The Soviet collectivization campaigns of the early 1930s brought death to millions throughout the USSR, but Ukrainians and Kazakhs suffered the most. According to best estimates, collectivization and the subsequent famine caused the deaths of between 2.6 and 3.9 million people in Ukraine between 1930 and 1933 (Himka Reference Himka2013, 426). In Kazakhstan, the campaigns to settle and collectivize the nomadic population caused at least 1.5 million deaths, of which at least 1.3 million were ethnic Kazakhs, or one-third of the Kazakh population at the time (Cameron Reference Cameron2016, 120).Footnote ¹ At least another 600,000 people fled the Kazakh Autonomous Republic to other parts of the Soviet Union, China, or other neighboring countries (Ohayon Reference Ohayon2013, 3).

Given these horrific numbers, it is striking how differently these events are commemorated in the two countries. In Ukraine, collectivization is widely remembered as Holodomor, a genocidal policy that targeted the Ukrainian nation. Most large cities have at least one memorial to the victims of the famine, and the memorials, plaques, and other physical objects in Ukraine commemorating their deaths number into the thousands (Nechet Reference Nechet2014). In Kazakhstan, the memory of the famine is far less prominent in the public sphere. Monuments have been erected only recently in Astana (2012), Almaty (2017), and a few other cities, while official accounts have largely avoided accusations of a deliberate genocide. Instead, the famine has more frequently been described as a tragedy or perhaps a crime perpetrated by overzealous local officials.

To some extent, these differing commemorative strategies reflect divergent scholarly interpretations of collectivization in the two countries. The argument for genocide in Ukraine maintains that nationalist resistance during the civil war, including declarations of independence, massive peasant revolts, and a brief alliance with the newly independent Polish state, led Stalin to believe Ukrainians would betray the Soviet Union in the event of another war, and so used mass starvation to preempt such treachery. Thus, in late 1932, knowing full well the impact collectivization was having on the Ukrainian countryside, the Kremlin made new, harsher demands that required local authorities to confiscate nearly all foodstuffs from peasant households, “blacklisted” hundreds of villages from buying manufactured goods or trading with neighboring villages, and prohibited starving peasants from fleeing to Poland or even to cities in Ukraine (Applebaum Reference Applebaum2017, esp. 186–204). But not every scholar believes collectivization was a premeditated attack on the Ukrainian people. Terry Martin (Reference Martin2001, 302–306) argues that the directives against Ukraine, and the purges that followed them, were not so much a deliberate design as a cruel response to local resistance, which was much greater in Ukraine (and in Kazakhstan) than in Russia. Thus, in Martin’s view: “The Kremlin’s ‘national interpretation’ was not a cause of the grain requisitions. Rather, it emerged as a consequence of it” (Reference Martin2001, 303).Footnote ²

In Kazakhstan, too, Moscow persisted in its murderous policies despite full knowledge of their consequences. In the Kazakh case, though, most scholars agree that “mass death … was not an objective the policy makers set out to achieve, but rather the price they were prepared to pay” to achieve their goals (Pianciola Reference Pianciola2001, 246; see also Cameron Reference Cameron2010; Ohayon Reference Ohayon2013; Payne Reference Payne, Alexopoulos, Hessler and Tomoff2011). Soviet ideology considered the nomadic practices of the Kazakh herdsmen both “backward” and unproductive. In 1929, the regime had ordered the nomads to be settled for more productive occupations, but, given the emphasis on rapid industrialization, the policy remained a low priority for local officials and was implemented slowly and inconsistently (Pianciola Reference Pianciola2018). The products of the Kazakh steppes, though, were important. Beginning in 1930, Moscow sharply escalated its demands for Kazakh livestock, partly for use in the collective farms, but mostly to provide meat for urban areas in central Russia (Pianciola Reference Pianciola2018). In response, many Kazakhs slaughtered their animals themselves. Overall, the procurement campaigns reduced Kazakh herds by almost 90%, and the herdsmen themselves were left to their own devices. Meanwhile, those people who were settled on collective farms often found themselves on marginal agricultural land and subject to many of the same harsh requisition measures found in Ukraine (Cameron Reference Cameron2010, 224). Under these circumstances, many Kazakhs left their home territories in search of food.Footnote ³

These divergent accounts do not explain the differing commemorative strategies in the two countries, however. Stalin’s intentions have been disputed in both countries, and in both countries public memory has changed several times since independence. Even so, the overall reluctance of the regime in Astana to call it a genocide remains striking. Evgeny Finkel (Reference Finkel2010) points out that accusations of genocide have been abundant in the former Soviet Union and Central East Europe since 1991, so why have Kazakhstani officials neglected to join in?

