Coup with Adjectives: Conceptual Stretching or Innovation in Comparative Research?

https://doi.org/10.1177/0032321719888857 Political Studies 1 –22 © The Author(s) 2019 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/0032321719888857 journals.sagepub.com/home/psx Coup with Adjectives: Conceptual Stretching or Innovation in Comparative Research? Leiv Marsteintredet1 and Andrés Malamud2 Abstract Was Brazil’s President Dilma Rousseff victim of a coup or removed through a legal process of impeachment? The heated debate on the 2016 ousting of Brazil’s president testifies to the growing controversy around the definition of coups. Focusing on Latin America, we show that the use of coups with adjectives have become more frequent in public and scholarly debates. Occurring at a time when coups are becoming rarer, we argue that this development is linked to prevalenceinduced concept change, meaning that when instances of a concept become less prevalent, the understanding of the concept expands. The meaning of coups has expanded through a proliferation of adjectives. Coups with adjectives are not new, but recent usage changes the concept from a classic to a family resemblance structure. Although this strategy can avoid stretching and increase differentiation, we urge caution and warn against harmful consequences, whether conceptual, theoretical, or practical. Keywords coup d’état, impeachment, concept formation, conceptual stretching, Latin America Accepted: 25 October 2019 “In Europe, with its more limited experiences of such matters, the expression coup d’état is not specifically distinguished from say, a coup de force, and we use the expressions, ‘military revolt’, ‘mutiny’, ‘rebellion’, ‘coup’, ‘revolution’ interchangeably without asking what precisely has happened. Latin Americans, with their closer acquaintanceship with the phenomenon, distinguish.” S.E. Finer (1962: 154) 1University of Bergen, Bergen, Norway 2Institute of Social Sciences, University of Lisbon, Portugal Corresponding author: Leiv Marsteintredet, University of Bergen, Christies gate 15, Bergen 5020, Norway. Email: [email protected] 888857PSX0010.1177/0032321719888857Political StudiesMarsteintredet and Malamud research-article2019 Article 2 Political Studies 00(0) Introduction Coups d’état —however defined—have traditionally been regarded as the greatest threat to the survival of democracy. The sudden, often military-backed removal of the government customarily meant the end of a democratic regime. Today, however, despite a broad consensus that there is a global democratic recession and that liberal democratic values are threatened (Diamond, 2015a; Inglehart and Norris, 2016; Lührmann and Lindberg, 2019; Mechokva et al., 2017), there are fewer classical coups than ever, whether globally or in Latin America (Belkin and Schofer, 2005; Derpanopoulos et al., 2016; Djuve et al., 2019; McGowan, 2003; Marinov and Goemans, 2014; Powell and Thyne, 2011; Singh, 2014). In fact, consolidated democracies have grown almost immune to them (Svolik, 2015). Yet, we witness a puzzling increase in the use of the term (whether coup d’état or golpe de estado, as in Spanish and Portuguese), frequently combined with a qualifying adjective, in both academic and non-academic texts.1 In this article, we analyze why coup with adjectives are on the rise at a time that coups occur less frequently, and what the analytical and conceptual consequences of this development are. “Coups with adjectives” are not a new phenomenon. Terms like “military coup” or “self-coup” (autogolpe) have been in regular use within and outside academia for decades (Cameron, 1998; Fitch, 1977). However, their use has mushroomed in recent decades, particularly in Latin America. We found evidence of the recent invention or adaptation of qualifiers such as “soft” (Pitts et al., 2016), “parliamentary” (Santos and Guarneri, 2016), “presidential” or “democratic” (Varol, 2017), “constitutional” (Helmke, 2017; Yarwood, 2016), “market” or “neo-liberal” (Mauceri, 1995), “electoral” (Hellinger, 2005), “slowmotion” (Polga Hecimovich et al., 2017), “civil society” (Encarnación, 2002), and “judicial” (Yavuz and Koç, 2016). Many of these terms are piled together in Gentili’s (2016) volume on the impeachment of President Dilma Rousseff in Brazil. Although Bello (2016) and Keating (2012) register the use of coups with adjectives outside academia, the list above shows that their usage is not restricted to politicians or activists who have incentives to use real or invented coup-plots to implement emergency rule, arrest opposition leaders, clean out the bureaucracy, or brand the opposition as undemocratic. We link the rise of “coup with adjectives” to the phenomenon of “prevalence-induced concept change” (Levari et al., 2018): when instances of a concept become less prevalent, the understanding of the concept expands to cover cases it previously excluded. Following Collier and Levitsky (1997), we argue that while earlier uses of coup with adjectives went down the ladder of generality or abstraction (increasing differentiation), new adjectives make up diminished or dismissive subtypes in which some constitutive elements of the concept are missing or only partially present. The examples of contemporary coups with adjectives mentioned earlier have in common that they do not satisfy all the criteria for the commonly accepted definition of a coup: the illegal overthrow of the government by other state actors. As a consequence, whereas a coup was formerly understood as a classical concept, the new usage approaches it as a family resemblance concept (Collier and Mahon, 1993; Goertz, 2006). The choice of how to conceptualize a coup is not to be taken lightly since it carries normative, analytical, and political implications. In particular, in today’s democracies there is a conceptual peril of conflating a coup with legal tactics for government replacement. Identifying a phenomenon as an “impeachment” as opposed to a “coup” involves widely different interpretations, moral judgments, and consequences, as the latter could trigger international sanctions whereas the former should