Conceptualizing Historical Legacies

Legacy Arguments

Researchers of post-communism have identified a vast number of communist legacies. Some of these can be labeled cultural, encompassing attitudes, beliefs, and knowledge inculcated during the communist period. Examples include a “ghetto” political culture, where the population views politics as dangerous and something to avoid; ¹ the skills to successfully navigate politics; ² the hybrid of nationalism and socialism that proved inimical to liberal values; ³ economic beliefs; ⁴ and trust in political parties. ⁵ Others might be termed material, such as the lack of infrastructure, the destruction of the environment, the dominance of the state sector, and excessive focus on heavy industry. ⁶ Still others could be called institutional, encompassing the persistence of old regime institutions, organizations, and elites throughout the economy, polity, and society. Examples include the bloated welfare system; ⁷ weak party systems; ⁸ communist-era constitutions; ⁹ and centralized economic planning. ¹⁰ This list is by no means exhaustive. Indeed, there have been some factors, such as ethnic fragmentation and natural resource endowment, ¹¹ that do not easily fit into the above categories, and there are certainly many other potential legacies. ¹²

Another important albeit less popular research area has been on pre-communist legacies. For example, Kitschelt et al. note how the choice of post-communist political institutional arrangements across East Europe is conditioned, ultimately, by the level of social and administrative modernization before communism. ¹³ Shugart illustrates how countries with a history of parliamentary governance tend to put greater authority in parliaments during the post-communist period than countries with no such history, which tend to center authority in the presidency. ¹⁴ Pop-Eleches reports evidence of the importance of interwar statehood for post-communist democratic success. ¹⁵ Darden links variation in pre-communist political socialization parts of the former Soviet Union with later patterns of nationalist conflict and voting behavior. ¹⁶ Kashin and Ziblatt trace the roots of depressed voter turnout across former East Germany to the presence of large landed estates in the nineteenth century. ¹⁷ Wittenberg shows how in Hungary post-communist patterns in support for parties of the Right resemble patterns established before communism. ¹⁸ For Bunce, the power of post-communist nationalism is a legacy of imperial rule in the region. ¹⁹ Jasiewicz traces the contemporary importance of religiosity in Polish voting behavior to processes set in motion in the nineteenth century. ²⁰

Abstracting away from the particulars of individual studies, including whether or not any given study has successfully made its case for the existence of a legacy, we can identify three components of a legacy-type argument. ²¹ One is an outcome (or pattern of outcomes) that appears inexplicable, or at least not fully explicable, given circumstances contemporaneous with that outcome. For example, why should populations in Eastern Europe after 1989 have been so mistrustful of politics and political parties? ²² One would not necessarily expect this given the fact that most of the parties were new, and that after roughly four decades of dictatorship the citizenry finally had an opportunity to determine its own fate. Jowitt argues that the mistrust was due to a Leninist legacy inherited from the past. Similarly, the propensity of some peoples to resist Soviet occupation more than others is likewise not readily explainable by reference to any obvious distribution of social or economic characteristics. ²³ Darden argues for national cohesion, a result of earlier literacy campaigns. The puzzling patterns of both voter turnout in the former German Democratic Republic ²⁴ and post-communist democratization ²⁵ have likewise been attributed to historical legacies.

A second component is a purported antecedent to the outcome that is identified as either a cause or a correlate of that outcome. The antecedent might take the form of a measurement of the outcome at a prior period, in which case it is claimed that the outcome has persisted. For example, Wittenberg found high correlations between post-communist support for rightist parties across Hungarian municipalities and electoral results from the last democratic national parliamentary election before the advent of state-socialism. ²⁶ In the case of interwar anti-Jewish discrimination in Germany, Voigtländer and Voth report a correlate in the pattern of anti-Jewish violence that occurred during the fourteenth-century Black Death epidemic, in which Jews were blamed for spreading disease. ²⁷

The identified antecedent might be a potential causal factor rather than a correlated outcome. For example, Grosfeld and Zhuravskaya link differences in contemporary support for conservative religious parties across Polish territories to whether the territory had once belonged to the Habsburg or to the Russian empire. ²⁸ Likewise, Becker et al. find that there is greater mass trust of public institutions in areas governed by the Habsburg Empire than in neighboring areas ruled by either the Ottoman or Russian empires. ²⁹ Peisakhin argues that Ukrainians on the formerly Habsburg side of the long-defunct border between the Habsburg and Russian empires have more antipathy to Russia (and greater sympathy for Ukrainian nationalism) than their Ukrainian neighbors who happen to live on the formerly Russian side, even though those areas had been in the Soviet Union (and undergone Soviet socialization) for decades. ³⁰

