely miscalculated the degree to which identity as a Russian (russki) implied loyalty to or desired membership in Russia (rossiane) and wrongly assumed that the Russian-speakers in eastern Ukraine were natural allies for his wider project of reconstituting the Soviet world (sovetskiy mir). Post-2022 surveys and focus groups in Ukraine revealed a growing unity across Russian- and Ukrainian-speakers on a ethnonationalist identity, an outcome that undermined Putin’s expectations (Kulyk Reference Kulyk2023). In a parallel manner, both post-Maidan governments in Kyiv miscalculated the Kremlin mission (and Putin’s obsession) to restore Russia’s historic boundaries, even if it involved major warfare.

Was the van Houten model discredited by history? I think not. Putin went on the offensive consistent with his belief that Russia’s reconstituted army as it functioned in the Georgia war of 2008 was ready to decimate Ukrainian forces. This view was not completely unfounded.Footnote ¹¹ And again consistent with van Houten, Kyiv overestimated its support from western democracies to counter a Russian invasion, given signals from the US Ambassador on future membership in NATO.Footnote ¹² The model correctly portrayed the strategic situation faced by those in the triadic configuration – Russia, Ukraine, and the Russian-speakers in the Donbas – but the parameter values supporting two decades of uneasy peace changed, incentivizing the risk of war.

I want to thank Harris Mylonas, editor of Nationalities Papers, for giving me a chance to rethink my analyses of the political aftermath of the Soviet world.