not. Therefore, academics as Marsteintredet and Malamud 3 well as politicians and pundits should proceed with care when choosing concepts to define as consequential an event as the removal of a president. The qualitative material for this article is based on a review of the academic literature on coups in general and in Latin America in particular, and on close reading of social media and news related to Latin American politics, our area of research. To substantiate our claim of an increased use of coups with adjectives, we use quantitative evidence from Google Books’ English and Spanish corpora for the period 1800–2008 with the Ngram Viewer (2013) tool.2 Furthermore, we mapped the academic use of coups using the Social Science Citation Index for English terms, and SciELO (Scientific Electronic Library Online) citation index for Spanish and Portuguese terms.3 We compare the use of the coup concept with coup data from Powell and Thyne (2011) and Przeworski et al. (2013). In the next section, we present empirical evidence to sustain our argument about the proliferation of the term accompanied by a variety of adjectives, and substantiate our puzzle: why does this occur during a period when actual coups in the world decline? In order to disentangle the puzzle, we then discuss the coup concept, its constitutive elements and structure, before we discuss strategies to avoid conceptual stretching, and show through examples that the current usage of coups with adjectives is shifting the concept structure from a classical to a family resemblance type. We argue that this shift can be understood as special form of prevalence-induced concept change. Finally, we discuss the implications and whether conceptual innovation has resulted in conceptual stretching or improved the utility of the term. Although we draw our examples mostly from Latin America, the discussion is expected to hold general validity. The Fall of Coups and Rise of “Coups With Adjectives” Globally, coup attempts have been declining since the latter half of the 1960s, and have been particularly rare after the Cold War. Based on data from Powell and Thyne (2011), Figure 1 displays the trend of coup attempts since 1950. The downward trend is uncontroversial and supported by several recent studies (Belkin and Schofer, 2003; Derpanopoulos et al., 2016; Djuve et al., 2019; Marinov and Goemans, 2014; Marshall and Marshall, 2018; Singh, 2014). In Latin America, with some exceptions such as in Honduras in 2009, coups have almost vanished. At the same time, the use of the term coup has seen a puzzling revival with recent developments in Latin America. Reading the news, registering conversations on social media, and searching the literature, we observed that academics, politicians, and the media increasingly use the Spanish term golpe connected with an adjective to describe events that would not fall under the classical conceptualization. A non-exhaustive list of the most prominent and prevalent new adjectives includes “soft,” “parliamentary,” “constitutional,” “neo-liberal,” “market,” “electoral,” “slow,” “civil society,” and “judicial.” While often these terms are used by presidents to taint their opponents, they are also used in academic works, and thus should be analyzed rigorously. First, however, we need to establish whether our hunch about an increased use of coups with adjectives holds when confronted with evidence. We do this by studying the prevalence of the use of the term coup d’état and coup with adjectives in the English and Spanish corpora in Google Books using the Ngram Viewer tool. The results are displayed in Figures 2 and 3. The use of the term “coup d’état” in English follows a pattern similar to that of the registered coup attempts (Figure 2), a steady increase until the 1960s, and after that a sharp decline until the final year registered (2008). The use of the term in books is highly 4 Political Studies 00(0) Figure 2. The Use of Coup d’état and Golpe de estado in Books. The bar and line present the share of each bigram (coup d’état and golpe de estado) of all bigrams in the database per year. Results come from case insensitive search of “coup d’état” in the English corpus (googlebooks-eng-all-20120701), and from case insensitive search of “golpe de estado” in the Spanish corpus (googlebooks-spa-all-20120701). Years covered 1800–2008. Source: Ngram Viewer (2013). Figure 1. Coup Attempts Worldwide and in Latin America. Source: Powell and Thyne (2011). Marsteintredet and Malamud 5 correlated with actual coup attempts (Table 1). Notably, a similar decline is not present in the use of “golpe de estado,” which increases steadily from the 1950s and, with exception of a dip in the years prior to 2000, ends with the highest registered score in 2008. Although the Google books database covers academic literature, it also includes fiction and thus tells us little of the popularity and content of academic studies of coups. Looking only at academic articles in the Social Science Citation Index and SciELO (Figure 3), we find signs that studies of coups have become more frequent (solid and dotted lines), especially since the early 2000s. As academic production has increased overall, we control by Figure 3. The Use of Coup d’état and Golpe de estado in Academic Articles. Unit of observation is years. Results from search of “coup d’état” as topic in Social Science Citation Index (N-258), and “golpe de estado” as topic in the SciELO citation index (N-187). Both available through Clarivate Analytics’ Web of Science (v. 5.30). Share calculated as number of articles per year with “coup d’état” as topic divided by number of articles per year with “political science” as topic (N-3592). The Social Science Citation Index goes back to 1956. SciELO has records from 1997. Table 1. Correlations actual coups and attempts and mentions in Google Books. Coups and Coup attempts since 1949 (Powell and Thyne, 2011) Coups since 1949 (Przeworski etal., 2013) Coups (Przeworski etal., 2013) Coup d’état .69 (.00) .61 (.00) .47 (.00) Coup with adjectives .01 (.92) –.08 (.53) .37 (.00) Golpe de estado –.19 (.14) –.12 (.34) .35 (.00) Golpe con