A third component is a mechanism (or at least a purported or hypothesized mechanism) that fills in the links leading from the antecedent to the outcome to be explained. For example, Voigtländer and Voth claim that medieval anti-Semitism disappeared in those German towns where trade openness raised the cost of discrimination against outsiders, and persisted into the interwar period, where such openness never took root. ³¹ Grosfeld and Zhuravskaya contend that support for religious parties in regions of Poland formerly in the Habsburg Empire can be traced back to Austria’s more tolerant attitude to the Roman Catholic Church, which led to higher church attendance and ultimately more conservative politics. ³² Likewise, Wittenberg maintains that pre-communist attachments to right-wing parties were more likely to survive state-socialism where the communists failed to destroy local church institutions. ³³ Peisakhin finds that pre-Soviet Ukrainian historical identities were transmitted even through the ideologically hostile Soviet regime if parents were consistent enough in the political messages they telegraphed to their children. ³⁴

To summarize, researchers construct legacy-type arguments with three components: an outcome that is not fully explainable from causes contemporaneous with that outcome, a cause or correlate that existed prior to the outcome, and potential (or at least speculative) links between the antecedent and the outcome. Whether or not the existence of a legacy has been established in any particular study depends, of course, on the quality of that study’s argument. How believable is the claim that the antecedent is correlated with or is causally related to the outcome? How plausible are the purported links between the antecedent and the outcome? By what criteria do we establish believability or plausibility? These are important questions whose answers are beyond the scope of this essay. What is important for present purposes is not the difficulty of identifying an antecedent or clarifying the mechanism but that outcomes, antecedents, and mechanisms form the scaffolding of legacy arguments.

Considering the three components together, we can identify two further characteristics of legacies. First, what we call the “legacy” is the outcome to be explained, not the antecedent or the mechanism linking antecedent and outcome. Contemporary Polish electoral behavior is, for Jasiewicz, a legacy of the nineteenth-century partitions. ³⁵ For Wittenberg, patterns of support for conservative parties in Hungary are a legacy of pre-communist partisanship. ³⁶ But an outcome qualifies as a legacy only if it cannot be fully explained except by reference to an antecedent cause or correlate. As a direct consequence of this temporal structure, all legacy arguments must feature at least an implicit division between a past period (when the antecedent cause or correlate are identified) and a later period (when the outcome occurs). Statements asserting legacies of the past in post-communist politics can thus always be stated in the form that some post-communist outcome (say, conservative voting behavior in Poland) is a legacy of some cause or correlate prior to post-communism (say, Habsburg policies toward the Roman Catholic Church prior to World War I). The label we give the legacy refers to the past period. We would say that post-communist conservative voting in Poland is a Habsburg or pre-war legacy. Although the lion’s share of work on historical legacies in post-communism features periods defined by different political regimes (e.g., Habsburg, Czarist, communist, Stalinist), nothing in the definition of a legacy requires that the periodization refer to regimes. We could just as plausibly identify a (prior) period of central planning and a (later) period of market allocation and investigate the legacies of central planning on later market practice.

Second, in cases where the antecedent is a cause (rather than a correlate), an outcome qualifies as a legacy only if that explanatory factor ceased to directly operate at some point before the outcome is observed. In the phrasing of Stinchcombe, the outcome must have “historical” causes, causes that produce the outcome at an earlier point in time but that do not reproduce that outcome over time. ³⁷ An example of such an explanatory factor is the religious policy of the Habsburg Empire, which Grosfeld and Zhuravskaya argue is indirectly related to post-communist support for conservative parties. ³⁸ Initially the policy directly increased support for religious parties in Habsburg Galicia. But with the passing of the Dual Monarchy there ceased to be a Habsburg policy, so there could have been no direct effect. Rather, support for conservative parties persisted even after the policy disappeared. Similarly, arguments about the contemporary lack of trust in post-communist political parties rely on Soviet-era factors such as state repression and domination of the public sphere that no longer exist (at least in much of Eastern Europe). ³⁹ Thus, although a legacy connotes continuity with the past, it cannot exist without a discontinuity of causal factors. A legacy can be characterized as an aftereffect of an antecedent cause that no longer operates. ⁴⁰

