Religious Competition and Indigenous Mobilization

Indigenous mobilization is measured by the annual number of protest events undertaken by indigenous populations and movements from any of the 883 Mexican indigenous municipalities between 1975 and 2000. To avoid biases associated with a count of protest simply based on the geographic location of events, I distinguish between the site of protest and the place of residence of protesters for purposes of statistical analysis. A protest event is assigned to a municipality every time indigenous villagers or groups from that municipality participate in a protest event, regardless of where the event takes place. Although MII surveys 3,553 protest events, I use as measure of the dependent variable the number of municipalities represented per protest event (N = 6,114). P_it expresses the intensity of indigenous protest in municipality i in year t. This panel includes 22,958 observations of which only 10.3% are nonzero counts.

I use the effective number of religions (ENR) as an indicator of Religious competition. ENR is defined as 1/∑r_i², where r_i represents the proportion of a municipal population that adheres to religion i as reported by the Mexican population censuses of 1970, 1980, 1990, and 2000. Mexican census information groups religious affiliation into six categories: Catholic, Protestant, Jewish, Other, None, and No Response. The Protestant category includes all non-Catholic Christian creeds. In indigenous municipalities, the Jewish category is generally empty, but the latter three categories are unusually high and, in some places, represent up to 30% of responses. Garma (Reference Garma1994) suggests that these residual categories may be a good proxy for la costumbre—a syncretic mix of Catholic and pre-Hispanic indigenous religious practices. I consider la costumbre as a religious affiliation distinct from Catholicism and, in calculating ENR, take the sum of the residual categories as a proxy for la costumbre. In the absence of competition, when membership was not an issue, the Catholic Church learned to “tolerate” costumbristas, but as soon as the Protestant expansion rendered membership relevant, Catholic clergy in competitive dioceses actively sought to win costumbristas back into the church.Footnote ⁷

The average ENR for the entire sample is 1.32. This is clearly a monopoly situation. But at one standard deviation from the mean, the average ENR is 1.65—a semicompetitive situation. Levels of religious competition changed from one decade to the next; while 8.8% of Mexico's indigenous municipalities experienced a semicompetitive situation in 1980, 23.5% had more than 1.65 effective religions by 2000. Religious competition varied also across regions. At one extreme, the average indigenous municipality in Chiapas was already experiencing a semicompetitive situation in 1980 (ENR = 1.68) and was fully competitive by 2000 (ENR = 2.1), whereas at the other extreme, Puebla remained a Catholic monopoly throughout the entire period (ENR = 1.2).

Because the Catholic Church waged a “two-front war” against Protestants and costumbristas, ENR is the most appropriate indicator of religious competition. Nevertheless, to show the independent effect of Catholic–Protestant competition on indigenous mobilization, I test for the difference in the proportion of Catholics and Protestants, Cath.-Prot. difference.

As previously discussed, given external shocks that motivate mobilization, the likelihood of contentious collective action depends on the existence of ideological frames, mobilizing structures, and political opportunities (Tarrow Reference Tarrow1998). Although this article focuses on the effect of religious competition on creating the mobilizing structures and ideological frames for mobilization, I test the religious competition hypothesis along with competing hypotheses about mobilizing structures, as well as control for political, economic, and demographic factors typically associated with social protest.

The literature on indigenous mobilization in Latin America suggests that major economic crises associated with the unraveling of Import Substitution Industrialization policies and the neoliberal policy shift motivated indigenous collective action (Van Cott Reference Van Cott2000; Yashar Reference Yashar2005). I test for the effect of stagflation and currency collapses and create a dummy variable (Economic crises) for the years 1976, 1982, 1983, 1985, 1986, 1987, and 1995. I also test for major Economic policy shifts and create a dummy variable for the years when Mexico entered General Agreement on Tariffs and Trade (GATT; 1986) and North American Free Trade Agreement (NAFTA; 1994) and when the government ended the country's six-decade-long program of land reform and liberalized land tenure (1992).Footnote ⁸