adjetivos –.41 (.00) –.39 (.00) .13 (.05) Cells show correlation coefficients, significance level in parenthesis. “Coup d’état,” “Coup with adjectives,” “Golpe de Estado,” and “Golpe con adjetivos” from Google Books, observations with 1 year lag. 6 Political Studies 00(0) mapping English language articles with “coup d’état” as topic as share of articles with “political science” as topic (dashed line).4 The results show that the study of coups has also increased from the early 2000s as a share of all articles in the discipline, even though it peaked in the 1990s. Recent popularity may be connected to the growth of available datasets on coups (cited earlier), and the recent surge of studies of authoritarian regimes and leadership succession. Our main interest, however, is with coups with adjectives. We searched Google Books for the above-mentioned coups with adjectives in English and Spanish, and found the following: in English there is no decline of their use as the number of coups declined, and in Spanish their use has expanded tremendously at the same time as the number of coups declined (Figure 4). Figure 4 displays the English and Spanish usage of coups with adjectives. Comparing Figures 2 and 4, we see that while the use of coup d’état has decreased with the numbers of coups, there is no dip in the use of coups with adjectives.5 Furthermore, while the mention of coup d’état is positively and significantly correlated with actual coup attempts, the use of coup with adjectives is not after 1949 (Table 1). In Spanish, however, coups with adjectives were an extremely rare phenomenon until the 1960, but have consistently been mentioned in books since the 1970s confirming that it is indeed a more recent phenomenon. Figure 5 visualizes the association between the usage of coup d’état, coups with adjectives in English and Spanish, and actual coups since the early 1800s. With exception Figure 4. Coup With Adjectives in English and Spanish in Books. Unit of observation is years. The lines present the share of the sum of bigrams (sum of coup with adjectives and sum of Golpe con adjetivos) of all bigrams in the database per year. Results from case insensitive searches of “electoral coup,” “judicial coup,” “slow-motion coup,” “soft coup,” “parliamentary coup,” and “constitutional coup.” Results for each of the coups with adjectives are added together into the variable “Coup with Adjectives.” In Spanish, results from case insensitive searches of “golpe electoral,” “golpe blando,” “golpe constitucional,” and “golpe institucional.” No hits on “golpe parlamentario.” Marsteintredet and Malamud 7 for the usage of the root term in English, the correlation between actual coups and usage of the term with or without adjectives seem to break down exactly when the number of coups decline (see also Table 1). Table 1 shows correlations between mentions of coups with or without adjectives in English and Spanish in Google Books and actual coups and coup attempts as measured by Powell and Thyne (2011) and Przeworski et al. (2013). When looking at the full period from 1804, all variants of coups in Google Books are positively correlated with actual coups, and mentions of “coup d’état” is strongly and significantly so. Since 1949, when Powell and Thyne’s data start, however, we observe that coups with adjectives in Spanish show a negative and significant correlation with coups and attempts, while for coups with adjectives in English or coups in Spanish, the correlation is either negative or zero and non-significant. Table 1 also confirms that the decoupling or negative correlation between coups with adjectives in English and Spanish and real coups and attempts, are recent phenomena. In sum, we find support for our initial observation that there is an increase in the use of coups with adjectives, especially in Spanish, and that several of the adjectives are of new date. What accounts for the increase in the use of coups with adjectives at a time when coups are declining? First, for Latin America we believe it is related to two key developments. Second, and most importantly, we argue that it is related to a change in the concept structure of the term coup, which constitutes a special type of prevalence-induced concept change (Levari et al., 2018). Two developments in Latin America help explain the rise of coups with adjectives in particular in the Spanish language. One is the new political instability, which refers to the many cases of early presidential termination by impeachment (Pérez-Liñán, 2007), or Figure 5. Actual Coups vs Mentions of Coups with or without Adjectives in Books. Unit of observation is years. Variables based on Google Books are standardized with a mean of 0 and standard deviation of 1. Lowess smoother used with bandwidth of 0.5. 8 Political Studies 00(0) through resignation after massive protests (Hochstetler, 2006). Not only was the use of impeachment, or impeachment-like procedures, new to the region, it was also new that democracy tended to survive the fall of presidents. Second, the revival seems linked, but is not exclusive, to the wave of leftist regimes in the region and the fall of presidents of the left. While the phenomenon Pérez-Liñán and Hochstetler analyzed mainly occurred with neoliberal presidents, a recent development is that left-leaning presidents have begun to fall (Presidents Rousseff in Brazil 2016, Lugo in Paraguay 2012, and Zelaya in Honduras 2009) or to be questioned for their democratic merits (Presidents Maduro in Venezuela and Ortega in Nicaragua). In a region where United States meddling and coups directed against the left were trademark, the fall of presidents of the left has revived bad memories.6 Yet, this explanation is incomplete, unless we can explain how events formerly not recognized as coups are now labeled as coups with adjectives. Our key hypothesis is linked to prevalence-induced concept change, or concept creeping, which is a theory from psychology that holds that “when instances of a concept becomes less prevalent, the concept may expand to include instances it previously excluded” (Levari et al., 2018: 1465). Over a series of experiments, Levari