One common explanation is that President Nursultan Nazarbayev has dampened discussion of the famine in order to maintain good relations with Russia, particularly given the large number of Russian speakers living in the country (Kindler Reference Kindler2012, 107–109). Indeed, in both countries, public representations of the famine have tracked closely with the representers’ attitudes toward Russia. In Ukraine, the most ardent advocates of the genocide argument portray Russia as a hostile, imperialist power, whereas Ukrainians who look on Russia more favorably have either rejected the genocide thesis or at least toned it down. Yet the correlation still raises a question of cause and consequence. Both countries have long borders with the Russian Federation and a large number of Russian speakers living in their territory. Both would have reasons to use this history either to placate or confront their powerful neighbor. Why, then, has Ukraine chosen a confrontational strategy whereas Kazakhstan chosen an accommodationist one?

This article focuses on public memory. Public memory, as Michael Bernhard and Jan Kubik (Reference Bernhard and Kubik2014, 8) define it, “combines most importantly the official memory strategies propagated by holders and competitors of state power, as well as the cultural memory propagated within the public sphere.” Unlike individual memories, public memory requires an active effort to present historical events in a particular way. Unlike memories circulating within families and other social networks outside the public eye, public memories are available to a broad section of the population, and, at times, even imposed on a public sphere that cannot easily avoid them (Forest, Johnson, and Till Reference Forest, Johnson and Till2004).

In short, public memory does not just happen; it is work. But what kind of work does it do (LaFreniere Reference LaFreniere2012, 10)? Public commemorations provide an overarching narrative upon which individual members of a population can attach their own life stories, thereby creating a “memory community” whose members share a common past and, implicitly, a common future. Such a narrative establishes the community’s roots deep in the past and lends it a degree of permanence and stability beyond the vagaries of contemporary politics. It also sets community boundaries and shapes the perceptions of others outside these boundaries. In short, the narrative helps frame the public debate to make some policies seem more appropriate and desirable, and others less so. Not surprisingly, the construction of a public narrative is often contested and politicized.

Though the politics of memory has become a prominent subject of research in such diverse fields as history, anthropology, sociology, and cultural studies, it has received strangely little interest in political science (Bernhard and Kubik Reference Bernhard and Kubik2014, 2). This article examines two important exceptions: (1) a model developed by Michael Bernhard and Jan Kubik (Reference Bernhard and Kubik2014) in the edited volume Twenty Years after Communism and (2) a model developed by Keith Darden and Anna Grzymala-Busse (Reference Darden and Grzymala-Busse2006) in their article, “The Great Divide.” Both models explain variation in how the past is mobilized in Central and Eastern Europe in the post-communist era. Both recognize the presence of contemporary political calculation as well as cultural and historical factors in determining differing patterns of public memory. Whereas the Bernhard and Kubik model tends to privilege an instrumental explanation, however, Darden and Grzymala-Busse emphasize historical factors.

The Politics of Public Memory

Bernhard and Kubik (Reference Bernhard and Kubik2014) developed their model to explain variations in how different countries in the former Soviet bloc marked the 20th anniversary of the collapse of communism. As their dependent variable, the authors identify three types of memory regime, defined as “the dominant pattern of memory politics that exists in a given society at a given moment in reference to a specific highly consequential event or process” (Bernhard and Kubik Reference Bernhard and Kubik2014, 4). The first, a fractured regime, occurs when the public memory around a given topic is contested. A society is most likely to have a fractured regime when there is at least one mnemonic warrior in the public debate. The warrior narrative presents an absolutist view of historical truth that draws a sharp distinction between those within the memory community and perceived hostile forces outside it.

Second, a unified regime exists where there are no public contests over the historical narrative regarding the event or process in question. A unified regime exists when all relevant actors are mnemonic abnegators, that is, pragmatists who regard the past as largely irrelevant to resolving contemporary problems. The abnegator’s choice may reflect widespread social consensus about the interpretation of the past, or it may reflect a pragmatic decision to avoid memory politics in the name of social stability. Finally, a pillarized regime describes a situation where competing narratives coexist, but that “toleration of differences of opinion or indifference over memory issues will prevent the partisan politicization of such interpretations” (Bernhard and Kubik Reference Bernhard and Kubik2014, 17).