The literature also argues that rural corporatist unions provided the organizational bases for indigenous mobilization. Le Bot (Reference Le Bot1994) and Yashar (Reference Yashar2005) argue that when states lost their financial capability to sustain traditional corporatist loyalties, underfunded corporatist rural unions became the main organizational vehicle for dissident indigenous mobilization. Yashar (Reference Yashar2005) argues that where state corporatism did not destroy indigenous communal autonomy, and served as a façade enabling the survival of indigenous languages, cultures, and political institutions, local peasant indigenous leaders used formerly corporatist organizational structures to engage in independent protest. In her influential account, these communities provided the cultural and ideological frames for the politicization of ethnicity, and unions and churches served as the vehicles for mobilization. I use as an indicator of Corporatism a multiplicative variable that combines turnout and vote share for the Institutional Revolutionary Party (PRI) in municipal elections. My assumption is that municipalities in which all eligible rural indigenous voters participated and all cast their vote in favor of the PRI were cases of successful corporatism, whereas those with low turnout and low PRI support were instances of weak corporatism. I take the proportion of indigenous language speaking population, Indigenous pop. (%), as a proxy of cultural survival (communal autonomy).

Following the political process approach to social movements, the literature suggests that democratization (Van Cott Reference Van Cott and Van Cott1994) and declining levels of repression (Yashar Reference Yashar2005) facilitated indigenous collective dissent. I use the effective number of parties in municipal elections as an indicator of Electoral competition and of the opportunities for protest, and a count of acts of government Repression to assess the effect of state threats on mobilization.Footnote ⁹ Finally, I include four control variables: Poverty (measured as the arithmetic average of the municipal proportion of illiterate population and of households earning less than one minimum wage), Indigenous pop. (ln), a dummy variable for the South (Chiapas, Guerrero, and Oaxaca), and a one-year lag of the dependent variable to control for the cyclical nature of protest.Footnote ¹⁰

I test for random effects (RE) and fixed effects (FE) negative binomial (NB) regression models. NB models are the most appropriate estimation technique because Mexico's cycle of indigenous protest, like most cycles of contention, presents a problem of overdispersion and because protest events are not independent (Cameron and Trivedi Reference Cameron and Trivedi1998). To check robustness and specifically to address the potential for bias due to the large number of zero counts, I distinguish between those municipalities with zero and positive protest counts, and run a logit model with a dichotomous (0, 1) dependent variable, where 1 indicates a municipality with a nonzero count. I subsequently run NB models only with those cases that have positive counts (the number of protest events ranged from 1 to 36).

Table 1 presents regression models of the total frequency of protest by indigenous municipality for the period January 1975–December 2000. For purposes of substantive interpretation, I rely on Model 1. Models 2 and 3 are robustness checks, and Model 4 tests for alternative indicators of religious competition. RE and FE regressions yield similar results. Poverty is the only variable for which the FE results (not significant) are systematically different from RE (significant). Models 2 and 3 are the only ones in which FE and RE results differ for two or three of the control variables. Note, however, that in most cases it is the RE coefficients that remain statistically significant. These differences can be partly explained by the inclusion in the models of a few variables that change very slowly over time (e.g., Indigenous pop. or Poverty), which renders the use of FE less appropriate. It is for this reason that I emphasize cross-sectional variation over temporal change and rely on RE models. I transform coefficients of RE models into incidence rate ratios (IRRs).

TABLE 1.

Negative Binomial Models of Intensity of Indigenous Protest by Mexican Indigenous Municipality, 1975–2000

*** Significant at the .01 level; ** Significant at the .05 level; * Significant at the .1 level. Entry cells are unstandardized coefficients; standard errors are in parentheses. TSCS, time-series-cross-section models; RE, random effects; FE, fixed effects; IRR, incidence rate ratios.

Results across models show that religious competition is a strong predictor of indigenous mobilization. Model 1 suggests that, all else being equal, for every additional effective religion that entered Mexico's indigenous municipalities, indigenous protest would increase by 96.2% (IRR = 1.962). This means that, relative to a Catholic monopoly, protest would nearly double in a municipality experiencing a Christian duopoly (with Protestants controlling one half of the market and Catholics the other half). Another way to understand the substantive effect of religious competition on indigenous mobilization is to compare a situation of perfect Catholic monopoly (ENR = 1) with the average level of religious competition for the whole sample (ENR = 1.32). A shift from 1 to 1.32 (one standard deviation from the mean) would imply a 29.4% increase in the intensity of protest.