et al. showed how the reduced prevalence of a phenomenon made the understanding of the phenomenon expand so that the participants over-reported its occurrence. This explains why even though modern societies have made progress in fighting poverty, hunger, and infectious diseases, people still believe the world is going in the wrong direction (Levari et al., 2018: 1465). We contend that the same takes place with coups; as they become less frequent and less dangerous to democracy, observers expand their understanding of what a coup is. To be sure, our findings are not strictly consistent with this theory, since it is the use of coups with adjectives, and not the use of the term coup alone, which has increased. This finding is even more puzzling since adding adjectives to a classical concept, such as a coup, reduces its extension (coverage) by increasing its intension (meaning) (Collier and Mahon, 1993; Sartori, 1970). To make sense of why the use of coups with adjectives increases in a period when actual coups are at historical lows, we argue in the remainder that the prevalence-induced concept change has not altered the constitutive elements of a coup, or what a coup is, but rather changed the concept structure of coups from classical to a family resemblance type, thereby reducing the threshold for an instance to qualify as a coup. What is a Coup? A coup is not an essentially contested concept like democracy or populism (Gallie, 1956; Mudde, 2004). Even though disagreement exists, there is a relatively broad academic consensus on what constitutes a coup. The term, however, is much used outside academia, and controversial because it is politically loaded and holds normative and practical implications. The United States, for instance, is bound to suspend foreign aid if a recipient experiences a coup, and the European Union upholds similar clauses (Marinov and Goemans, 2014: 805). In Latin America, the Inter-American Democratic Charter of the Organization of American States calls for immediate action in case of a coup (“unconstitutional interruption”), and foresees the potential suspension of the member state in question (see, OAS, 2001: arts. 19–21). In addition, a distressed president has incentives to brand his belligerent opposition as coup-mongers, while the opposition would like to cloak themselves in innocence as fighters for freedom and democracy against an authoritarian-minded president. Likewise, coup-makers often attempt to legitimize their own Marsteintredet and Malamud 9 coups by calling their actions “revolutions,” and themselves a “movement.” Just as “one person’s terrorist is another person’s freedom fighter” (Wardlaw, 1989), one person’s coup-maker can be another person’s democratizer. In the seventeenth century, a coup was the exceptional and illegal act of a prince against other parts of society in order to deal with an extraordinary situation. Its goal was to restore order, and the prince’s illegal act was considered legitimate since its aim was to benefit the general interests (Bartelson, 1997; Martínez, 2014).7 With the end of absolutism and the establishment of the principle of popular sovereignty, which separated the ruler(s) from the State, the coup concept changed into how we know it today (Bartelson, 1997). One classic definition comes from Luttwak who defined a coup as “the infiltration of a small but critical segment of the state’s apparatus which is used to displace the government from the control of the remainder” (Luttwak, 1969: 12). Coup-makers use the state machinery to take control over the political levers. According to Luttwak, a coup shares characteristics with revolution, putsch, rebellion, pronunciamiento, civil war, and national liberation. Occupied with differentiating the coup from these phenomena, Luttwak focused on the state turned against the state, and distinguished a coup from other take-over strategies in its being performed by a small group. Another classic definition that seeks to distinguish a coup from a barrack revolt (or cuartelazo) is Finer’s (1962: 154),8 “the seizure and elimination of the person of the head of state.” Finer (1962: 155) also pointed out that a coup is “speedy and immediate,” but did not include these characteristics as part of the definition. In fact, the term coup in French or golpe in Spanish means a quick movement, a stroke, a blow, or a sudden attack. We stick, however, to Finer’s parsimonious strategy, and exclude speediness and other common traits such as secrecy and small group size of coup-makers from our definition. Had we added these criteria the distinction between the old and new conceptualizations of coups would become even starker. Based on the triple criteria of target, perpetrator, and tactics, Powell and Thyne summarized the literature and arrived at the following precise definition: “[a coup attempt includes] illegal and overt attempts by the military or other elites within the state apparatus to unseat the sitting executive” (Powell and Thyne, 2011: 252). Neglected in classic definitions, but emphasized by Powell and Thyne and McGowan (2003), is that the tactics used must be illegal. Powell and Thyne (2011: 251) write, “this is important because [it] distinguishes coups from political pressure, which is common whenever people have freedom to organize.” This illegal act also normally involves a suspension of the constitutional order, whether democratic or autocratic, thus connecting this criterion to the original root of the concept in the seventeenth century. The earlier omission of illegality was due to three factors which are as follows: (1) the definitions aimed to differentiate between illegal power grabs—thus the criterion was superfluous; (2) the authors were occupied with military coups—again, the criterion was superfluous; (3) many coups occurred in regimes that failed to meet the criteria for rule of law and democracy, which made the legality/illegality distinction less important. The legal-illegal distinction is especially relevant in a democratic context, and therefore in today’s Latin America from where we draw many of our examples. In contrast to previous