11—
The Military State and Its Future:
Adventures and Misadventures of Demilitarization

A judgment on the possible future evolution of military-dominated regimes depends on the theories adopted to explain their unprecedented emergence in Latin America. If we view contemporary militarism as a type of culturally produced anachronism that provides temporary resistance to the supreme political good—representative democracy—this will result in a theory of a predictable and practically inevitable unilinear evolution. Structural interpretations of the appearance of modern authoritarian regimes also emphasize their transitory character. Functionalist determinism, by establishing a more or less instrumental correspondence between the dominant economic actors and the type of regime, predicts an end to the authoritarian systems when their supposed "objectives" have been fulfilled. Because of the "inevitable" or indispensable character of the authoritarianisms for peripheral capitalism in the present period, their disappearance is also historically determined. These two contradictory views have in common a fixed dogmatic attitude concerning the "exceptional" character of authoritarian regimes. In effect both those who interpret Latin American history in terms of the "struggle for democracy" and those who view the field of politics as directly subordinated to the needs of capital assume the inevitability of liberalization.

We should note that the supporters of these two theses generally pay little attention to the fact that the great majority of

the Latin American authoritarian regimes are military. The first group does so because they have decided once and for all that armies in politics are a thing of the past. Since the mark of modernization is representative democracy and functional specialization, the obstacles to popular sovereignty are the result of the weight of the traditional past. On these premises there is no attention given to the bureaucratic modernity of professionalized military institutions and the need to discover its relevant effects on politics. "Economistic" views also ignore the military component. Their approach neglects an institution that is at the center of power because it is assumed to be nothing more than the expression or instrument of other socioeconomic forces. In short, its special characteristics and particular processes form a sort of epiphenomenon.

A nonreductionist approach that focuses on the real holders of power in the political systems under military domination while taking into account group differences, alliances, and civilian support, as well as extrainstitutional political resources within a framework of the structural constraints experienced by national societies, cannot accept as predetermined the types of political organizations that will succeed authoritarian regimes. This does not mean that military power will last forever, but that it has a logic of its own. The successive waves of militarization and demilitarization that the continent has experienced since 1945 argue for caution.

In fact, while in 1954 twelve of the twenty republics were governed by military men who had taken power by force, only one remained by the middle of 1961, Stroessner in Paraguay. In seven years, revolution or assassination had ended the rule of ten military presidents, while another in Peru had "retired."^[1] It is true that those generals headed very different regimes, including democracies, and the removal of the leader did not always change the system, as the situation in Nicaragua after the assassination of the rather unmilitary dictator, Somoza, in 1956 demonstrates. Often these systems are only military in terms of the original profession of their president; however, they have evolved in contrasting ways. Should one attribute to an antimilitary movement the deposition of Perón, who had been legally reelected constitutional president in 1951, or those

of the Venezuelan tyrant, Pérez Jiménez, General Magloire in Haiti, or Colonel J. M. Lemus in Salvador, even though it is true that all those military men along with Batista in Cuba and Rojas Pinilla in Colombia were, for the time being at least, indeed the candidates of the military for the presidency? What are we then to say after those changes concerning the military tidal wave that ended civilian regimes in nine countries of the continent between March 1962 (Argentina), November 1964 (Bolivia), and June 1966 (Argentina again)? Furthermore, how are we to situate—as a continuity or a new phenomenon—the series of coups d'état at the beginning of the 1970s that struck countries with solid traditions of civilian government such as Chile and Uruguay, while in Argentina a new military intervention exhibited a violent character that was unheard of in the history of the nation?

Nevertheless, beginning in 1976–77, democracy seems to have been gaining ground. It appears to be time for a liberalization of regimes based on force and a return to civilian government. If we judge only by the figures, in 1978 twelve popular elections took place on the continent. That intense electoral activity seems to indicate that there will be a return of representative procedures. In fact, that figure covers both authoritarian votes and competitive elections or ambiguous maneuvers. The referendum in Chile and the reelection of Stroessner for the sixth time do not seem to indicate—far from it—the end of despotic systems. In Venezuela and Colombia elections that are customary in those model democracies do not constitute anything special. In Brazil the legislative elections took place within the framework of the military system under restrictions and manipulations that were aimed at guaranteeing its continuation in power, but nevertheless they had unfavorable results for the government. However, in Peru, Ecuador, and Bolivia those consultations were aimed at preparing for the return of the civilians to power, the free play of democratic institutions, and an orderly withdrawal of the military to their barracks.

This historical survey does not argue for a single simultaneous interpretation of military rule, as we have explained above. Nor do we believe that these continental movements condemn the states of the subcontinent to an indefinite alterna-

between civilian and military regimes. Rather they indicate that the forms of demilitarization are complex and diversified and that they may have their limits. Recurrences and retreats, rather than confirming a priori generalizations, invite us to examine the realities of demilitarization and therefore the real impact of the militarization of the state. Does it consist in a simple interlude with no institutional consequences after which there is a return to the previous regime once the army is back in the barracks? Or, on the contrary, do the military withdraw only when they believe that they have eliminated the political obstacles to a civilian regime and created the socioeconomic conditions favorable to the normal functioning of democratic institutions? These are questions that we can answer only by looking at what happens after military rule.

Controlled Usurpation

The instability of governments based on force has often been noted. Institutionalized military regimes, even when they appear to be the rule in a country, still remain exceptions, as paradoxical as that may seem. In fact, the dominant official ideology throughout the continent is liberal and pluralistic. The constant changes in military systems and the short duration of noncivilian governments are partly related to their lack of legitimacy, as perceived by those involved. In the cultural and normative context of Latin America those who hold military-based power always know, whatever they may say, that above them there is a higher legitimacy, that of constitutional legality that they may claim to possess, but to which they must finally appeal.^[2] Military regimes are only really legitimized by the future. While elected governments are legitimate because of the way they originate, de facto governments only acquire a legitimacy in exercise, from their performance, so to speak. The past can justify the arrival in power of the military; the customary references to social and political chaos, to the vacuum of power, to threats of all kinds, nevertheless become transformed into objectives to be achieved. Military regimes thus look to the future. They are essentially transitory. Also, a permanent military sys-

tem is a contradiction in terms.^[3] An army can govern directly for an extended period of time only by ceasing to be an army. Also, it is precisely the following government, the successor regime, that is the basis for the legitimacy of the military usurpation of power.

Even if we work with the relatively arbitrary distinction between provisional (or caretaker ) and constituent military governments, in neither case there is an avowed and declared intention to create a new type of state, a definitive and lasting power. A democratic regime is always more legitimate in Latin America than the omnipresent "state of exception." Contemporary military regimes in Latin America differ in this respect from the modern dictatorships that Europe and other continents have known precisely because of their constitutive weakness. They do not claim to create a new legitimacy or to put forward a new system of political values among the ruins of the old. The authoritarian regimes of Europe between 1920 and 1945 aimed at the creation of a "new order," even a "thousand year Reich," as opposed to liberalism and democracy. The military dictatorships of Latin America today are first of all regimes without an ideology. The "national security doctrine" that those institutionalized military governments share to a lesser or greater degree furnishes a rhetoric that conceals their illegitimacy, rather than providing a new source of legitimacy. That doctrine was above all a way to forge a mobilizing consensus within the military institution around an image that was related to their professional alarmism. Their theories of war, by enlarging the spectrum of threats and locating them within the nation itself, gave a corporate basis for the army's intervention in politics, but they did not explain it. They could justify a lasting presence in positions of command in the state, but they did not establish a new power. In a word, the theory of national security in no way takes the place of an ideology, not in its consistency, or its diffusion, or its constitutive function.

This is why representative democracy is always on the horizon of these regimes. They appeal to it both in their legitimation and in their objectives, proposing to improve it, to strengthen it, to amend it, even to protect it, but never to abolish or destroy it, as was the case in other areas. This tells us something about the

Brazilian "system" that has always retained (with proper safeguards) parties, elections, and a congress—and even about the archaic militarism of Stroessner who, like all the classic dictators of the continent, keeps having himself reelected to the presidency and tolerates under strict surveillance a decorative but genuine multiparty system. The proclamations, declarations, programs, and maneuvers of the military in power in Uruguay or in Argentina refer to no other political system, to no other legitimacy than the traditional one of liberalism. Perhaps this is only a facade, but it operates against military messianism and prevents any program of remaining permanently in power. The military in power, however central their position in the political system and however great their autonomy, are participants in a political culture of the dominant internal and external classes who share a self-interested liberalism that acts as a check on the organic ambitions of the men of the barracks. Everything happens as if the dominant classes believed that the reestablishment of the market in economic affairs could only truly be legitimized in the name of a certain reestablishment of the political market.