As an illustration, consider the Mayan Tzeltal municipality of Chilón in Eastern Chiapas and the Zoque municipality of Ocotepec in Northwestern Chiapas. Between 1970 and 2000, the two municipalities shared many socioeconomic and electoral features but differed notably in one: the level of religious competition. Chilón was one of the most successful U.S. Protestant missionary fields in twentieth-century Mexico and, in 1976, had 1.53 effective religions. During the same year, Ocotepec had 1.27 effective religions. Model 1 predicts that protest in Ocotepec in 1976 would have been 25.5% more intense had the municipality experienced Chilón's level of religious competition. If we take a longitudinal view and analyze both municipalities two decades later, by 1996 Chilón had 2.24 effective religions, but Ocotepec had experienced little change. Model 1 predicts that protest in Ocotepec in 1996 would have been 81.3% more intense had the municipality experienced the same level of religious competition as Chilón did. Over the course of three decades, while Chilón experienced 103 protest events and became an epicenter of protest in Chiapas and a site of recruitment for the Zapatista Army of National Liberation (EZLN), Ocotepec did not partake in any of the multiple cycles of mobilization in Chiapas.

Results across models suggest that economic policy shifts, rather than economic crises, were the main external shocks motivating poor indigenous villages into mobilization. They also show that indigenous protest was more intense in municipalities with relatively weaker corporatist controls and less intense where corporatism remained strong. It is unclear from these results whether weak corporatist peasant unions served as mobilizing structures for independent indigenous mobilization or if the erosion of corporatist controls simply lifted an important control for independent mobilization. What is clear, however, is that more culturally autonomous communities did not serve as mobilizing structures for protest. The negative sign associated with the proportion of indigenous language speakers would in principle cast some doubt on the argument that culturally autonomous communities led the cycle of protest. In fact, a closer inspection of the overall results shows that smaller municipalities with greater proportion of indigenous language speakers and more homogenous religious and political structures were less likely to engage in protest. In contrast, larger and more heterogeneous municipalities with greater linguistic, religious, and political diversity were more likely to mobilize. This does not mean that external agents mobilized otherwise passive indigenous communities; rather, as I show in the qualitative section, it means that the penetration of new religious and political actors into the indigenous world created new local leadership structures, organizations, and ideological frames that facilitated indigenous mobilization.

In sum, statistical results across models show that even if we consider the effect of other potential sources of mobilization (corporatism and culturally autonomous communities), political opportunities (electoral competition) and threats (repression), and the impact of economic structures and fluctuations, religious competition appears to be a strong predictor of indigenous mobilization.Footnote ¹¹ This result is consistent even if we use a binary transformation of the dependent variable (Model 2), test only for positive counts (Model 3, RE), or rely on alternative indicators of religious competition (Model 4).

Religious Competition and Ethnic Claim Making

Mexican indigenous communities and movements did not publicly raise the ethnic banner for the first time until the early 1990s—eighteen years into the cycle of peasant indigenous protest! Although ethnic claims never came to dominate, by 1997 ethnic autonomy rights had become a critical demand of Mexico's indigenous movements, accounting for nearly 40% of indigenous protest claims.

I measure ethnic identification as the number of ethnic claims made in indigenous protest events. Yet, following the same criteria used to measure the number of protest events in the previous section, I do not take the raw number of ethnic demands (N = 462) as a measure of the dependent variable, but the number of municipalities making ethnic claims per protest event (N = 1,142). The most common ethnic claims range from demands for public funds for ethnically differentiated groups (10%), to constitutional recognition of Mexico as a multicultural state (40%), and indigenous peoples’ rights to autonomy (50%). To be sure, this is a different set of events from the ones tested in the previous section. For one thing, this is not a measure of protest events but one of claims. The MII reports on average 1.26 claims per protest event. For another, this indicator only includes ethnic claims (13% of total protest event claims) and excludes economic (29%) and political (58%) demands.