periods, most countries in Latin America since the Third Wave are democracies with at least a veneer of rule of law. Therefore, it is imperative to distinguish coups from legal removals of the heads of state such as impeachments or popular recalls, which have become increasingly common (Helmke, 2017). Before the Third Wave, the pitfall of coup definitions was conflating coups with other illegal attacks on the government, for example, 16 Political Studies 00(0) turned against the government, but unless the parliament illegally removed the president (by for instance not having the required number of votes to depose the president), this term does not satisfy the tactic of a coup since it was deemed legal by the relevant domestic and international authorities. Again, cases are the impeachments of Presidents Rousseff and Lugo.12 The constitutional coup describes the use of constitutional measures to remove a president, or prolong his stay in power. The latter case is used particularly in the African context (Camara, 2016; Kotze, 2017),13 and is similar to the judicial coup mentioned earlier. In Latin America, the term has mostly been used by political actors and the press, but appears frequently in Spanish in Google Books since the 1980s. The term constitutional coup does not indicate an actor, and it admits to the (controversial) constitutionality of the action; hence, the term only satisfies the criterion that the chief executive is the target, and where it is used to describe a president’s expansion of term limits, the term actually does not satisfy any of the criteria for a coup even though the actions are identified by that term. Adding Adjectives, Subtracting Meaning: Discussing Coup With Adjectives The cases presented earlier lack either one or more of what have been considered the necessary and jointly sufficient attributes of the root concept. Therefore, they constitute diminished subtypes. The increased use of coups with adjectives has changed the concept structure of coups from a classical to a family resemblance category and driven its prevalence upward in a time when coups are diminishing globally, consistent with prevalenceinduced concept change. In some cases, the terms describe actions by masses instead of a small group (e.g. electoral coup), slow instead of quick processes (e.g. slow-motion coup), or actions that are not shredded in secrecy (e.g. constitutional or parliamentary coup). Therefore, several of these examples also fail to satisfy logical but omitted criteria from modern definitions (secrecy, speediness, and small group size). Diminished subtypes may increase differentiation and avoid stretching (Collier and Levitsky, 1997). The root concept is left intact, but it is applied with qualifications to cases that are not considered “full and complete.” This strategy allows for flexibility, so that new forms and shapes of mandate interruptions can be captured by a well-known concept. For example, the coup concept itself has changed considerably since Naudé’s definition (Martínez, 2014), and the Third Wave democratization and the new political instability in Latin America contributed to conceptual changes that reflected empirical developments, for instance by insisting that coups should consist of illegal actions (Powell and Thyne, 2011). For terms such as democracy (Collier and Levitsky, 1997) or populism (Weyland, 2001), changing the concept structure from classical to family resemblance has been a successful strategy to capture empirical developments. And although Sartori (1984) was skeptical toward family resemblance concepts, these are now well established in the social sciences and their use does not inhibit positivist, causal analysis (Goertz, 2006). Nevertheless, we identify three problems with the use of diminished subtypes and the coup concept with a family resemblance structure. The first is that it inhibits the recognition of political evolution. Second, it creates the danger of causal and conceptual confusion. Third, getting a coup wrong may entail serious political consequences. Marsteintredet and Malamud 17 First, contrary to other concepts such as populism—which expanded to account for the increasing presence of populists (Rooduijn, 2018)—we link the changing structure of the coup concept to its decreasing empirical presence. The relative scarcity of coups, we argue, has led to prevalence-induced concept change. By adding adjectives, observers have increased the extension while reducing the intension of the root concept, therefore overreporting coups at the same time as their occurrence ebbed. For Latin America in particular, changing the concept structure inhibits observing the real progress made in reducing coups as threats to democracy. Even though scholars and the media mention coups (with adjectives) more often than before, the phenomenon is minimal in the region. Second, applying a family resemblance structure has created causal and conceptual confusion. By conflating coups with other concepts, the plethora of event that coups with adjectives refer to are also assumed to share the same or similar causes and consequences. Figure 6 shows that for every subtype of coup with adjectives under the family resemblance structure, there exists established concepts such as revolution and impeachment. The strategy of applying diminished subtypes allows for differentiating the key concept while traveling, but it also creates confusion between, for example, a civil society coup and a revolution, and between a parliamentary coup and an impeachment. It further creates pseudo-equivalences between phenomena as different as an impeachment and a revolution by placing both under the conceptual umbrella of a coup, which inhibits rather than improves our understanding and leads to faulty conclusions. We argue therefore that applying a family resemblance structure and diminished subtypes obfuscate rather than clarify the coup concept. Even though adding adjectives can enhance the precision in order to capture new developments, such as the new political instability or new ways of democratic deterioration, one problem remains, the recent usage of