Thus in Argentina every corporatist and antiliberal intention on the part of the military in power, from Uriburu in 1930 to Onganía in 1966–70, produced a reaction in the economic and social establishment and the replacement of the "anticonstitutionalist" generals by liberal military men.^[4] In Uruguay, Bordaberry, the civilian president of a military dictatorship imposed by the gradual coup d'état of 1973, was removed by the general staff in June 1976 for having been accused of favoring "new institutions" that were opposed to democracy.^[5] In fact, he had proposed the elimination of the party system and the establishment of an authoritarian "new state," the legitimacy of which would be guaranteed by the armed forces alone. However much they militarized the real exercise of power and however strongly the overpowering presence of military institutions and its representatives marked the whole life of the society, the Uruguayan generals found it difficult to give up the fiction of a civilian president: the garrison-state of Uruguay had a nonmilitary president until 1 September 1981 as well as a government from which the officers were practically absent. The parties were only "suspended" and the text submitted in

the constitutional referendum of 30 November 1980, while it formalized the participation of the armed forces in the executive, provided for the legalization of the two traditional parties and the return of a restricted and purified version of representative procedures. The rejection by the electorate of that proposal after the semblance of a campaign was able to demonstrate that the military had been correct not to underestimate, even after seven years of prohibition and hostile propaganda, the strength of the support for the party system. This was also demonstrated in other countries such as in the Peruvian elections of May 1980 and the Argentine elections in 1973 after, respectively, twelve and seven years of the suspension of institutionalized political life.

The government in Chile, of which General Pinochet has been president since 1973, figures among the most antiliberal military regimes in the continent and is among those that concede the least to democratic rhetoric. On the contrary the authoritarian language of the Chilean military, with its insistence on the need for new institutions, has Francoist accents. Corporatist tendencies are expressed without concealment by the advisers and leaders of the "hard" line of the regime—the "renovators," to use their expression—who absolutely reject the institutional system that was in force until 1973. After the coup d'état General Pinochet himself announced a new constitution that was supposed to "banish forever the politicians, sectarianism, and demagogy."^[6] The minister of the interior declared in September 1975, "All the political parties only divide the citizenry, favor their supporters in a demagogic way, and undermine the soul of the nation." The influential El Mercurio, spokesman for the moderates (blandos) and supporter of a moderate opening, commented on his statement as follows: "The government desires the destruction or the progressive disappearance of the parties."^[7] In the constitutional debate, while the goals and time periods announced in the plan of Chacarillas (July 1977) reflected the desires of the "hardliners" for the establishment of an "authoritarian democracy," the constitution submitted to the plebiscite of 11 September 1980, apart from its gradualism and the restrictions on freedom that it contained, called for the establishment in the relatively distant future of a representative sys-

tem that would include parties, a congress, and a president elected by universal suffrage. It is obvious that that juridical apparatus is aimed above all at justifying Pinochet's remaining in power. However, the utilization of a constitutional text that is not corporatist in inspiration and the fixing of a time limit to the state of exception is sufficient proof that even in the Chilean case the antiliberal temptation, the desire to exclude the "vanquished" forever, has given way to accommodation with the dominant democratic ideology.

The attempt to place representative practices under strict surveillance differs fundamentally from the ways and means adopted by dictators in other continents who had the same objectives. If we compare the regime of General Franco with that of General Pinochet the similarities are obvious, but the differences are no less clear. The two counterrevolutionary systems intended to break with the previous political situations, to refuse civil rights to political dissidents,^[8] to keep the "vanquished" definitively out of power by prolonging the victorious (by coup d'état or civil war) coalition through the unlimited personal authority of the military leader of the counterrevolution. However, in Franco's case there was no concession to pluralism for forty years except at the summit of the state in his bourgeois-technocratic coalition. Liberal democracy was rejected forever regardless of internal changes and the international context. Franco, caudillo of Spain por la gracia de Dios never put into question even incidentally his remaining in power permanently. Neither the referendum of 1947 nor that of 1966 posed the question of the choice of the chief of state nor the length of his mandate. The opposition also finally adopted the idea that the dictatorship was for life and that a change of regime could only take place after the death of the caudillo.^[9] General Pinochet, however, specified the time period of his provisional regime (it is true, after only four years), whatever may have been his real intentions for the future, and he did not exclude the possibility of the return of the parties and of competitive elections, modified of course by various prohibitions aimed at "protecting democracy." Thus, Pinochet proved that however much one might desire it one cannot create a new form of legitimacy in an environment that is hostile to such ideological adventures. Having made these remarks de-

scribing the limits of the militarization of the state in Latin America, let us see in what way demilitarization has been carried out—as well as on what level, to what degree, and with what kinds of regimes.

The Postmilitary State and the Forms of Institutionalization

The withdrawal of the army from power involves very diverse phenomena. The "civilianization" of the military state, however complete it may be, does not necessarily mean a return to "normal democracy." In order to analyze comparable situations we will only examine genuine systems of military domination, that is, regimes established by a coup in which the sovereignty of the armed institutions is exercised collectively over the selection of the chief executive and over all the major decisions of national importance, apart from the extent and content of civilian alliances or the background of the members of the government. We will therefore leave aside authoritarian regimes of other kinds, whether patrimonial or party, even when coercion and the participation of the officers play a major role.

We can dismiss at the outset a first type of demilitarization, the one secured by force through a civilian pronunciamiento. In fact, it is generally the military who overthrow the regimes of their peers by violence (more often, through the threat of the use of violence). Certainly many personal dictatorships, patrimonial autocracies, and postmilitary tyrannies have been removed through civilian uprisings allied at times to groups within the armed forces. Without going back to nineteenth-century Peru or to the civilian montoneras of Nicolás Pierola, a civilian-military revolution in Guatemala in 1944 overthrew General Jorge Ubico and his short-lived heir. In the same year in El Salvador students and soldiers ended the dictatorship of Hernández Martínez, whom his army no longer supported. Guerrillas, that is, civilians, defeated the National Guard of Somoza in 1979, and ended the dynasty in Nicaragua, repeating thus in different circumstances the Cuban precedent. However, among institutional military governments, only that of

Bolivia in 1952 was overthrown by civilians. The military junta that annulled the electoral victory of the Movimiento Nacionalista Revolucionario (MNR) was in fact driven out in the streets of La Paz. The militarization of power that in time, at least, had been relatively limited, was followed by a drastic demilitarization with the Bolivian army being largely demobilized and its leadership violently purged so as to remove any possibility of its being a threat to the new revolutionary government.

The most frequent forms of demilitarization consist in leaving the military system in place but removing the army from power. Since for reasons that are both internal and external to the armed forces direct military government cannot be permanent, a number of methods are employed to maintain the continuity of martial power. They can be classified into two main tendencies—that of personalization, and of legalization—with the two models linked or not to a real or only apparent opening toward democracy.

The transfer of power to a military leader who takes control over those who put him in power constitutes one way of subordinating the armed forces to the executive and returning the army to its professional tasks. The transfer of power from impersonal institutions to the person of one man never takes place very easily. That process is evidently less difficult when the army is less statist and bureaucratic. Somoza, jefe director of the National Guard in Nicaragua, and Trujillo, generalisimo of the Dominican army, "personalized" the neocolonial military institutions that were entrusted to them before seizing power in their own names, and not as representatives of the army. It was different in Bolivia where Barrientos could overcome his rivals while having his power as first among equals ratified through an election and also drawing on an historical-military legitimacy (the tradition of the Chaco war) and a quasi-personal popular support (the military-peasant pact). The installation of Barrientos as constitutional president prolonged both the military junta and at the same time the preceding legal regime of which Barrientos was vice president. General Banzer had less luck in this area, it seems, than his predecessor. Having come to power through a coup d'état in 1971, he governed until 1974 at the head of a conservative coalition that was made up of part

of the political class. In 1974, when he reorganized his government and replaced the politicians of the MNR and the Falange with generals, he seemed to emerge with increased personal power, but in fact the army took over the state once again.^[10] General Banzer, after having announced on several occasions beginning in 1974 that there would be presidential elections, in 1978 had to resign himself to not being a candidate as a result of pressure from the army. He supported Juan Pareda, his former interior minister, while the armed forces, which were divided, announced that they would remain neutral. The elections of July 1979 were followed by a coup d'état by the "official" winner, who remained in office after being improperly elected and clumsily institutionalized by the military.