I use ENR as an indicator of Religious competition and alternatively test for the level and Growth rate of religious competition on ethnic claim making. Recall that my theoretical claim was that the Protestant strategy of social service provision to converts would serve to channel competition for souls toward social and redistributive issues, and that the quintessential Protestant strategy of translating Bibles into historically stigmatized indigenous languages would contribute to a linguistic revival. The inability of the Catholic Church to respond to the Protestant strategy of decentralizing ecclesiastic hierarchies would force Catholic authorities to oversupply material and cultural rewards to their parishioners to avoid a defection en masse. In the cultural realm, this would result in the “inculturation” of the Gospel, the rise of Indigenous Theologies, and the mobilization of ethnic claims.

However, the mobilization of ethnicity in the public realm requires more than networks and ideology; it also requires material incentives and political opportunities for “ethnic” minorities to express their renewed ethnic consciousness in the public plazas.

Although some anthropological accounts of the politicization of ethnicity in Latin America still hold to primordial arguments, for the most part the literature on indigenous ethnic mobilization assumes that identities are socially and politically constructed (Eisenstadt Reference Eisenstadt2009). Yet, the process by which ethnic identities become politicized remains an open question. One of the few agreements is that the end of the region's experiences of land reform, the liberalization of land tenure, and the dismantling of traditional agricultural policies based on subsidies and price supports for agricultural producers provided peasant indigenous populations with powerful incentives to give up on class as the master cleavage for mobilization (Trejo Reference Trejo2008; Yashar Reference Yashar2005). At the same time, the emergence of a new international regime of indigenous rights motivated Indian peasants to embrace ethnic claims and identities (Brysk Reference Brysk2000). What remains under dispute, however, is the mobilizing agent that facilitated the politicization of ethnicity. In some accounts, peasant indigenous leaders of underfunded rural corporatist union organizations and indigenous communities that received collective land and succeeded in safeguarding indigenous cultural autonomy served as mobilizing agents for the politicization of ethnicity (Yashar Reference Yashar2005). In others, after the collapse of Communism in the Soviet sphere and the discrediting of class identities, Leftist opposition parties in rural indigenous regions became active promoters of ethnic identities (Van Cott Reference Van Cott and Van Cott1994). In most accounts, the advent of democracy opened a window of opportunity for ethnic mobilization (Van Cott Reference Van Cott and Van Cott1994; Yashar Reference Yashar2005).

I test the religious competition hypothesis against four alternative mobilization vehicles for ethnic claim making: corporatism, communal autonomy, Left electoral opposition presence, and the presence of the EZLN and other rebel groups. I use the same indicator of Corporatism as in Table 1 and continue to use Indigenous pop. (%) as an indicator of cultural autonomy. Because leftist parties were the main opposition to the PRI in rural indigenous regions up to the late 1990s, the effective number of municipal parties captures both dynamics of Electoral competition and leftist presence. I use a three-point scale of Guerrilla activity by municipality drawn from Trejo (Reference Trejo2008). I also include controls for Economic crises and Economic policy shifts as in Table 1 and introduce a dummy variable for the 1992–2000 period, Post-1992 dummy, to distinguish between the years before and after the liberalization of land tenure and the emergence of a new international regime of indigenous rights. Finally, I use a count of government Repression as an indicator of political opportunities and threats, and add economic (Poverty), demographic (Indigenous pop., ln), and geographic (South) controls, as well as a one-year lag of Economic, Ethnic, and Political protest to assess whether ethnic claims grew out of prior ethnic or nonethnic claims.