Legacies and Non-legacies

When is an outcome not a legacy? For arguments where the antecedent is a causal factor this is a troublesome question. Consider an outcome of interest. Suppose a researcher investigates a potential antecedent and finds no evidence that it is related to the outcome. That researcher is warranted in concluding that the outcome is not a legacy of the (potential) antecedent cause. For example, if the outcome were post-communist popular demand for high welfare expenditures and the potential antecedent were, say, the coming to power of Leonid Brezhnev in the Soviet Union, then the researcher would conclude that demand for welfare expenditures is not a legacy of Leonid Brezhnev’s coming to power. But it would not be correct to conclude that demand for welfare expenditures is not a historical legacy at all. The reason is that any conceivable outcome is a product of some prior causal factor, though that factor may well be incredibly difficult to identify. After all, what else could an outcome be a product of? There is of course a value in identifying which causal chain leads to an outcome. To continue with the example, it would be important to know that demand for welfare expenditures is rooted in popular experience with the state-socialist nanny state rather than, say, interwar Church–State relations. But this is a matter of distinguishing among different legacies rather than the existence of a legacy per se. In short, all outcomes are legacies in the context of antecedent causes. The only question that makes sense to ask is what the outcome is a legacy of.

Non-legacies make more sense when the antecedent and the outcome are instantiations of the same phenomenon measured at two different periods of time. Where the outcome is correlated with the antecedent then we can say the phenomenon persisted and that the outcome is at least a potential legacy (of itself at an earlier period of time). For example, Jowitt argues that the popular post-communist view of politics as something to avoid was inherited from the state-socialist period. ⁴¹ Where there is no correlation between outcome and antecedent, or where the outcome is new and thus has no phenomenologically equivalent antecedent, then there is discontinuity and a legacy can be excluded. An example of such a post-communist non-legacy would be free and fair elections, for which in parts of the former communist world there was no historical precedent. Unless otherwise noted, all legacies referred to in the remainder of this essay are of the persistence rather than the causal chain form.

Outcomes in post-communism may thus be divided into those that are potential legacies and those that are new, having never appeared in the past. But these two possibilities do not exhaust the possibilities for understanding post-communism in a historical perspective. Research on historical continuity to post-communism has given short shrift to phenomena that existed before post-communism but have not persisted into post-communism. ⁴² Examples would be the cult of personality or show trials, which existed under state-socialism but have failed to materialize in post-communism (at least in Eastern Europe). Such phenomena, which might have become historical legacies but did not, prove that the past, however large it looms, does not necessarily have to be a prologue. They are aborted legacies.

Pathways to Post-Communism

The temporal structure of continuities and discontinuities between post-communism and the past may be visualized in Table 1, which exhibits the relationship between the presence or absence of a phenomenon in different historical periods and the legacies and non-legacies of the past. The rightmost three columns in this table represent the conventional time periods from the post-communist legacies literature: pre-communism, communism, and post-communism. I employ this particular periodization and labeling for convenience only; the argument would work just as well with a different periodization as long as there are three periods. An “X” in a particular column means that a phenomenon of interest was present during that period. Each row represents a potential pathway to post-communism for a phenomenon of interest. Rows in which an “X” appears in both post-communism and at least one prior period are potential legacy pathways because those are instances in which something that existed before post-communism continued into post-communism. The table refers to potential legacies because although persistence of a phenomenon across at least two time periods that include post-communism is necessary for the post-communist outcome to be a legacy, it may not be sufficient. The post-communist phenomenon would still need to be inexplicable given post-communist circumstances.

OPEN IN VIEWER

Table 1

Pathways to Post-Communism

	Pre-Communism	Communism	Post-Communism
Uniquely Post-Communist			X
Potential Communist Legacy		X	X
Potential Pre-Communist LegacyA	X	X	X
Uniquely Communist		X
Pre-Post-Communist	X	X
Uniquely Pre-Communist	X
Potential Pre-Communist LegacyB	X		X

The top path (“Uniquely Post-Communist”), with an “X” only in the post-communism cell, represents features of a post-communist polity that are new in the sense that they had never appeared before the fall of communism. As noted above, for most countries in the region such novelties include free and fair elections. With the exception of interwar Czechoslovakia and perhaps the royal elections of the Polish-Lithuanian Commonwealth, no country in East Europe could boast of having had very many fully democratic elections before 1990. In many countries of the region, constitutionally guaranteed freedoms appeared for the first time after 1989.