To explain why political actors choose the strategies they do, Bernhard and Kubik assume that political elites draw on public memory to enhance their influence. They propose two sets of constraints that affect elite choices: structural constraints, which reflect contemporary political realities, and cultural constraints and opportunities, which originate from preexisting discourses within society (Bernhard and Kubik Reference Bernhard and Kubik2014, 7–36). The structural constraints include (1) the type of state socialism existing prior to the transition, (2) the mode of exit from state socialism toward something else, and (3) the depth of political cleavages that exist at the time of commemoration. The first two address the likelihood that political contestation of national narratives would have occurred prior to independence—producing hypotheses that a more reformist communist regime or a more contested exit from communism would be more likely to result in a pillarized or fractured regime. The third variable, the depth of cleavages at the time of commemoration, speaks to the ability and motivation of political actors to produce contesting narratives in political competition. The greater the divide between the actors, the more likely they will be motivated to adopt a warrior stance (Bernhard and Kubik Reference Bernhard and Kubik2014, 19–22).

Proceeding from these hypotheses, the authors further maintain that statist regimes dominated by a single “party of power” are likely to select a strategy that will favor abnegation over conflict or pluralism. Because parties of power often include a significant population of former members of the nomenklatura, it makes sense they would play down the repressive elements of the previous regime and focus instead on the need for stability and future economic development (Bernhard and Kubik, Reference Bernhard and Kubik2014, 27–29). Although Bernhard and Kubik limit their model to electoral democracies, this particular hypothesis could be extended to authoritarian regimes as well. Diana Kudaibergenova (Reference Kudaibergenova and Laruelle2016), in comparing Kazakhstan and Latvia, argues that both regimes had incentives to promote a nationalizing discourse after independence—that is, a discourse that bolsters a regime’s claims to rule in the name of a particular national group (Brubaker Reference Brubaker1995, 114). She further argues, however, that in a polity containing multiple ethnic identities, an authoritarian regime likely would choose a strategy of abnegation to demobilize the population, whereas a more open electoral process would create incentives for politicians to mobilize supporters with a warrior narrative.

These structural constraints help explain why political actors decide to contest, tolerate, or suppress conflicting narratives, but they do not explain what the content of these alternatives might be, what values they hope to project, or why some narratives are more successful than others. To explain this, Bernhard and Kubik offer a set of cultural constraints on political actors defined as “a historically formed repertoire of cultural (mnemonic) forms and themes” (Reference Bernhard and Kubik2014, 22). If politicians present narratives outside this repertoire, their appeal will seem alien to the public and likely fail. Unfortunately, Bernhard and Kubik do not adequately develop this variable. They do suggest that a fractured regime is more likely to occur with the existence of linguistic, ethnic, or religious cleavages, but those categories are too general to be useful.

One of the most powerful models explaining how cultural and historical constraints influenced the strategies available to politicians in post-communist countries can be found in a study by Darden and Grzymala-Busse (Reference Darden and Grzymala-Busse2006). These authors proceed from the widely held argument that the most important factor leading to a successful transition from communism to democracy was the presence of an organized liberal opposition before the communist regime collapsed (Vachudová Reference Vachudová2005, 13). Observing that these opposition movements all shared a strong nationalist discourse that identified communist rule as foreign and unacceptable, they then asked what explains the emergence of such a discourse. They answer this question by ascertaining first whether mass literacy in the community emerged before or during the communist regime, and second what values the nationalist ideology taught to that first literate generation.

The emergence of a literate population before communist rule, they argue, enabled elites to construct a memory community associated with political nationalism. Only with mass literacy could the elite create more standardized discourses and communicate them to a large portion of the population without face-to-face contact. Mass literacy also “dramatically increased a society’s capacity to record and convey history, literature and myth” (Darden and Grzymala-Busse Reference Darden and Grzymala-Busse2006, 98). The national identity inculcated into this first literate generation then “immunized” that generation and its followers against subsequent efforts to introduce alternative versions of the national narrative (Darden and Grzymala-Busse Reference Darden and Grzymala-Busse2006, 100).