Table 2 presents negative binomial regression results of the total frequency of ethnic claim making by indigenous municipality for the period 1975–2000. Models 1 and 2 show RE and FE regression results covering the period 1975–2000, and Models 3 and 4 aggregate all ethnic claims made in the 1990s and run cross-sectional regressions. I concentrate on the 1990s for two reasons: (1) the large substantive effect of the Post-1992 dummy on ethnic claim making in Model 2 suggests that this is indeed a critical time period for analysis, and (2) as Figure 1b shows, practically no ethnic claims were made prior to the 1990s. By focusing on this segment of the cycle, Models 3 and 4 partially reduce the potential for bias due to the large number of zero counts. In testing for the 1990s, I introduce one additional variable (not shown), Communal Land (the municipal proportion of nonprivate land plots, 1990), to further assess the mobilizing role of autonomous communities on ethnic claim making. Note that I average all explanatory variables for the 1990–2000 period, the only exception being the protest lags, which result from the aggregation of all protest events in the 1980s. The Rate of growth of religious competition variable includes the period 1975–2000. I continue to rely on RE rather than FE for the reasons previously stated.

TABLE 2.

Negative Binomial Regressions of the Intensity of Ethnic Claim-Making by Mexican Indigenous Municipality, 1975–2000

*** Significant at the .01 level; ** Significant at the .05 level; * Significant at the .1 level. Standard errors in parentheses (Models 3 and 4, robust standard errors). CS, cross-section models. Lagged economic and ethnic protest, Communal Land, and Guerrilla are not shown.

Models 1 to 4 show that the level and growth rate of religious competition are strong predictors of indigenous ethnic claim making. Model 1 shows that for every additional religion in an indigenous municipality's religious marketplace, ethnic claims are likely to increase by 114.5% (IIR = 2.145). Model 4 reports, moreover, that not only levels but also a history of religious competition is an important predictor of ethnic mobilization.Footnote ¹² As an illustration, consider the neighboring Tzotzil municipalities of Teopisca and Venustiano Carranza in Chiapas. Both municipalities shared many economic and demographic factors and had similar histories of intense social mobilization, and electoral competition. One major difference that drew them apart, however, was their histories of religious competition: in the 1990s, the mean level of religious competition in Teopisca was 1.85 and 1.32 in Venustiano Carranza. Another difference was that in the 1990s, 42% of protest events in Teopisca involved ethnic claims but only 18% in Venustiano Carranza. In the absence of competitive pressures from Protestant missionaries, Catholic authorities in Chiapas did not develop similar theological and pastoral practices of “inculturation” of the Gospel in Venustiano Carranza as they did in Teopisca, and hence, peasant indigenous protest turned ethnic in the latter but not in the former.

Results show that economic policy shifts, particularly the liberalization of land tenure in 1992, were major stimuli for ethnic mobilization in impoverished indigenous villages. Yet, ethnicity was not politicized in the more culturally and politically homogenous indigenous municipalities but in the more heterogeneous ones—in larger and more racially mixed municipalities, where religious and partisan competition, as well as political protest and the presence of the EZLN, had been more intense. Results also suggest that corporatism depressed, rather than encouraged, ethnic mobilization. And neither cultural autonomy nor communal structures of land tenure provided the structural context for the politicization of ethnicity.Footnote ¹³

In sum, statistical results across models show that religious competition is a strong predictor of indigenous ethnic mobilization. This result is consistent after controlling for major economic policy shifts in agricultural policies, alternative channels for mobilizing ethnicity, political opportunities and threats, and whether using levels or rates of growth of religious competition.

Three Life Histories

Bishop Ruiz. On a cold, misty Sunday afternoon in January 1960, in a ceremony that resembled the crowning of a medieval prince, Samuel Ruiz was ordained bishop of San Cristóbal de Las Casas (SCLC) in the heavily indigenous highlands of Chiapas. As a young, conservative mestizo Catholic clergyman from the central state of Guanajuato, who had trained at the prestigious Gregorian University in Rome, Ruiz was candid in his first encounter with the press about his plans for the predominantly Catholic diocese: to civilize the Mayan Indians, fight communism, check the secularizing tendencies of the PRI, and fight Protestantism (Benítez Reference Benítez1967). Forty years later, Tzeltal, Tztotzil, Chol, and Tojolabal Indians took center stage in a colorful ceremony that celebrated Ruiz as he entered retirement. In his last sermon, the bishop spoke eloquently on the ethical superiority of Mayan cultures and ways of life, and of the Christian responsibility of the Catholic Church to contribute to the empowerment and liberation of indigenous peoples. Over the course of four decades, Ruiz underwent a dramatic transformation from a socially conservative, paternalistic cleric to the most influential promoter of Indigenous Theologies in Latin America (Meyer Reference Meyer, Ríos and Anaya2000; Ríos Reference Ríos2005). At the same time, the diocese of SCLC became a major institutional promoter of peasant indigenous movements and of indigenous peoples’ rights to autonomy and self-determination. With Bishop Ruiz's full support and sponsorship, these networks came to serve as the organizational infrastructure on which peasant indigenous movements in Chiapas launched the most powerful subnational cycle of protest in Mexico. These organizational resources also became—without Ruiz's consent—a critical source of recruitment to the EZLN.