The second path (“Potential Communist Legacy”) describes phenomena that came into existence during the communist period and also exist in the post-communist period, more commonly referred to as communist or Leninist legacies. For example, many of the steel factories built during communism would qualify because they were built during communism and continued to exist right into the post-communist period. We need not even qualify them as potential legacies. They exist in post-communism because they existed at the end of communism and were not dismantled. Post-communist circumstances did not (re)create them. Another example, mentioned above, is the excessive popular expectation of the willingness and ability of the state to provide for social welfare. Such expectations surely did not exist before the communist period, when states were too weak and too poor to provide the cradle-to-grave welfare that came to be seen under communism as a right rather than a privilege.

“Potential Pre-Communist LegacyA” in path three portrays features of the region that existed before communism, during communism, and after. One example of this is “backwardness.” As Janos notes, Eastern Europe as a whole has been economically marginal vis-à-vis Western Europe for centuries. ⁴³ Another example would be ethnic fragmentation, which began with the fall of the Habsburg and Ottoman empires at the end of World War I and has continued, if more attenuated, until the present day. Some of the phenomena researchers have identified as Leninist legacies are in fact pre-communist phenomena of this type. For example, attitudes inimical to liberalism, a “ghetto” political culture, and deference to authority, while undoubtedly features of communism, were also prominent in pre-communist East Europe. The same might be said of étatist developmental strategies.

The fourth path (“Uniquely Communist”) describes features of communism not present in either the pre- or post-communist phases. From the standpoint of post-communism, this path represents an aborted legacy, a phenomenon that might have become a legacy but failed to do so. Many such features can be identified depending on the country in question, including the fusion of Party and State, ⁴⁴ the soft budget constraint, ⁴⁵ and features of Stalinism such as the cult of personality and central planning.

The fifth path captures what is awkwardly termed “Pre-Post-Communist.” It illustrates a feature that is present until the collapse of communism, but is extirpated under post-communism. Rigged elections would fall into this category for the many post-communist countries that experienced liberal democracy for the first time only after 1989. National sovereignty would be another for those states that first gained independence after the collapse of the Soviet Union. From the standpoint of post-communism, this trajectory represents another aborted legacy. However, if we shift our reference period from post-communism to communism, then our interpretation changes. Specifically, for an observer in the communist period, the pre-communist phenomenon is a potential historical legacy. An example of this in most countries in Eastern Europe would be authoritarian rule, which certainly existed prior to communism and continued into the communist period. Some outcomes, such as peripheral status in the world economy, are potential pre-communist legacies for both the communist and post-communist periods.

The sixth path represents a “Uniquely Pre-Communist” pathway. In this category would be features of pre-communist systems that were wiped out under communist rule and have not been revived. Examples for Eastern Europe include the political power of the land-owning class and the economic influence of the Churches. The communists relegated, seemingly permanently, both the large landowners and the Churches to a political status far inferior to what they had enjoyed before the advent of communism.

The seventh path (“Potential Pre-Communist LegacyB ”) represents features of these polities that are present in both the pre-communist and post-communist periods but not during the communist period itself. Many such candidate legacies have been offered, usually as “revivals of the past” or “return of history.” One example might be the so-called “frozen conflicts” that raged before the communists came to power and then reemerged after the fall of communism. Another example could be the post-communist revival of political parties that competed in elections before the communists assumed power. Metaphysically speaking this is the most controversial path because it is not clear what it means for a phenomenon to disappear and then come back. How do we know that the phenomenon in the post-communist period is the same phenomenon as the one that existed before communism. What do we mean by “the same”? Addressing these questions is beyond the scope of this paper. ⁴⁶

The utility of Table 1 is that its categories are exhaustive. Its columns cover all periods of a former communist country’s history, and its rows cover all possible pathways to post-communism. ⁴⁷ As noted earlier, there is nothing sacred about the “pre-communist”–“communist”–“post-communist” categorization. Indeed, from a descriptive perspective, it is quite flawed. The pre-communist era is composed of multiple regimes, in some countries including fascist, traditional dictatorship, and monarchy. Moreover, with the entry of much of Eastern Europe into the EU and NATO and the even more recent return of Russia as a great power, we have arguably entered the “post-post-communist” era. The discussion has also ignored (and not even labeled) the transitional periods themselves, which we know can leave their own legacies on the subsequent regime. ⁴⁸ However, for purposes of understanding the temporal structure of legacies, nothing is gained by employing more than three periods. The three basic trajectory types are those in which an outcome appears in only one period (those paths prefaced with “Uniquely”), originates in one period and appears again in the successive period (“Potential Communist Legacy,” “Pre-Post-Communism,” and “Potential Pre-Communist LegacyA”), or originates in one period and reappears after an absence (“Potential Pre-Communist LegacyB ”). More complicated trajectories resulting from the addition of additional periods can be broken down into one of these three fundamental pathways.