The presence of mass literacy prior to communist rule is necessary but not sufficient to explain the rise of a nationalist opposition in the late 1980s. One must also discern the content of national values prevailing in that formative period, particularly how these values pertained to Russia and the Soviet Union: “First, was communism a domestic development or was it the imposition of an occupying force? Second, was its sponsor, the Soviet Union, a historical enemy or a more friendly power? Third, and most important, did communism represent a political, cultural and economic advancement for the nation or was it an antimodern step backwards” (Darden and Grzymala-Busse Reference Darden and Grzymala-Busse2006, 102). In fact, the authors found a strikingly high correlation between those societies where mass literacy had been achieved prior to communist rule and the emergence of nationalist oppositions during the Gorbachev era. Still, it is unclear if the onset of mass literacy is always a causal variable or rather a useful proxy for a much broader process of nation-building among a mass audience. In western Ukraine, for example, Jean-Paul Himka (Reference Himka1988, 217) describes how in Eastern Galicia—now Western Ukraine—civic organizations, local notables, and the penetration of economic markets gave rise to a Ukrainian nationalist movement in the countryside before most of the population could read or write.Footnote ⁴ For this reason, this article looks not only at mass literacy, but more broadly at the institutions and discourses in place when the national identity of a given community first became a category of social significance beyond a small elite.

Also, despite the predictive power of the model, it does not explain well how the ideas and values entrenched in the formative years of nation-building endure. As Fredrik Barth (Reference Barth1969) argued, the boundaries between groups may remain stable even as the content of what defines them changes significantly. To establish continuity in the discourses of that first literate generation, then, it is necessary to demonstrate how they have been reinforced even in the face of changing institutions and practices. If the continuity is broken, one can then investigate when and how this happened.

Memory Regimes and Structure in Kazakhstan and Ukraine

Combining these two models explains a great deal of the variation in Ukraine and Kazakhstan’s memory regimes around collectivization. This section examines the progress of public commemoration of the famine in Ukraine and Kazakhstan in light of the structural constraints posited in Bernhard and Kubik’s model. Given that both Ukraine and Kazakhstan had been part of the Soviet Union, the first of these structural constraints, namely the nature of the former communist regime, was nearly identical.

The Soviet narratives of collectivization in the two republics were also similar. Collectivization was described as a progressive process that introduced more advanced property relations into the countryside and extracted the resources necessary for rapid industrialization. The issue of mass starvation rarely appeared in public, and when it did, it was usually attributed to the resistance of kulaks, bad weather, or, particularly in Kazakhstan, the overexuberance of local officials (Mark Reference Mark2004, 120; Vasil’ev Reference Vasil’ev2004, 165). Only during Gorbachev’s perestroika did historians and journalists begin to explore the extremity of the suffering in both these countries. Even then, official accounts usually described the brutality of forcible collectivization as a departure from Leninist norms that could be attributed to Stalin’s personal failings rather than from more systemic failings of the regime itself (Vasil’ev Reference Vasil’ev2004, 166).

Historical Constraints on Public Memory

The previous discussion shows that the structural constraints proposed by Bernhard and Kubik (Reference Bernhard and Kubik2014) do a good job of showing how instrumental political calculations can track commemorative strategies within a particular cultural field. In both Ukraine and Kazakhstan, the victory of former communists, oligarchs, and industrialists after independence led to an abnegating strategy that favored the depoliticization of memory over mobilization. In Ukraine, the emergence of a strong, organized opposition during the transition period created the structural preconditions for the fractured memory regime that emerged in 2004. The political contests during this latter period also support the hypothesis that political polarization exacerbates memory wars. Yushchenko’s efforts to make the nationalist narrative the standard history throughout the country provided new ammunition to his opponents at the local level, with the result that public memory became an important fault-line in the political landscape, even when many citizens were not committed to either narrative. In Kazakhstan, by contrast, Nazarbayev’s firm grip on power, both before and after the transition, ensured that contests over memory remained limited and that his vision of unity and harmony remained dominant.