diminished subtypes has not led to an agreement over a more precise concept. The new subtypes point toward different processes which cannot be captured by a single new criterion and several adjectives such as “slow” or “constitutional” contradict the original meaning, constituting dismissive subtypes. In some versions, state actors are not involved, in other the actions are legal, in some the president is the target, in other all institutions but the presidency are the target. Unlike similar developments with the concept of democracy, which generated a more precise definition, coups with adjectives that constitute diminished subtypes obfuscate rather than precise the root concept. The conflation of coups with adjectives with other established concepts and phenomena lead to causal confusion and flawed analyses. Scholars explaining democratic backsliding and recession often highlight the differences between today’s dangers to democracy and classical coups. Causal processes leading to democratic backsliding are slow and incremental rather than quick, led from the government rather than against the government, and weakening institutions rather than aborting them (see for example, Bermeo, 2016). As Svolik (2015) points out, democracies grew immune to coups, but not to incumbent takeovers. Conflating these processes is likely to mislead causal analysis and policy prescription since when the diagnosis change, so should the therapy. Finally, classifying a change in government as a coup may hold real-life consequences. During the Cold War, the military in Argentina, Brazil, and Paraguay classified their coups as revolutions and their organizations as movements to legitimize their actions. Today scholars, observers, and presidents on the losing end of an impeachment classify such events as coups (with adjectives) to delegitimize their opponents. A crisper definition is in order since labeling an event as a coup may generate political actions of grave consequences such as the withholding of aid, the suspension from international organizations, 18 Political Studies 00(0) the triggering of international sanctions, and even foreign military intervention. To reduce the dangers of wrongful categorization of coups, and to distinguish politically motivated accusations or whitewashing, a clear definition with a set of jointly necessary and sufficient conditions is more useful than a family resemblance structure. Conclusion Whereas the threat of coups is lower than ever before, the dangers of concept misformation of coups is rising. Even though we can identify several recent interesting and positive conceptual developments in the literature that distinguish between different types of coups (Aksoy et al., 2015) or between coups and mutinies (Dwyer, 2015; Johnson, 2018), our main concern has been with the proliferation of diminished subtypes of coups. In today’s democracies, we particularly warn against conflating coups with legal tactics for government replacement. We have shown how the use of coups with adjectives, in particular in the Spanish language, has increased considerably in both the scholarly and nonscholarly literature while the prevalence of coups has decreased both globally and in Latin America. We argue that the reason for this puzzling development lies in prevalenceinduced concept change, which has come in the form of changing the concept structure of coups from classic to family resemblance. Originally suggested as a way of preventing conceptual stretching and creating analytical differentiation, we showed that when applied to coups, this strategy prevents concept stretching only partially and at a high cost. The risk is that proliferating adjectives end up diluting the noun: a coup becomes tantamount to just any mandate interruption and, as such, is virtually meaningless without adjectives. This further creates conceptual and causal confusion, and inhibits observers from identifying real world improvements. Politically, the spreading of coups with adjectives can also have damaging consequences. Classifying as coups phenomena that belong to another class, such as impeachments, raises a set of normative and, more consequentially, policy-related questions, as powerful actors could set measures in motion according to the label that best fits their interests. We conclude with a humble reminder for scholars and a warning for practitioners. The reminder, however obvious, is that not every removal is a coup, as a Latin American political activist ironically quipped, “golpes eran los de antes” (“real coups were those of yesteryears”). The warning should alert political leaders and activists that stretching the concept by accusing today’s adversaries of coup-mongering may hit them back tomorrow. Authors’ Note A previous version of this article was presented at the 9th Congress of Asociación Latinoamericana de Ciencia Política (ALACIP), Montevideo, Uruguay, 26–29 July 2017, and at the American Political Science Association Annual Meeting, Washington DC, 29 August–1 September 2019. Acknowledgements The authors thank Carlos Gervasoni, Mariana Llanos, Miguel De Luca, Michael Alvarez, and two anonymous reviewers for comments, and Octavio Amorim Neto and Michaël Tatham for conceptual and technical input. Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Marsteintredet and Malamud 19 Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Andrés Malamud acknowledges the financial support of FCT project UID/SOC/50013/2013. ORCID iD Leiv Marsteintredet https://orcid.org/0000-0002-5811-2641 Notes 1. Unless specified, we use “coup” as shorthand for coup d’état, and golpe as the Spanish shorthand for golpe de estado. This is not to say that coups in other institutions or organizations do not exist, but here we only deal with the coup d’état. 2. We use the English 2012 (googlebooks-eng-all-20120701) and the Spanish 2012 corpora (googlebooksspa-all-20120701), which cover titles from 1800 to 2008. See https://books.google.com/ngrams/info 3. Both available through Clarivate Analytics’ Web of Science (v. 5.30). 