Democratic endorsement can also make it possible for there to be a legal resolution of a deadlock in a regime of the military that allows them to survive. In Argentina in 1945 the regime that emerged from the 1943 coup d'état was caught in the cross fire of internal and external opposition that was intensified by the defeat of the Axis powers, but the "workers' colonel" was at the height of his popularity. Critically regarded by a part of the army that was opposed to his proworker behavior and his political ambitions, Perón provided by his candidacy in free presidential elections an honorable way out for the institution that had brought him to power and which he claimed to represent. The "revolutionary" officers, even those hostile to the lider, had a choice between allowing the return of the traditional parties and a continuity represented by the former vice president of the military government. During his first presidency General Perón was careful to recall his military origins and to appear as the inheritor of the "revolution of June 4, 1943." Thus, by electoral endorsement of a candidate of the army, or what could be described as such, the military institution found coherence again and in theory ceased to make decisions. The vertical hierarchy of discipline was established once more, recreating the internal unity that had been disrupted. The process of demilitarization can stop there or it can go further and be extended using the alternative political resources available to the military leader to the point that he can sometimes cut himself off to a large—and dangerous—degree from

his bases within the armed forces. This is what happened to Perón beginning in 1951.

The transfer of power to a military head of state can allow demilitarization without immediate recourse to dangerous electoral procedures. The military takeover thus ends up with a one-man dictatorship. This seems to be the pattern today in Chile. From 1977 the prolongation of the role of the military in response to the tutelary role that the armed forces had assumed strengthened the absolute power of General Pinochet. His irresistible rise, which relegated the junta simply to a legislative and constitutive role, was cleverly ratified by the success of the unexpected referendum of January 1978; the text of this document, which was imposed on the other members of the junta, stated, "I support General Pinochet."

In the Chilean case we may conclude that the high level of professionalization and the lack of political experience of the armed forces were factors that had much to do with that process of personalization-cum-institutionalization of the military regime. Hierarchic discipline took the place of political consensus, and fear of the return of the "vanquished" cemented cohesion around the leader who was a symbol of a counterrevolutionary policy that no one in the army questioned. This is why we can understand the lack of response within the institution to the criticisms of the political proposals of General Pinochet by General Gustavo Leigh, the air force representative in the junta, and to Leigh's subsequent removal and early retirement, as well as to the resignation of eighteen of the twenty-one air force generals. The slow pace of the "constitutional timetable" and Chile's international isolation operated in the direction of reinforcing military support for an institutionalization without an opening that gave the army a guarantee of what was essential to them. While the army no longer governed it was never far from power, and, more important, it felt that its needs were understood.

In most cases what is called the institutionalization of military regimes amounts to their legalization in conformity with the prevailing constitution. That change, which has some elements in common with a return to democracy and produces a certain liberalization, means that military power is exercised

purely and simply within the institutional framework that is considered to be legitimate and makes use of it. At the same time, of course, the major sources of uncertainty that are inherent in the democratic process are removed. These procedures can produce "military governments that are both elected, constitutional, and anti-democratic"^[11] as in Guatemala. This process of legalization is carried out generally in one of two ways: the formation of a controlled and coercive multiparty system, or the creation of a dominant military party .

An example of the latter is the system in force in El Salvador from 1950 until October 1979, the date of the overthrow of General Carlos Humberto Romero by a civilian-military junta. The military in power in 1948 tried to create an official party, the Revolutionary Party of National Unification (PRUD),^[12] a veritable party of colonels modeled on the Institutional Revolutionary Party of Mexico but without its popular bases. The Party of National Conciliation (PCN) that succeeded it was the party expression of the military institution as well as its electoral extension.^[13] However, it was also the state party in which under the aegis of the army, arrangements were made between the civilian or military bureaucracies and the dominant class. Alternately allowing or forbidding political competition (when the PCN lost ground) the "military party" controlled political life, obtained a majority in the Congress, and had a colonel or general elected to the presidency—although not without a certain amount of fraud in 1972 and 1977. The defeat of the PCN by the opposition in 1972 produced the disintegration of the semiopen electoral system. The fraud, repression, and the limitation of electoral competition that followed demonstrated the impotence and decomposition of the machinery that had been created to provide a legal guarantee of the continuismo of the military state.

The institutionalization of the nationalist military regime of General Torrijos in Panama seems to have followed a course parallel to that of the Salvadoran colonels, apart from their differences in political orientation. The Democratic Revolutionary Party (PRD) established by his supporters nearly ten years after the coup d'état carried out in 1968 by the national guard against the traditional oligarchic parties also seemed to aim at becoming a Mexican-style institutional party. Its success in the congres-

sional elections of 1978 enabled the new civilian president who was elected by the assembly to democratize the regime without taking great risks.^[14] Would the revival of political life take place at the expense of the PRD and would the process of democratization go as far as the acceptance of a possible defeat of the official party? By keeping the command of the National Guard, General Omar Torrijos, the strongman of Panama, had reestablished the classic pattern on the continent of a military caudillismo that does not allow such a possibility to be predicted. Also, it was whispered in Panama that Aristides Royo had only a six-year lease on power, granted by Torrijos.^[15] However, Torrijos's accidental death in August 1981 opened a political vacuum that the National Guard could be tempted to fill.

The very fluid political-military situation in Honduras offers a special example of institutionalization through a traditional two-party system. Although, as in Peru, after the military reformers came to power in December 1972, they faced conservative forces who called for a return to institutional normality. Only following the removal of General Lopez Arellano, and then of his successor, Melgar Castro in August 1978, did the government of the armed forces begin a third period in which it ended the cycle of reform. The National party (conservative) that supported the new government offered to play the role of "military party, that is, of a civilian organization through which the military could continue to exercise power."^[16] For that purpose elections were necessary. They took place on 20 April 1980, but unexpectedly the Liberals, the traditional adversaries of the National party, were victorious. That vote of protest against the military—thanks to the goodwill of the Liberals and to international circumstances—did not result in a coup d'état to annul the unforeseen electoral results: liberal and conservative members of the Congress joined in voting to elect as provisional president of the republic until the next elections (after the adoption of a new constitution) General Paz García, the head of the military junta,^[17] and the party that won the elections found itself in a minority in the government! Finally, after many uncertainties the candidate of the Liberal party was elected president with a comfortable majority in the general elections of November 1981. But President Roberta Suazo Cordova had given so

many guarantees to the ultras in the army that the new civilian regime could not upset the military.

In Guatemala the state was profoundly militarized. The army not only exercised power but also many civilian functions, constituting a veritable bureaucratic bourgeoisie,^[18] while the chiefs of the general staff supervised the nominations to all posts of responsibility. Despite the use of more or less regular competitive elections, there was no military party. However, in 1978 all three candidates for the presidency were generals, and since the overthrow of the progressive civilian president, Jacobo Arbenz, in 1954 by Castillo Annas, "anti-Communist" governments supported by the army have been in power with or without popular ratification. Beginning in 1970—in a climate of increasing violence—generals regularly succeeded one another in the presidency as a result of elections that the army always won. The same scenario was repeated with some variations: the armed forces chose a candidate who was necessarily to become the chief executive and then they negotiated with one or two parties of the right or the extreme right that could provide the electoral bases. Pluralism was limited by a "constitutional range" from which the left parties were banned by definition.^[19]

In 1970 Colonel Carlos Arana Osorio was elected president with the support of the Movement of National Liberation, "the party of organized violence" and of counterterrorism. In 1974 General Kjell Laugerud was the candidate of a coalition of the MLN and of the Institutional Democratic Party (PID); in 1978, the so-called Revolutionary party allied with the PID to elect General Romeo Lucas García. It seems that only Arana Osorio actually won the elections and his successors gained power by fraud or the use of force by the preceding government. Thus General Laugerud definitely obtained fewer votes than General Rios Montt, but the government had his election ratified by the congress.^[20] Rios Montt, not having enough support in the army, had to leave the country. These constitutional governments are therefore the expression of a military state that is legalized along the lines of "controlled coercive multipartyism." However, they also represent forms of demilitarization that can alternately close and open toward the establishment of less exclusionary pluralist systems.

The evolution of Brazil shows both the ambiguity and the ease with which a controlled redemocratization can be carried out by the military when it has not eliminated democratic procedures, even if it has emptied them of their content. The policy of "decompression" and of opening initiated in 1974 by General Geisel and pursued by his heir, General Figueiredo, clearly produced a liberalization. The elimination of the dictatorial powers given to the president by Institutional Act No. 5, the ending of censorship, an amnesty, the return of the political exiles, the reestablishment of direct election of governors and of most of the senators, were so many steps in a "gradual" redemocratization carried out by the government at a speed of its own choosing. The reactivation of civil society and the enlargement of the recognized political arena (as demonstrated by the multiplication of publications of the extreme left that circulate legally) can still be viewed as involving a new strategy of institutionalization following the failure of the coercive two-party strategy adopted in 1965.