However, the instrumental approaches do not help us understand where different narratives come from, which ones succeed, and which ones do not. For example, Bernhard and Kubik’s model does not explain why the Ukrainian genocide narrative should gain so much currency in Western Ukraine, which did not experience the famine, but not in Kharkiv, where the Holodomor was particularly devastating. Nor does it explain why the resistance to Nazarbayev’s policy in Kazakhstan was not stronger. Indeed, according to Shirin Akiner, many Kazakhs remembered the horrors of collectivization and sedentarization “with a curious sense of detachment, as though it were not recorded in actual, recorded time, but rather in some abstract past” (1995, 63; see also Ohayon Reference Ohayon2013, 7).

These questions are better addressed by the arguments of Darden and Grzymala-Busse (Reference Darden and Grzymala-Busse2006). The timing of mass literacy is particularly useful in explaining why early pressure for commemorating the famine in Kazakhstan was not stronger. The low rates of literacy among famine victims greatly assisted Soviet efforts to suppress and shape the memory of collectivization later on. First, as mentioned above, it ensured that relatively few eyewitness accounts from the victims’ points of view survived the famine. Second, destroying the society’s nomadic way of life made the survivors both accessible and vulnerable to state institutions. The primary schools and other public institutions enveloping the newly literate Kazakhs all conformed to and propagated the Soviet historical narrative of progress and economic development. In the urban areas, moreover, the preferred language of instruction was Russian, and the incentives for adopting a Soviet identity were huge. Kazakh language and culture were often stigmatized as backward, while a knowledge of Russian opened the doors to social mobility and was “seen as providing a broader vision and a sense of empowerment” (Dave Reference Dave2007, 48; see also Akiner Reference Akiner1995, 40–43).

This legacy is evident even today. In March 2018, a survey of over 2,000 respondents conducted by Demoscope found that 44 percent of the population still felt positively about Stalin because he helped make the Soviet Union a superpower, and only 17 percent condemned Stalin for his actions against the people of Kazakhstan; 37 percent agreed with the abnegating position that “this is the historical past that does not affect our lives” (Demoscope 2018). It is not surprising, then, that many in Kazakh society, particularly in the early stages of the post-communist era, should view the nomadic past with “ambivalence”: while nomadism remains “the center of Kazakh identity and group cohesion,” it also conflicts somewhat with the modern and cosmopolitan representation of contemporary Kazakhstan (Ohayon Reference Ohayon2013, 7).

The argument by Darden and Gryzmala-Busse also gives us insight into the geographic divide in Ukraine. Both the Russian and the Austro-Hungarian imperial state had begun to extend their legitimacy to a mass population in the last decades of the 19th century. In those parts of Ukraine under Habsburg rule, as noted earlier, imperial policies tolerated and even encouraged the formation of political parties, gymnastic clubs, and other civic institutions organized around ethnolinguistic identities, such that Ukrainians, Poles, and Jews in the region before World War I “increasingly saw themselves in modern national terms” (Amar Reference Amar2015, 29; see also Bartov Reference Bartov2018, 6–36).

In those parts of Ukraine under Russian sovereignty, by contrast, the imperial government sought to integrate the Ukrainian (or “Little Russian”) population into a broader East Slavic identity that centered on Russian language and culture. Ukrainian was regarded a dialect of Russian rather than a separate language and was banned both from primary education and even from most printed material. These integration efforts were most successful in the urban areas of what is now south and east Ukraine, where the beginning of industrialization had attracted Russian workers and enabled the state to penetrate more deeply into society. By contrast, the state had not penetrated so deeply in rural areas of Central Ukraine, where literacy rates remained low (Miller Reference Miller2003, 254–258). According to the imperial census of 1897, “81% of Ukrainian speakers over the age of eighteen often could not read or write” (Liber Reference Liber2016, 35). Civil organizations in the Russian empire were also less dense than in Galicia. Although both Ukrainophile and All-Russian activists worked to mobilize a mass electorate for elections to the Duma after 1905, most rural districts elected local notables rather than members of a national party (Hillis Reference Hillis2013, 190, 204–205). When rural Ukrainians learned to read under Bolshevik rule, many of them were taught in Ukrainian, but the content of instruction praised Soviet rule as the vanguard of world history.

Thus, regardless of whether the Ukrainians of Imperial Russia learned to read and write before or after the Bolshevik Revolution, the expectations arising from Darden and Gryzmala-Busse are the same. In both cases the predominant discourses at the time of mass literacy would likely see Bolshevik rule as a domestic as opposed to an occupying force, and perceive Russia (or the Soviet Union) to be a friendly power and the communism experience to have brought economic progress from what existed earlier.