4. For Spanish and Portuguese we only have data from 1997. The first observation is from 2002 and we do not control for a possible increase in production articles over time. 5. The Google books database ends in 2008, and we believe, but are unable to confirm, that the use of coups with adjectives has increased since then. 6. See interview with political scientist Fabiano Santos likening the impeachment of President Rousseff with the coup in 1964 (“Impeachment revela que a base da democracia brasileira ainda é frágil,” El País, edición Brasil, 29 March 2016). 7. The first reference to coup d’état comes from Gabriel Naudé’s “Considerations Politiques sur les Coups d’éstat” in 1667, who defined a coup as “those bold and extraordinary acts that princes are forced to undertake in difficult and hopeless matters, contrary to common law and regardless or any justice, putting particular interest at stake for the benefit of the general one.” Quote translated and cited in Bartelson (1997: 324). 8. A cuartelazo is a barrack revolt involving a larger military group, and often more violence. 9. This distinction still holds relevance in the study of military mutinies and how they are distinct from (but may evolve into) coups (Dwyer, 2015; Johnson, 2018). 10. Whether coups may foster democracy, and whether coups after the Cold war more often end in democracy, are important issues in recent research (Derpanopoulos et al., 2016; Marinov and Goemans, 2014; Thyne and Powell, 2016; Varol, 2017). 11. The Peruvian case is so important for the term that the Spanish version autogolpe has been accepted in English-written academic works. The term appears in the Oxford English Dictionary since 2011. Although its popularity rose with Fujimori, autogolpe first appeared in Melville and Melville (1971). 12. See Santos and Guarneri (2016) for Brazil, and Marsteintredet et al., (2013) for references to the Paraguayan case. 13. The term soft-coup has also been used to describe this phenomenon in Africa (Van Woudenberg and Sawyer, 2015). References Aksoy D, Carter DB and Wright J (2015) Terrorism and the Fate of Dictators. World Politics 67: 423–468. Bartelson J (1997) Making Exceptions: Some Remarks on the Concept of Coup d’état and Its History. Political Theory 25: 323–346. Belkin A and Schofer E (2003) Toward a Structural Understanding of Coup Risk. Journal of Conflict Resolution 47: 594–620. Belkin A and Schofer E (2005) Coup Risk, Counterbalancing, and International Conflict. Security Studies 14: 140–177. Bello (2016) When a “coup” Is Not a Coup. The Economist, 9 April. Available at: https://www.economist.com/ the-americas/2016/04/09/when-a-coup-is-not-a-coup Bermeo N (2016) On Democratic Backsliding. Journal of Democracy 27: 5–18. Camara K (2016) Here’s How African Leaders Stage “Constitutional Coups”: They Tweak the Constitution to Stay in Power. The Washington Post (Monkey Cage). Available at: https://www.washingtonpost.com/ news/monkey-cage/wp/2016/09/16/heres-how-african-leaders-stage-constitutional-coups-they-tweakthe-constitution-to-stay-in-power/ 20 Political Studies 00(0) Cameron MA (1998) Self-Coups: Peru, Guatemala, and Russia. Journal of Democracy 9: 125–139. Collier D and Levitsky S (1997) Democracy with Adjectives Conceptual Innovation in Comparative Research. World Politics 49: 430–451. Collier D and Mahon JE (1993) Conceptual Stretching Revisited Adapting Categories in Comparative-Analysis. American Political Science Review 87: 845–855. David SR (1987) Third World Coups d’état and International Security. Baltimore, MD; London: The Johns Hopkins University Press. Derpanopoulos G, Frantz E, Geddes B, et al. (2016) Are Coups Good for Democracy? Research & Politics 3: 1–7. Diamond L (2015a) Facing Up to the Democratic Recession. Journal of Democracy 26: 414–155. Diamond L The Use of Coup d’état and Golpe de estado in Books. (2015b) In Search of Democracy. London: Routledge. Djuve VL, Knutsen CH and Wig T (2019) Patterns of Regime Breakdown Since the French Revolution. Comparative Political Studies. Epub ahead of print. DOI: 10.1177/0010414019879953. Dwyer M (2015) Tactical Communication: Mutiny as a Dialogue in West and Central Africa. Africa Spectrum 50: 5–23. Encarnación OG (2002) Venezuela’s “Civil Society Coup”. World Policy Journal 19: 38–48. Finer SE (1962) The Man on Horseback: The Role of the Military in Politics. London: Pall Mall Press. Fitch JS (1977) The Military Coup d’etat as a Political Process: Ecuador, 1948–1966. Baltimore, MD; London: The Johns Hopkins University Press. Gallie WB (1956) Essentially Contested Concepts. Proceedings of the Aristotelian Society 56: 167–198. Gentili P (2016) Golpe en Brasil.Genealogía de una farsa. Buenos Aires: CLACSO, Octubre Editorial. Goertz G (2006) Social Science Concepts: A User’s Guide. Princeton, NJ: Princeton University Press. Hamilton LC and Hamildon JD (1983) Dynamics of Terrorism. International Studies Quarterly 1: 39–54. Hellinger D (2005) When “No” Means “Yes to Revolution”. Electoral Politics in Bolivarian Venezuela. Latin American Perspectives 32: 8–32. Helmke G (2017) Institutions on the Edge: The Origins and Consequences of Inter-Branch Crises in Latin America. Cambridge: Cambridge University Press. Hochstetler K (2006) Rethinking Presidentialism: Challenges and Presidential Falls in South America. Comparative Politics 38: 401–418. Inglehart RF and Norris P (2016) Trump, Brexit, and the Rise of Populism: Economic Have-Nots and Cultural Backlash. HKS Working Paper no. RWP16-026 (HKS Faculty Research Working Paper Series), 29 July. Cambridge, MA: Harvard Kennedy School/John F. Kennedy School of Government. Johnson JM (2018) Things Fall Apart: The Determinants of Military Mutinies. Doctoral Dissertation —Political Science, University of Kentucky, Lexington, KY. Keating JE (2012) Coups Ain’t What They Used to Be. Foreign Policy. Available at: https://foreignpolicy. com/2012/06/27/coups-aint-what-they-used-to-be/ Kitson D (1982) Jamaica and the Electoral Coup of 1980. Race & Class 24: 169–178. Kotze JS (2017) Africa Faces a New Threat to Democracy: The “Constitutional Coup.” The Conversation. Available at: http://theconversation.com/africa-faces-a-new-threat-to-democracy-the-constitutional-coup-72011 Levari