The continuing electoral advance of the tolerated opposition front (the Movimento Democratico Brasileiro) beginning with the legislative elections of 1974 and the weak hold of the official ARENA party, put the government in a delicate and possibly unstable situation. A well-ordered opening, on the other hand, could assure continuity while ending the situation of a "plebiscitary impasse"^[21] into which the system had been driven through its identification with the government party and the dual choice offered to the electorate. Some observers believed that the return of the pre-1964 leaders to political activity and the reestablishment of a multiparty system were measures that would lead to a split in the MDB and would therefore weaken the opposition by freeing it.^[22] While for the time being the new party law certainly did not succeed in isolating the left by producing basic political realignments, it resulted at least in the formation of two conservative parties: the Partido Democratico Social, the president's party, and a progovernment opposition Partido Popular Brasileiro (PPB), incorporating the moderates. That new political spectrum was to make it possible to have an alternation without risk that was acceptable to the military provided that the opposition was divided and, better still,

atomized. However, the prohibition of "electoral alliances" aimed at preventing the establishment of a united opposition front forced the PPB to combine with the PMDB, the new version of the MDB, while the unexpected rise of the "Workers party" (PT) further complicated the programmed opening.

We may therefore conclude that that opening constituted a new effort at legitimation of an isolated regime in crisis that was seeking to enlarge its base. The "slow and gradual" democratization was not supposed to be a prelude to the transformation of the "system," but a continuation of its long-term practice of changing the rules of the game when they were unfavorable to it. This new manifestation of casuismo and of the flexibility of a regime that was a past master at elections in which the loser won had an effect on the nature of the system itself,^[23] despite all its built-in safeguards. As F. H. Cardoso correctly points out, up to this point "the system legitimated the parties,^[24] but now the parties had become essential elements for the regime to function, so that the head of the state was presented as the leader of a party. In this setting liberalization develops its own dynamic. The utilization of authoritarian measures to channel a democracy that has been conceded in this way is no longer possible. It is only by playing the political game that the program can succeed and provide the regime with what it expects. A return to authoritarianism, which is always a possibility, would result in the loss of the political dividends of the whole strategy. However, political liberalization is not supposed to produce a social opening, as long-repressed and restricted demands emerge almost spontaneously. The repression of the large-scale strikes of April and May 1980 and of free trade unions seemed to indicate that the regime did not intend to give up its powers of control over "the dangerous classes" that were a legacy of the Vargas era and that, it is true, the "democratic experiment" of 1946 to 1964 had not changed. Will that authoritarian restriction remain and does it indicate the limits beyond which liberalization cannot go—even the price that must be paid for it to be permanent? Whatever the case, the regime does not intend, it seems, to surrender power or to lose the initiative. It has taken every step to ensure that democracy will operate in its favor. Rather than a relative democracy, there-

fore, it is a democracy in which those in power cannot lose control of what comes after them.^[25] The test that remains is evidently the presidential succession and the way that it is carried out.^[*] The revival of civil society, and the reactivation of the parties and of congressional life by narrowing the scope of authoritarianism, have reduced the space for the exercise of military sovereignty. The regime has changed its nature, but to whom does power belong?

Civilian Government and Military Power

We have seen the ambiguous character of a liberalization that is carried out without a break, but we have also seen that the means that are provided in order to maintain a facade of democracy already imply a certain form of demilitarization. In the recent history of Latin America noninstitutionalized military governments have generally agreed to withdraw from power only in exchange for certain guarantees. They make efforts to fix the rules of the game. Better still, when the situation permits it, they do not hesitate to demand a place for the military institutions in a democratic constitutional order that enables them to exercise a permanent right of supervision over political decisions. The draft constitution submitted by the Uruguayan military to a referendum in November 1980 in order to provide a juridical basis for their de facto power stipulated that the National Security Council (COSENA) made up of superior officers would have the right to make accusations against the members of the executive and the legislature without being responsible themselves to any other body, that it could intervene in "activities relating to national security," and could even (with the president) declare a "state of urgency" without referring to the congress, except after the fact.^[26] That tutelary democracy, as we know, was rejected by the voters after being condemned by all the parties from the Frente Amplio on the left to the traditional Colorados and Blancos.^[27]

[*] For the events that led to the election of a civilian president in Brazil in 1985, see the Epilogue. (Translator's note).

In Argentina in 1972, the military, who had been in power since 1966, decided, in order to avoid a social explosion, that elections would be held from which, for the first time since 1955, no group would be excluded. However, they wanted to avoid a "leap into the void" that might permit, in their words, a return to the "fatal errors of the past." To that end, General Lanusse, president of "the government of the armed forces," wished to conclude an agreement with the political forces on guarantees that were to be implemented by the army. Looking for an honorable way out, the army even made the holding of elections conditional on the conclusion of a "Grand National Accord" by all the political groups under its aegis. A military transitional candidate of national unity would not have displeased the general staff. After the political forces rejected any institutionalization of the participation of the military in a reestablished democracy and the attempts at an official candidate failed, the military put in place in extremis a double insurance policy by amending the electoral law to prescribe a two-round balloting system for the election of the president and a residence requirement that would prevent Perón, who had been forbidden to participate in political life since 1955, from becoming a candidate. The accumulation of protections and stratagems imposed by the de facto regime did not secure the support of the political forces. Finally, the junta of commanders in chief, in the absence of an agreement, issued a declaration that listed the principles that the military wished to have respected. Its text provided that the armed forces would be opposed, among other things, to an "indiscriminate amnesty" for crimes of subversion, and stipulated that they ought to "share in the responsibilities of government."^[28]

In reality the regime had lost the initiative. The electoral victory of the Peronist candidate swept away the checks put in place by the outgoing government. The slogan "Campora in government, Perón in power" was a direct challenge to the proscription imposed by the generals. In addition, despite the provisions of the electoral law, the military declared Perón's candidate, Hector Campora, elected with only 49.5 percent of the vote in order to avoid a second defeat in a presidential electoral round that would be more agitated and more mas-

sively hostile to the government. The two political groups against which the 1966 coup d'état was carried out, the Peronists and the Radicals, received 70 percent of the votes. A semiofficial candidate did not even get 3 percent of the votes. All the candidates who represented continuity with the military hardly received 18 percent of the votes.^[29] What is more, the new government promulgated an immediate general amnesty, and the president-elect rejected all the institutional suggestions as to who was to represent the armed forces. The leadership of the army was even decapitated by the nomination of a commander in chief who did not belong to the cavalry, which had dominated the army since 1960.

Similarly, in Ecuador the army that had seized power in 1972, when it withdrew tried to secure the acceptance of demands analogous to those of the Argentines. When the Ecuadorians decided to hand over power to the civilians after a palace revolution in 1976 had removed General Rodríguez Lara, they announced their wish to give the country a truly representative democracy. Nevertheless, the junta took some precautions, or rather tried to institute a democratic system that conformed to the image of the military. The process of democratization therefore was characterized by a prudently slow pace. It took no less than three years and began with the prohibition of the candidacies of the three most representative candidates who were considered to be dangerous demagogues by the army. A made-to-measure electoral law adopted in February 1978 provided that the future president was forbidden to have occupied that post in the past—which blocked the way to Velasco Ibarra, the eternal caudillo who had already been president five times, as well as to Carlos Julio Arosemena. Another ad hoc provision provided that the future president must be an Ecuadorian and the son of an Ecuadorian. The requirement was directed against Assad Bucaram, head of the Concentration of Popular Forces, one of the largest parties in the country, who was the son of a Lebanese. That populist leader and moving speaker, who enjoyed a large support in the subproletariat of Guayaquil, was the main favorite in the electoral race, as he had been in 1972 at the time of the coup d'état.

The use of the veto on the selection of candidacies, which

was contrary to democratic principles, augured badly for the reestablishment of a constitutional regime. The utilization of the French-style two-round election system, which only allowed the two front-running candidates to remain in the second round, seemed to be aimed at the establishment of a right wing front in the second round. The separation of the two rounds by nearly ten months and the numerous incidents during the campaign did not lead one to expect that the military would respect the results if they did not correspond to their wishes. They supported the conservative candidate, Sixto Duran, in an almost open way, while Bucaram, who had been excluded, was represented by his nephew by marriage, Jaime Roldos. However, Roldos won the elections and became constitutional president of Ecuador in August 1979 without any attempt by the military to question the election results that emerged at the end of a difficult and uncertain process.