Table 1 highlights other important features of legacies. First, we can see that scholarly focus on legacy pathways misses more than half of the trajectories leading from pre- to post-communism. Of the seven paths in Table 1, only three involve phenomena that exist in post-communism and were carried over from at least one prior period. Three involve phenomena that historical precedent suggests might have existed in post-communism but never appeared (the “Uniquely Pre-Communist,” “Pre-Post-Communist,” and “Uniquely Communist” pathways), and one pertains to phenomena new to post-communism. An interesting and relatively unexplored research question is why some outcomes become legacies whereas others do not.

Second, the number of time periods in which a phenomenon exists has a bearing on how the corresponding legacy should be temporally labeled. For phenomena that exist in only two time periods, there is no ambiguity about how to label a potential legacy. Suppose, for example, the historical record were divided into a post-1989 period and a pre-1989 period (that included the events of 1989 themselves). If there were historical legacies in post-1989 politics then those could (by construction) only be pre-1989 legacies. But if there are more than two periods when the phenomenon exists, then the potential legacy should be labeled according to the period in which the phenomenon is first identified. Consider once again the avoidance of political involvement (“ghetto political culture”), which Jowitt argued as being an important communist (“Leninist”) legacy in post-communist politics. ⁴⁹ As noted earlier, it was also an important phenomenon of pre-communist regimes, and should therefore be labeled a pre-communist rather than a communist legacy.

Third, historical legacies are not merely out there in the empirical world waiting to be discovered, but also conditional on how researchers choose to periodize history. This is trivially true in the sense that how we label a period will determine the labeling of the legacy. But it is also true in a less trivial way. Consider popular post-communist demands for welfare transfers, widely considered to be a legacy inherited from the communist period. If we were to reperiodize history in such as way as to merge the communist and post-communist periods into one longer period, then there would no longer be a legacy because the phenomenon would only exist within one period, analogous to the paths with only one “X” in Table 1. A similar logic works if instead of merging two periods into one, we break one period into two. For example, one might argue that the recent reemergence of right-wing populism in Eastern Europe is a legacy of hardships endured and compromises made on the road to qualifying for membership in the European Union. In effect, this argument is making an implicit distinction between different phases of the post-communist period, demarcated by the point at which EU entry became inevitable. Such a legacy is not visible in Table 1, where post-communism is not further subdivided.

Conclusion

This article illustrates that a great deal can be learned from an analytic rather than empirical approach to the study of historical legacies. Although for expository purposes the focus was on the former communist world, the arguments apply to legacy arguments more broadly. The big takeaway concerns how we conceptualize and identify legacies, about which there is scholarly disagreement. Historical legacies come in two flavors. One is as the endpoint of a causal chain that began at some point in the past. Understood this way, all outcomes are historical legacies because any outcome can be conceived as the end of a causal chain beginning in the past. Within this causal chain branch of research, the key question is thus not whether or not an outcome of interest is a legacy, but what kind of historical legacy an outcome is.

Another kind of legacy researchers identify is a phenomenon that persisted from the past. Understood this way, a phenomenon qualifies as a historical legacy under two conditions. One is that it has to exist in at least two historical periods, at a minimum a “past” and a “present.” Phenomena unique to one historical period cannot be historical legacies unless that period is itself divided into at least two subperiods. The other condition is that in the most recent of the periods in which the phenomenon exists, the phenomenon should not be explainable with contemporaneous causal factors. Further research is needed on how we know whether it is the same phenomenon that appeared in each time period, particularly in regard to legacies that represent so-called revivals of the past.

The article has also drawn attention to the importance of non-legacies. These are not merely phenomena that exist in only one historical period. They are also phenomena from earlier historical periods that did not persist into later periods. Why some outcomes become legacies and others do not remains a tantalizing question.