DE, Gilbert DT, Wilson TD, et al. (2018) Prevalence-Induced Concept Change in Human Judgment. Science 360: 1465–1467. Lührmann A and Lindberg SI (2019) A Third Wave of Autocratization Is Here: What Is New about It? Democratization 26: 1095–1113. Lunde TK (1991) Modernization and Political Instability: Coups d’etát in Africa, 1955–85. Acta Sociologica 34: 13–32. Luttwak EN (1969) Coup d’état: A Practical Handbook. Cambridge, MA: Harvard University Press. McGowan PJ (2003) African Military Coups d’etát, 1956–2001: Frequency, Trends and Distribution. The Journal of Modern African Studies 41: 339–370. Marinov N and Goemans H (2014) Coups and Democracy. British Journal of Political Science 44: 799–825. Marshall MG and Marshall DR (2018) Coup d’état Events, 1946–2017. Center for Systemic Peace, May 2. Available at: http://www.systemicpeace.org/inscr/CSPCoupsCodebook2017.pdf Marsteintredet L, Llanos M and Nolte D (2013) The Politics of Impeachment: The Fall of President Lugo in Comparative Perspective. Journal of Democracy 24: 110–123. Martínez R (2014) Subtipos de golpes de Estado: transformaciones recientes de un concepto del siglo XVII. Revista CIDOB d’afers Internacionals 108: 191–212. Mauceri P (1995) State Reform, Coalitions, and the Neoliberal Autogolpe in Peru. Latin American Research Review 30: 7–37. Marsteintredet and Malamud 21 Mechokva V, Lührmann A and Lindberg SI (2017) How Much Democratic Backsliding? Journal of Democracy 28: 162–169. Melville M and Melville T (1971) Guatemala: The Politics of Land Ownership. New York: Free Press. Mudde C (2004) The Populist Zeitgeist. Government and Opposition 39: 542–563. Ngram Viewer (2013) Google Books Ngram Viewer. (ed Google Research). Available at: http://books.google. com/ngrams OAS (2001) Inter-American Democratic Charter. Lima, Peru: OAS. Pérez-Liñán A (2007) Presidential Impeachment and the New Political Instability in Latin America. Cambridge: Cambridge University Press. Pitts B, Joyce R, Sheptak R, et al. (2016) 21st Century Golpismo: A NACLA Roundtable. NACLA Report on the Americas 48: 334–345. Polga Hecimovich J, Sagarzazu I and Sánchez Urribarri R (2017) Venezuelan Overview. LASA Forum XLVIII: 35–37. Powell JM and Thyne CL (2011) Global Instances of Coups from 1950 to 2010: A New Dataset. Journal of Peace Research 48: 249–259. Przeworski A, Asadurian T, Bohlken AT, et al. (2013) Political Institutions and Political Events (PIPE) Data Set. New York: Department of Politics, New York University. Roberts K (1995) Neoliberalism and the Transformation of Populism in Latin America: The Peruvian Case. World Politics 48: 82–116. Rooduijn M (2018) State of the Field: How to Study Populism and Adjacent Topics? A Plea for Both More and Less Focus. European Journal of Political Research 58: 362–372. Santos F and Guarneri F (2016) From Protest to Parliamentary Coup: An Overview of Brazil’s Recent History. Journal of Latin American Cultural Studies 25: 485–494. Sartori G (1970) Concept Misformation in Comparative Politics. American Political Science Review 64: 1033– 1053. Sartori G (1984) Social Science Concepts: A Systematic Analysis. Beverly Hills, CA: SAGE Publishing. Singh N (2014) Seizing Power: The Strategic Logic of Military Coups. Baltimore, MD: Johns Hopkins University Press. Smith WC (1990) Democracy, Distributional Conflicts and Macroeconomic Policymaking in Argentina, 1983– 1989. Journal of Interamerican Studies and World Affairs 32: 1–42. Stein E (2012) Revolution or Coup? Egypt’s Fraught Transition: Survival. Global Politics and Strategy 54: 45–66. Stone Sweet A (2007) The Juridical Coup d’état and the Problem of Authority. German Law Journal 2: 915–928. Svolik MW (2015) Which Democracies Will Last? Coups, Incumbent Takeovers, and the Dynamic of Democratic Consolidation. British Journal of Political Science 45: 715–738. Thyne CL (2010) Supporter of Stability or Agent of Agitation? The Effect of United States Foreign Policy on Coups in Latin America, 1960–1999. Journal of Peace Research 47: 449–461. Thyne CL and Powell JM (2016) Coup d’état or Coup d’Autocracy? How Coups Impact Democratization, 1950–2008. Foreign Policy Analysis 12: 192–213. Tovias A (1984) The International Context of Democratic Transition. West European Politics 7: 158–171. Van Woudenberg A and Sawyer I (2015) Africa’s Softer, Gentler Coups d’etat. Foreign Policy. Available at: https://foreignpolicy.com/2015/11/03/africas-softer-gentler-coups-detat/ Varol O (2017) The Democratic Coup d’état. Oxford; New York: Oxford University Press. Wardlaw G (1989) Political Terrorism: Theory, Tactics, and Counter-Measures. Cambridge: Cambridge University Press. Weyland K (2001) Clarifying a Contested Concept: Populism in the Study of Latin American Politics. Comparative Politics 34: 1–22. Yarwood J (2016) The Struggle Over Term Limits in Africa. The Power of Protest. Journal of Democracy 27: 51–60. Yavuz MH and Koç R (2016) The Turkish Coup Attempt: The Gülen Movement Vs. the State. Middle East Policy XXIII: 136–148. Author Biographies Leiv Marsteintredet holds a PhD in comparative politics and is an Associate Professor at the Department of Comparative Politics, University of Bergen. His research focuses on Latin America, in particular topics related to political institutions, presidential crises, impeachments, and falls. Currently, he is studying the development 22 Political Studies 00(0) and history of succession rules and the vice presidency in Latin America. He has published in journals such as Comparative Politics, International Political Science Review, Journal of Latin American Studies, European Political Science, and Journal of Democracy. Andrés Malamud (PhD European University Institute, 2003) is a senior research fellow at the Institute of Social Sciences of the University of Lisbon. His research focuses on comparative regional integration, foreign policy, Latin American politics, and EU Studies. His work has been published in journals such as the Latin American Research Review, Cambridge Review of International Affairs, Latin American Politics and Society, Journal of European Integration, Latin American Perspectives, and European Political Science.