It does not always happen in this way. The military seem to agree to retire only if the civilian government is very similar to their own or their candidate wins the elections. In all other cases the election may be invalidated, either immediately or after a period of observation that is more or less long, and when the circumstances are appropriate. The increase in the number of "contentious elections," to use the term of F. Bourricaud, is a result of that continuista behavior. The agitated political life of Bolivia from 1978 until 1980 is a good example of that tendency. General Juan Pereda, the official candidate of General Banzer, "unelected" in the July 1978 election, was the originator of a coup d'état on 21 July in order to guarantee a "victory" that was very much disputed, especially by the candidate of the moderate left, Hernán Siles Suazo. In November 1978, the legalist sector of the army led by General David Padilla overthrew General Pereda and called new elections for June 1979. When those elections did not produce a clear majority, the president of the senate was put at the head of the state. The process of constitutionalization was continuing when Colonel Natusch Busch seized power on 1 November 1979, but he was forced to resign two weeks later. He was replaced by the president of the Chamber of Deputies, Lydia Gueiler. New elections took place on 29

June 1980 which demonstrated a clear movement to the left. Siles Suazo, who came in first with a center-left coalition, was supposed to be named president by the congress on 4 August. General Banzer, who was a candidate in the elections, only received 15 percent of the votes. On 17 July 1980, a victorious and bloody coup d'état installed General García Meza in the presidency. Those who carried out the coup no longer spoke of elections. Their primary objective, "the extirpation of the Marxist cancer," postponed any form of institutionalization for the indefinite future, while military instability continued.^[*]

Lacking the power to impose a government to their liking that would keep them in power, the armed forces can make use of measures of corporate defense that are far from promoting the reestablishment of civilian supremacy in all areas. Also, the "postmilitary" civilian regime does not completely resemble those that have preceded it if the elected authorities lose the upper hand in the nomination of those responsible for the army. In fact, the affirmation of military autonomy is often the legacy of the militarization of power or the price to pay for the return of the military to their barracks. In Peru President Belaunde^[*] , who had been elected after a military interlude in 1962—and who was to be deposed in October 1968, and then reelected in 1979 following the return to democracy—was forced in 1963 to designate as head of each branch the officer who was at the top of the promotion list and to choose his military ministers in conformity with the wishes of the High Command. In Ecuador, shortly before the first round of the presidential elections in July 1978, the military amended the organic law of the armed forces and declared that the future president would be required to name as minister of defense the officer who was highest in the hierarchy.^[30]

A military defeat at the ballot box that is accompanied by a complete rejection by an exasperated public opinion, such as happened in March through May of 1973 in Argentina, does not guarantee a return to full representative democracy, even

[*] In 1982 Siles Suazo reassumed the presidency for the remainder of his term. See the Epilogue (Translator's note).

when the army respects the results of the elections. The fact that the government has been demilitarized does not mean that the military have given up power in cases in which they have become quasi-legitimate players in the political game. Thus, from 1973 to 1976 the military, who had apparently been removed by the electoral tidal wave of Peronism, followed the development of the political situation step by step. Perón removed his lieutenant, President Campora, only after the general staff had returned to him his rank of general and had signaled approval. The army was not absent from the public stage under a series of commanders in chief, who, whatever their inclination to neutrality, were faced with a regime that was rapidly coming apart after the death of the leader. The desire on the part of the government of Isabel Perón for the participation and, at the outset, legitimation of the military was to produce a very serious crisis in August 1975 that was a prelude to the fall of the civilian government. The ostentatious apoliticism of the Argentine general staff turned out in March 1976 to be one of the more subtle forms of military intervention. The theory of the "ripening of the fruit" and the willingness of the military to let the situation deteriorate argue that the 1976 uprising was neither spontaneous nor accidental.

The fact that the Argentine army never left power completely does not mean that if a country has experienced military intervention even once in the modern period it is condemned to its continuing recurrence.^[31] There is no doubt that with a half-century of military domination Argentina is an extreme case of a military-dominated political system. Yet, who would deny that the withdrawal of the armies is never definitive and that the postmilitary state, however democratic, always lives in the shadow of the barracks? The burden of that shadow affects the conduct of the actors, whether they wish to avoid a coup or to produce one. However, there is nothing inevitable about it. The longer military intervention does not take place, the more civilian power is reinforced, military usurpation made more difficult, and the political system demilitarized. On the other hand, the threat or continuing fear of a coup already amounts to actual intervention, as we see in Spain, where, since the death of Franco, for good reasons references to the "tolerance" of the

military are common in political life, while the ghost of General Pavia's horse haunts the parliament.^[32][*]

Demilitarization therefore is a matter of degree. The return of the civilians to government is not automatically equivalent to the "civilianization" of power, even after free and representative elections. We must ask why and under what influences and circumstances the military hand over power to the civilians; but also it is necessary to understand the limitations on the process of "extraction" of the military.

The Hour of the Civilians

The many theories that can be suggested concerning the causes of the transition from military authoritarianism to representative government in Latin America do not clarify the problem. The political, social, and economic factors that are generally cited actually apply to all kinds of authoritarianism, not just to the military version. Furthermore a number of them seem to be of little explanatory value because they can be turned in the opposite direction and thus seem to possess a "mythological" character. This is the term that is used by Wanderley Guilherme dos Santos when he criticizes the contradictory "economistic" explanations of authoritarian intervention. As he writes:

Thus economic recessions are cited both as an explanation of the erosion of authoritarianism, since they make it impossible according to these theories for these regimes to coopt the masses and/or the elites by the distribution of benefits, and yet the same recessions are presented as an explanation for the survival of authoritarianism because only the conditions of authoritarianism will permit those regimes to suppress demand in conditions of extreme penury.

Conversely high rates of economic accumulation and growth are used to explain the maintenance of authoritarianism because thus the governments can anesthetize the population, especially the masses, by

[*] In 1874 General Pavía organized a coup against King Alfonso in Spain (Translator's note).

distributing new gains to them, as well as to explain the erosion of authoritarian systems, since the social groups that benefit selectively from economic growth begin to demand greater political participation. Both the erosion and the continuation of authoritarianism as political phenomena are thus "deduced" as easily from economic growth as from economic decline. When opposing processes are used at the same time to explain opposite results, they can be characterized as "mythological."^[33]

Infrastructural theories of more direct application are no more convincing or operative. This is true of the interpretation of the recent wavering of the military regimes of Latin America and their tendency to move toward opening and institutionalization as the result of the completion of the process of "authoritarian restructuring of capitalism" that produced them.^[34] If we consider that General Pinochet's Chile is the most thoroughgoing example of such a transformation, to the point that we can speak of it as a veritable "capitalist revolution," the recent evolution of the Chilean situation amounts rather to a disproof of that thesis. That regime, like a certain number of its fellow regimes that were endowed with "constitutive" powers, declared that it had "objectives" (metas ) to complete rather than deadlines (plazos ) to meet. However, the initiation of the "seven modernizations" aimed at privatizing and "modernizing" by removing essential sectors of national life from state control^[35] (in order to establish the domination of the market and to change mental attitudes) has not prevented the regime from fixing a calendar for the progressive establishment of an institutionalized juridical system.

While it is evident that the actions and expectations of the different actors, the range of resources at the disposal of the military, their duration in power, and the initial justification of their emergence are factors to be taken into consideration, the continent-wide situation and the internal processes within the military institutions seem to be the most important elements in determining the political changes within military-dominated systems. Two mechanisms that are not exclusive but are rather often complementary or alternative can give the most complete explanation of those transformations. The first, involving the

will and intention of the constituent military actor, goes back to the overall legitimacy of which we have spoken and to the need to avoid or prevent the risks of democracy. The second, which includes many social factors and the particular functioning of the "military parties," emphasizes the problematical and nonprogrammed or indeterminate character of demilitarization that is the result of a whole series of "unintended effects," of chance, or of the mistakes or errors of those involved.

The effect of the hemispheric situation on the expansion and the involvement, as well as the orientation, of the military does not need any lengthy demonstration.^[36] The hemispheric policy of the United States, which has alternated between anti-Communist vigilance and a concern with democratization has produced an almost clockwork rhythm in the phases of autocracy and the waves of demilitarization. This does not affect the particular dynamics of the most autonomous Latin American states, but in their case produces adaptations that are formal and "cosmetic." If the overthrow of President Frondizi in March 1962 was a response to strictly national conflicts that went back at least to 1955, the military who carried out the coup borrowed their justification from the defense perspectives developed by the Pentagon in the framework of the post-Cuban strategic changes. They dressed up their illegitimacy in a legal cloak—putting Vice-President Guido in the presidency—in order to satisfy the need for democratic respectability required by the Alliance for Progress. In this case the contradiction between the Kennedyite civilian reformism and the counterrevolutionary antireformism of hemispheric defense of the Pentagon permitted a dualistic reading of political-military events on the basis of two levels of interpretation.