The Issue of Continuity

Darden and Grzymala-Busse assert that the identities formed in the first generation of mass literacy were “robust to the enormous upheavals of war, genocide and population transfers that surrounded World War II” (Reference Darden and Grzymala-Busse2006, 101). But how these identities are passed down from one generation to another is left underdeveloped.

In the case of Kazakhstan and eastern Ukraine, the continuity in attitudes toward Russia and the Soviet Union are easily explained by the persistence of Soviet power from the time of mass literacy through independence. The question of Western Ukraine is more puzzling: how and why should attitudes constructed under Habsburg rule remain robust in Ukraine today? It is true that the nationalist rivalries between Ukrainians, Poles, and Jews that emerged under Habsburg rule were reinforced repeatedly before the onset of communism by the discriminatory policies of the interwar Polish state, the multiple occupations of World War II, and the internecine conflicts that happened after the war. It is also true that the population of Western Ukraine experienced Sovietization as an imposition from a foreign invader, and that many holders of that identity saw the imposition of rule from Moscow as a developmental step backward (Amar Reference Amar2015, 77; Bartov Reference Bartov2018, 129–157, 284–288). Finally, it is clear that, at least in the countryside and smaller towns, many people still remember the brutal Sovietization struggles of the early postwar years (Bartov Reference Bartov2018, 284–302).

In other respects, however, West Ukraine in 1990 differed utterly from Eastern Galicia of 1939. As William Jay Risch notes, Sovietization in Western Ukraine was both thorough and, for the most part, successful: after the devastation of the World War II, “this region lacked the prewar elites, institutions and practices that blunted the impact of Sovietization in the Baltics. Soviet Lviv did not have the resources of a republic capital like Riga, Tallinn or Vilnius” (Reference Risch2011, 45). Many of the people who populated Lviv before the war had fled, or had been murdered, deported, or expelled by 1946; the majority of the new Lviv residents were newcomers from the nearby countryside, eastern Ukraine, the Russian Federation, and other Soviet Republics.

Indeed, Risch (Reference Risch2011) and Amar (Reference Amar2015) have both argued in recent years that the distinct identity of West Ukraine (and they both agree that such an identity exists) was related only indirectly to its imperial past. The preservation of this identity through the Soviet years, and its content, was largely an “outcome of Soviet policies” (Risch Reference Risch2011, 254). In this argument, after World War II Soviet officials took great pains to Ukrainize what is now Western Ukraine; they wanted to divorce the region completely from its Polish past and root the new identity permanently in Soviet Ukraine. At the same time, however, Soviet officials regarded local Ukrainians as socially backward, as people who had not experienced the revolutionary transformation of the Stalinist 1930s, and who had to be instructed in the ways of the more advanced Soviet civilization. They also continued to suspect local residents of harboring nationalist views in opposition to Soviet rule (Amar Reference Amar2015). Newcomers from the East often got the best positions, the best apartments, and other perks of official life.

There were also more tangible differences between Western Ukraine and the rest of the Republic that contributed to a sense of difference. The region never became Russified to the extent that other parts of Ukraine were, and the use of Ukrainian was much more common in everyday interactions than elsewhere in the Republic. Lviv continued to have more connections to Poland and other countries in Eastern Europe than other parts of Soviet Ukraine. Even the architecture testified to Western Ukraine’s distinctive Habsburg history. As a result, the regime’s efforts to incorporate the region into Soviet Ukraine often highlighted the region’s distinct character.

Still, as Risch notes, the distinctive identity of the West Ukraine region did not turn “explicitly anti-Soviet” until the perestroika era (Reference Risch2011, 254). That anti-Soviet sentiment should arise at that time is no accident. Social ferment against the regime was not unique to Ukraine at the time, and the distinct West Ukrainian identity probably made it easier to overcome the difficulties in coordination and trust associated with the rise of social movements. The nationalist activists also had at their disposal a more or less coherent narrative around the Holodomor, and one that had been endorsed by the U.S. Congress, that they could adopt as a means to strengthen their demands for more autonomy from the center. Thus, the nationalists of Western Ukraine adopted the diaspora’s narrative on the Holodomor even though their agendas were quite distinct one from the other.