More recently, the defeat in Bolivia of the coup d'état of 1 November 1979 and the success of the coup of 17 July 1980 were not unrelated to the continental situation and the policy of the United States. Colonel Natusch Busch was compelled to resign after two weeks under pressure from the Carter administration, which was promoting the process of democratization. The other member countries of the Andean Pact, acting as a veritable democratic bloc,^[37] supported Washington in not recogniz-

ing the usurping government. In July 1980 President Carter, at the end of his term and in the midst of an electoral campaign, could do no more than issue a moral condemnation of a determined and brutal military intervention that was counting on his defeat. Indeed, observers have noted that the coup by General García Meza took place immediately after the presidential nomination by the Republican convention of Ronald Reagan—the hope of all the conservative forces in the continent.

In addition, the proper thing for a military regime to do is to demilitarize and legalize its situation—both because of the overall ideology that we have described and also because of the specific characteristics of the military apparatus in relation to power. The internal tensions produced by military participation in government weaken corporate cohesion and thus the defensive capabilities that are precisely the basis for the (temporary) legitimacy of the military takeover; in addition, participation reduces the political resources of the institution that then becomes dangerously "desacralized." Also for the military in power and those who support them, direct and nonlegal military government is neither a necessity nor a good solution. It corresponds rather to a stage, to a moment of political domination. Legalization is the next step. According to a cost-benefit analysis, the military in power act as if they are trying to find a balance between the social cost of the risk of democracy, and the institutional cost of military authoritarianism. This is why institutionalization is only rarely the same thing as the withdrawal of the military from power, and why legalization often does not have full and complete democracy as its goal. The withdrawal of the military is rather a mark of continuity and of the accomplishment of the mission announced at the time of the intervention. The calling of elections, even when they are not limited in an authoritarian manner, does not ipso facto involve the restoration of the establishment of authentic democracy. If we adopt the definition of procedural democracy of Joseph Schumpeter according to which "the people are able to accept or reject the men who are to govern them,"^[38] the postmilitary state is more likely to organize elections that are without surprises or without results. The true holders of power are not affected.

Furthermore, the leaders of conservative military systems and their ideologues and allies openly reject the uncertainties of the democratic game. Their avowed ideal, "protected democracy," involves a search for an absolute guarantee against the risk of the arrival of adversaries of the status quo by legal means to power. One of the ideologues with the most influence over successive military regimes in Argentina wrote after the overthrow of the civilian government in 1976 that the new governments in the Southern Cone were in the process of founding "future democracies on the solid base of order and development" so as to make them more stable.^[39] For their part the hardliners of the Chilean regime aim at the establishment of permanent safeguards against democratic subversion, since, to quote one of them, "One cannot live always on one's guard."^[40] However, the best "protection" for democracy is in fact the utilization—devious, distorted, or directed?—of the procedures of democracy to legimate authoritarianism. The postmilitary state that is truly established and stable, like all lasting authoritarian governments in Latin America, involves a semicompetitive system. This is a system in which political competition, whether open or controlled, remains contained on the periphery of power while those who hold power keep electoral competition far away.^[41] This system gives those who utilize it the advantages of the legitimacy of a representative regime without the risks of "alternation." It is evident that systems of military domination tend to move in the direction of formulas of this kind when they can, and if they have not lost the initiative. The conservatives among the military have no monopoly on it, if we are to believe the experience of Panama under Torrijos, which gradually became an exemplary semicompetitive system.

Brazil under General Figueiredo seems to be moving toward this formula through alternating relaxation and opening. Certainly the convergence of forces favorable to liberalization—from the tolerated political opposition to the industrial bourgeoisie, from the new middle class to the old political class—has played a role, but the system had a choice of methods and timing. The institutionalization program of President Geisel not only tried to break the opposition front by ending the system of two parties but also aimed at making the system independent of

the army. General Figueiredo was chosen by Geisel as his successor in opposition to the military apparatus. The army lost its role as the great elector. Since, as the election results showed, the legitimacy of the military presence was questioned by civil society,^[42] it was no doubt appropriate to provide a legal basis for the system without appealing to the army. Demilitarization without taking risks was also evident in the care taken by General Geisel, Figueiredo's successor, and the "palace group" that surrounded him to separate those in the army who were responsible for government from the military leaders (chefia contra liderança )—to use the distinction of Rizzo de Oliveira)^[43] —in order to establish the hegemony of the bureaucracy over the armed forces and especially to prevent the emergence of political-military leaders with a legitimacy of their own.^[44] This nondemocratic program may get out of the control of the sorcerer's apprentice who establishes it. The "perfect political crime," as an opposition deputy put it, may not be committed. There is a dose relationship between risk and legitimacy. The more uncertainty—and therefore electoral fair play—there is, the more legitimacy. Thus in Brazil the game is not over despite the precautions that have been taken (election law, division of electoral districts, weakening of the opposition). If the opposition parties could unite to defeat the official candidates in the future direct elections of the governors (if they take place), the "system" would have some difficulty in designating a president without taking account of that situation, and in any case the semicompetitive system would be destroyed.^[45] This is all the more true from now on because the tensions within the military structure created by the awakening of civil society may produce some sudden shifts. However, it is unlikely that those in power will take that risk. The well-timed prohibition of electoral coalitions will no doubt be sufficient to prevent this and to assure the "legal" survival of an exhausted regime.^[*]

Most of the time the internal processes in the military apparatus determine the timing of the phases of demilitarization and open the way to democratic alternation. Weakness in the military apparatus or a serious corporate conflict discredits the in-

[*] For an analysis of the subsequent transition to civilian rule in Brazil, see the epilogue. (Translator's note.)

stitutionalization program. Then an appeal can be made to the civilians and to democratic approval to resolve the deadlock and to overcome the destabilizing cleavages. We do not want to say that the behavior of other actors does not have a role, or that the completion of the process of demilitarization and institutionalization does not depend on other factors, such as the length of time in power of the noncivilian government, the circumstances in which it was established, and the level of violence that it has applied to the society. However, the return of the military to the barracks is first of all a military problem and it would be paradoxical not to view it from that decisive angle. It is evident that since the exercise of power is more demoralizing for a state institution such as the army than for a party,^[46] economic and social schisms increase the conflicts that divide the institution on nonmilitary questions.

A civilian restoration through elections without conditions or prohibitions is frequently the consequence of a palace revolution that results from a political change within the armed forces. The program that justified the military exercise of power is thus abandoned after a period of uncertainty that can last several years (three years in Argentina after 1970, three years as well in Ecuador after 1976, but five years in Peru from 1975 to 1980) and the military then have only to prepare an orderly and "honorable" withdrawal. The rejection of a single political orientation or of a caudillist attempt was at the origin of such shifts. Thus in Peru and Honduras in 1975 and in Ecuador in 1976, the conservative sector of the army was opposed to the reformist military men in power, resulting respectively in the fall of Velasco Alvarado, of López Arellano, and of Rodríguez Lara. However, a refusal in the name of the corporate functioning of the military in power to give a blank check to the man whom the army has put in the government results in the same consequences. The two causes may sometimes reinforce one another as in Peru. In the name of institutional rotation of those who hold the title of chief executive—similar to that applied in Brazil since 1964 and Argentina after 1976—the Peruvian general staff deposed Velasco Alvarado, who wished to maintain his power beyond the time prescribed by military regulations and tried to acquire personal support. The change of policy of the military

party, which could be explained according to some observers by the economic crisis and the urgent need to negotiate with social forces,^[47] led to the restoration of democracy. In the absence of the resources of a charismatic leader and because it refuses to mobilize the support of a party, a bureaucratic system that lacks support and lacks a program can only retire or collapse. The regime of General Morales Bermudez, which lacked party support or the will to acquire it, still lasted five years—providing an unprecedented example of political levitation, of course, but also illustrating the difficulties of organizing an orderly transfer of power in an internally nonconsensual corporate situation.

In Argentina, after the overthrow of General Onganía who had placed no time limit on his power and claimed to be leaving the army out of government, General Alejandro Lanusse, commander in chief and king maker, for a short period enthroned General Roberto Levingston. Levingston then broke with the policy of economic liberalism of his predecessor without having the means to do so or specifying the aims of his government. The only thing that was left for the general staff to do was to recognize the defeat of the "Argentine Revolution" and to prepare the army's withdrawal. Acute internal divisions and intense social tension did not allow any solution other than the transfer of power to the civilians and the termination of the proscriptions that had both undermined the authority of elected presidents since 1955 and encouraged the development of political violence—or to engage in increased repression that the divisions among the military did not permit.

In such cases everything happens as if the recourse to the civilians, the opening of the democratic game without any guarantees for those who are in power, appears to be the only solution that would make it possible to reestablish the internal front of the armed forces. Faced with the danger of a breakdown and decomposition of the institution, the elections relieve the tensions, and reunify a military apparatus that has been torn by contradictory tendencies. It is not a taste for paradox that leads us to say, parodying the rhetoric of the military, that in such a case civilian intervention puts an end to military dissension. In the absence of a minimal consensus, if not of a program of the

armed forces, the solution is formal demilitarization by means of democracy. However, in order for the tactical retreat to be effective there still must be a minimum agreement on the maintenance of neutrality or else the politicization of the military will produce a Bolivian-style cascade of coups and countercoups. In addition, since the divisions among the military are linked to differences among the civilians, demilitarization can only result if a majority of the political forces are convinced that it is a necessity and the military do not see in the return of the civilians any direct danger or possibility of revenge.

Tomorrow the Military? How to Get Rid of Them?

There are many obstacles to the departure of the military from the public scene, that is from the government, that impede or make impossible the return of freely elected civilians to their business. These derive mostly from a corporate logic. The continuation of the threat that justified the coming of the army to power obviously represents the most frequently occurring reason. An outburst of urban terrorism or an unsubdued center of rural guerrilla activity produces military reactions that are not propitious for a democratic relaxation. The overall invocation of the "danger of Communism" or of "the Marxist cancer" that must be rooted out before returning to the normal functioning of institutions is only viable as long as the specter of subversion has a concrete character for a part of public opinion. The logic of counterrevolution can only be fed by the memory of the revolutionary threat. The memory of the three years of Popular Unity is still the surest foundation of the dictatorship in Chile. However, in Brazil, sixteen years after the overthrow of the populist regime and the establishment of control over the popular forces, the leaders of the "system" that created the Manichean doctrine of "ideological frontiers" have muted that worn-out source of legitimation so that it now has little effectiveness. In Argentina, on the other hand, the chaotic conditions of the government of "Isabelita" and the "subversive aggression" that undermined the values of democratic coexistence gave the

counterterrorist regime that was installed in 1976 a considerable capital of acceptance.

The level of official violence is a decisive factor. A military regime that employs little violence enjoys a great degree of freedom of maneuver. A terrorist government, on the other hand, will see a demand for an accounting on the part of the nation. The violations of human rights and the problem of those who have "disappeared" in the course of the struggle against subversion call for explanations, if not punishment, of those who were responsible when the situation returns to normal. In Argentina the specter of Nuremberg haunts the barracks and explains its present immobility. "Argentina only confesses to God,"^[48] declared the minister of interior of General Videla. The demoralization and defense reflexes of an army involved in the "dirty job" of revolutionary war can lead to the indefinite prolongation of the military in power. It is therefore unlikely that the Argentine military can abandon the game as they did in 1973. This time the stakes are too high for them to leave the initiative to the civilians.^[*] It is a question of the future and of the honor of the institution. In Brazil, despite an adroit amnesty that whitened the "stains" of repression, public revelations and denunciations of the responsibility of the officers for the assassination of members of the opposition produced a very strong reaction on the part of the military ministers in February 1981, who warned against any "revanchist" attempt that might impede the process of opening. "The honor of the military is more important than human rights," said the headline of an opposition weekly.^[49] At the very least, liberalization must deal with this demand.

On this question the strategies of the civilians come into play, but their margin of maneuver is narrow. The search for a compromise, the acceptance of the "law of silence"^[50] imposed by the military, can give the party forces and those of democracy the opportunity to gain ground. Avoiding direct confrontation and dissipating any personal or corporate concerns among the officers most involved in the repression can in a curious

[*] For the transition to civilian rule in Argentina, see the Epilogue. (Translator's note.)

way facilitate a gradual movement toward the rule of law and representative procedures. However, this also means restoring legitimacy through an act of weakness, guaranteeing that there will be no punishment for usurpation—in a word, placing the military apparatus in the role of irresponsible arbiter, therefore demilitarizing the government while maintaining the militarization of the political system. It is the eternal dilemma of the capable and principled—accommodation or intransigence. But there is also a fundamental difference between a transition as a result of concessions and a democratic rupture, and it perhaps has to do in fact with the evolution of the relation of forces.

In combination with the preceding characteristic, the nature and duration of the military government also affect the process of demilitarization. While democracy involves both competitive procedures for the choice of those who will govern and a substratum of freedoms that make that possible on a regular basis, some Latin American military systems eliminate the first, but only place limitations on the second. The restriction of party and trade union freedom, and to some degree of freedom of expression, was not in fact a major characteristic of the Peruvian and Panamanian regimes after 1968, or that of Ecuador between 1972 and 1979. Argentina under Onganía, Levingston, and Lanusse (1966–73) in comparison with other earlier or later regimes experienced a remarkable level of tolerance of the opposition. The continued vitality of civil society no doubt encouraged the different forms of demilitarization of the government undertaken by those regimes.

The duration of noninstitutionalized military power and the corruption that results from absolute power is likely to make political alternatives more improbable. The Bolivian case of an army that is divided into clans in which promotion to the rank of officer is a means of social mobility perhaps explains this phenomenon. It has even been hypothesized that their repeated refusal in 1979–80 to recognize the results of elections that did not guarantee the continuity of the military was based both on the fears of many officers that they would have to explain the origin of their increased wealth to public opinion or to the courts, and on the desire of those in the secondary levels of the army to have a share in the corruption. However, beyond

these psychological and anecdotal explanations,^[51] it is true that in the Bolivian case there are more profound elements relating to the militarization of the whole political system.

In Bolivia, while the defense of an institution that believed that it would be threatened by the return of the civilians and especially by a victory of even the moderate left blocked a transition, this also took place because the army is the area where political struggles are carried out. In that "praetorianized" system civilian political sectors are always involved in military interventions. It is rare that a military group launches a coup without party backing or an alliance with civilian groups. The overlapping of civilians and the military, the permanent interrelation of the two spheres, make the "extrication" of the military and the "civilianization" of the government difficult. In such a militarized system every military uprising has or does not have the public support of civilians who are involved. It seems in Bolivia that "praetorianization" is related to the absence of a political majority in the recent elections. In Argentina, where the army has dominated political life for the last fifty years, the demilitarization of governments still does not change the system or end "praetorian regression." The disengagement of the military and a lasting return to a liberal constitutional model in civil-military relations seems unlikely in the short or medium term.^[*] This situation, which is an expression of a structural crisis and especially of a social blockage, cannot be overcome in Argentina without a profound transformation of national society.^[52] This does not mean that the conduct of the actors is insignificant or inconsequential, but that behavior and tactics are not programmed and are themselves conditioned by social reality that recurrent military interventions both affects and distorts.

A Farewell to Arms?

There is no doubt that it is easier to demilitarize the government than to remove the military from power. The various openings and institutionalizations often represent tactical re-

[*] To understand why this prediction turned out to be false, see Epilogue. (Translator's note.)

treats that permit new interventions after the military apparatus has reconstituted its political instruments. It may also provide a way to guarantee a juridical basis for the continuation of a regime that was established by force. An objection can be made citing examples of successful demilitarizations. Without referring to the examples of Chile and Uruguay, let us examine the model democracies of today that for twenty years have been spared the military storms that have shaken their neighbors. If we examine the relations between the civilians and the military in Mexico, Costa Rica, Venezuela, or Colombia, without thinking of what might happen tomorrow in one or another country, we can still inquire about the methods and stages of civilian preponderance. We should ask ourselves first if these countries have experienced phases of militarization in the past and how they have overcome them.

In fact, only Venezuela and Colombia ended a military dictatorship with the restoration of civilian rule. However, in the Colombian case the brief interlude of General Rojas Pinilla in 1953 had the support of the majority political groups who had called him to power in order to put an end to the violence that was destroying the country.^[53] The rapprochement between the two traditional parties that took place in 1957 sounded the death knell for the military government, just as the disagreement between them had enabled it to be born. In Venezuela, which in 1948 had barely emerged from decades of dictatorship by caudillos, the army removed from the government the civilian reformists that it had put there at the outset; however, the rise to supreme and absolute power of General Pérez Jiménez brought the officers who had lost power close to the democratic opposition. The coup attempts from both extremes that punctuated the presidency of Romulo Betancourt after 1958 underlined the difficulties of civilian supremacy. Nevertheless, the military base of the Acción Democrática party helped to strengthen the democratic regime that was all the more sure because Pérez Jiménez had discredited the intervention of the army in political life.

In Mexico the generals of the revolutionary army were part of the power elite and then of the dominant party. Also, the stabilization of the revolutionary order that was in their collective interest facilitated the absorption of a spontaneous and

predatory military caudillismo. In a way the "generals" had to recognize the civilian power in which they participated in order to assure their political preeminence. Finally, in Costa Rica, which has not had a real military intervention since 1917, the army was abolished in 1948, but even before its legal abolition, the permanent military apparatus was already on the way to disappearing.^[54] Therefore, in that case as well, there was no transition from military domination to civilian preponderance.

Are we saying that the uprooting of militarism only takes place as a result of a miracle and in exceptional historical circumstances? Or, as some observers of the Cuban and Sandinista revolutions think, only "total politicization of the military . . . will exclude any militarization of politics in the future"?^[55] Certainly an army that is the guarantor of the revolutionary process that has created it and is led by political commissars and selects its officers on the basis of nonmilitary qualifications presents little danger for those in power.^[56] The maximization of the power of civilians produces a sort of "subjective control"—to use the phrase of Samuel Huntington—that is very reliable. However, besides the fact that we should not mix governmental types and that we are now no longer in the context of a liberal democracy that is characterized by pluralism and alternation, the elimination of the distinction between the civilians and the military often amounts to the militarization of all of the life of the society. Among the Cuban ruling elite that fusion has taken place, it seems, to the advantage of military concerns. Also, the model of the civic soldier that results from that fusion, according to Jorge Dominguez,^[57] does not avoid a conflict of roles.

If we remain with the capitalist societies of the continent and the solutions arrived at within the framework of pluralism and constitutionalism, it is evident that there is no preestablished scenario for the reconstruction of democracy. Apart from the revolutionary schema that we have just described that is based on the liquidation of the army of the state, there are only limited precedents that can give us the first approaches of a model of demilitarization. Let us note, however, that the armed route to "civilianization" does not always involve systems that reject capitalism over the short or medium term. Civilian su-

premacy in Mexico originated in the dissolution of the army of Porfirio Díaz and its replacement by revolutionary armies that were closely linked to the future of the new regime. However, the same scheme, when it was applied in Bolivia in a different international context, led to a defeat. The 1952 revolution purged the army to the point of practically annihilating it, but instead of creating a political and lower-class-based army the MNR governments, which were frightened by social agitation and worker militias that they did not control, hastened to reconstitute the classic army with the aid of a United States military mission.^[58] In Bolivia, rather than favoring demilitarization, the fear of the dissolution of the military as an institution became one of the resources of militarism.

The liberalization of military regimes often gives the impression that it is a stratagem. Some sacrifice is made in order to survive. The temporary character of military power is emphasized in order to disarm the opposition. This gives the opposition a difficult choice—whether to accept the loaded dice of the regime, thus assisting it in acquiring legitimacy, or to pull out and paralyze its institutional program. In fact the distinction between a sham election and an opening that is acceptable to the civilian forces is not measured by the degree of competitiveness of the elections. Elections without surprises and won in advance by the government can advance demilitarization, first, by endorsing the system apart from the action of the military, and second, in giving civil rights to the forces of the opposition. The decisive criterion is not the level of competition, but rather that of civil rights and freedoms. The holding of elections that appear to be pluralistic establishes a facade of legality that the authoritarian nature of the government does not change. Acceptance of the game of politics requires a space of freedom that can bring about a "qualitative change." The logic of the two actions is different and the risks are not the same. In the last case if the opening has any content and even if it does not result immediately in a "democratic breakthrough," it seems that the "low profile" tactic of political forces that are moderate—but not moderately democratic—and capable of temporary compromises can be effective. It allows an improvement in the correlation of forces.

In that case the precarious character of the civilian regime that is under intense military surveillance means that the construction of democracy must come before a change in society and that political maneuvering must be limited to allow a political agreement that the politicians will not resort to the military to resolve political conflicts.^[59] This was what was agreed to by the Venezuelan and Colombian parties in the 1950s and it is also what has emerged from the behavior of the parliamentary political forces of the right and the left in Spain since 1976.^[60] On this basis several steps can be envisaged, without prejudging the order. One of them consists in the democratization of the institutions of the state (army, police, courts) and the other, which is almost contradictory to the first, is the creation in a undramatic way of the conditions for the alternation in power. This latter condition is the very expression of true pluralism and constitutes at that time the real "democratic breakthrough."^[61] That long and uncertain road to democracy is based on a wager: one accepts the game proposed by those in power in order to beat them at their own game. To do that it is helpful if the entire political class and the majority of the participating social sectors give a special position to democratic values and procedures and accept the uncertainty of the ballot, while the political and social forces bid a definitive farewell to arms.

We have tried to give a comparative presentation of the many diverse manifestations of military power in what is called Latin America. It seemed to us that to clarify the various methods of militaristic usurpation and the individual kinds of military domination would enrich reflection and give it order. You may reproach the author for his pessimism and even accuse him—why not?—of a shameful inclination in favor of the military. There is no doubt that it is more comfortable to study the positive heroes of history than the villains, and easier to denounce than to understand. But who knows in advance the cunning of history? Nevertheless, while we have shown that exceptional conditions are necessary for democracy to resist the tensions of development or the crisis of capitalism in the Latin American context, we have also recalled rather forcefully that there is no inevitability, whether geopolitical or historical, about the militarization of

the state in the subcontinent. To explain reality by means other than prefabricated schemes does not mean that you consider it to be rational—and still less desirable.

This book has devoted all its attention to the mechanisms of militarization and the process of military hegemony precisely because they are usually passed over in silence, no doubt to avoid giving too much respect to regimes that should be condemned. Similarly, we have concentrated our attention, to the degree that the documentation permitted, on the principal protagonists, the army and the military, who are misunderstood and ignored by those who oppose them.^[62] In lectures and seminars to informed audiences even in Latin America we have been impressed by the fact that all the questions concentrate on knowing the why of militaristic usurpation. The impatience to know the immediate causes of praetorianism seems all the greater because its manifestations have been insufficiently examined and the empirical knowledge of the subject is relatively limited. On the other hand, it is paradoxical to state that in Western Europe, and especially in France, where political science possesses a formidable arsenal of analyses and data on the functioning of the political system and where electoral studies have reached an impressive level of sophistication, no one ever asks why we enjoy a stable pluralist representative system. Nevertheless, the problem of the "invention of democracy," never directly addressed, should merit some consideration, and no doubt would be easier to analyze in France than that of the emergence of military power in the republics which are our concern.

We have also left aside sweeping generalizations to look more closely at the actors, their environment, and their behavior. We have not asked whether the militarization of the state in Latin America is a response to the need for capitalist accumulation in a context of underdevelopment, or whether we are confronted with a universal march toward authoritarianism that would affect our "industrial democracies" in a different way. It is indeed possible to ask if the Latin American military are the local agents of a transnational authoritarianism that, in the view of some, will establish in our countries a "new internal order," or a "soft fascism"^[63] that will place our

freedoms in peril. Are they rather the representatives of a "Neo-Bismarckian" model in which the state will authoritatively allocate resources with a view to accelerating development?^[64] Are the authoritarian systems that we have examined peripheral and local manifestations of the antilibertarian declarations of the Trilateral Commission^[65] or the antiegalitarian lucubrations of the Club de l'Horloge that proclaim the agony of the welfare state and the destruction of political democracy in the countries of the center? Or, on the other hand, do they correspond to a temporary situation of industrial catching-up and accelerated modernization that is preparing the terrain for the future emergence of the "hundred flowers" of a delayed but still promised democracy? This impassioned and passionate debate has not yet begun. In fact whatever paradigm is chosen, the same question remains. Why in either theory in Latin America are the armies the specific institutions that direct development and put it in operation, since neither in yesterday's conservative modernization nor in the planetary neoauthoritarianism of today do the military play a preponderant role?

The reader who is attracted to unity and coherence will no doubt have also noticed that in accounting for change we have emphasized the conditioning structures and the conduct of the actors—the role of organizations and that of their leaders. We know that there is a raging debate between those who hold the two approaches, and our culpable eclecticism demands an explanation that will only underline the importance that we attach to the specificity and the rich diversity of national situations. When the conditions for a democratic "takeoff" are met over a long period in a society, it is evident that the role of the actors in the preservation of stability or the return of a pluralist system is decisive. On the other hand in countries in which structural blockages exist the virtues and abilities of political men are not a factor. What would Romulo Betancourt have done in Argentina, and what would Ricardo Balbín have not done if he had been a Venezuelan? Besides knowing something about what appears to be an exotic phenomenon, there is still therefore the central question of the conditions for democracy. Is it a privilege that belongs to the Atlantic area or to the north-

ern hemisphere, or is it a form of public life that can be universalized? Neither political philosophy nor history has a satisfactory answer to that question. However, that is no reason to ignore it. The answers, if there are any, must involve the dynamic relationship and dialectic interaction between national and international socioeconomic conditions and the capabilities of the actors. May this book be useful in advancing in that direction since, in politics at least, knowledge of oneself comes from the knowledge of others.

October 1981