Cardinal Corripio. After serving as bishop in his northern home state of Tamaulipas for almost a decade, Ernesto Corripio was named Archbishop of Antequera in 1967 in the heavily indigenous southern state of Oaxaca. A traditional and conservative cleric also educated at the Gregorian University, Corripio experienced a personal transformation in Oaxaca, embraced the cause of the Indian poor, developed an incipient Indigenous Theology, and became one of the founders of CEDIPIO, the state's pastoral center for training catechists in progressive religious, social, and cultural practices (Hernández Díaz Reference Hernández Díaz2001; Norget Reference Norget, Cleary and Steigenga2004). However, when Pope Paul VI transferred him to the central state of Puebla in 1976 and then promoted him to Archbishop of Mexico City, Cardinal, and Primate of Mexico, Corripio quickly returned to his traditional conservative identity. During his brief stay in Puebla, he did not develop any Indigenous Theology, and in Mexico City, he did not promote progressive practices. It stands as a great irony that in Puebla, the host state of the 1979 Latin American Bishops’ Conference, in which Latin American Catholic authorities famously confirmed their “preferential option for the poor,” neither Corripio nor any of his successors ever opted for the Indian poor and never encouraged indigenous ethnic mobilization.

Archbishop Castro. A traditionalist cleric from the western state of Michoacán, Manuel Castro trained under the influence of the conservative clergy from his home state and from neighboring Guanajuato, and then attended the Pontifical Pio Latin American College and the Gregorian University in Rome. Pope Paul VI appointed Castro archbishop of the heavily Mayan southeastern state of Yucatán in 1969. A practitioner of Catholic conservative orthodoxy among middle- and upper-class Yucatecos, in the late 1970s Bishop Castro became an enthusiastic supporter of charismatic and spiritual forms of Catholicism among poor urban and rural Mayan populations, at the expense of materialist or culturally driven theological and pastoral practices. Even though he promoted the translation of the Bible to Yucatec Mayan, Castro never embraced Liberation or Indigenous Theologies but fought to suppress them. Economically and socially conservative, Castro encouraged the adoption of heterogeneous practices for the Mayan poor, including charismatic preaching, spiritual healing, and psychological self-improvement. Rather than organize economic cooperatives or indigenous movements, Castro's diocesan team encouraged the Mayan poor to adopt a spiritual approach to everyday problems as part of the statewide Movement of Catholic Charismatic Renewal in the Holy Spirit (Garma Reference Garma1994; Várguez Reference Várguez2008).

In this section, I show that the intensity and nature of religious competition is one of the critical factors in explaining the differences in theological, pastoral, and social practices followed by Bishop Ruiz, Cardinal Corripio, and Archbishop Castro, and that these practices, in turn, contribute strongly to explaining the development (or absence) of the organizational infrastructure and ideological discourses for indigenous mobilization and ethnic claim making. Based on the statistical results, we can situate the case of Bishop Ruiz and SCLC as one of high religious competition and high ethnic mobilization, that of Cardinal Corripio and Puebla as one of low competition and low mobilization, and that of Archbishop Castro and Yucatán as one of low mobilization despite high levels of religious competition.

Bishop Ruiz and Chiapas: Protestant Competition and Indigenous Mobilization

Luther Goes to Indian Mexico. Since the mid-twentieth century, the indigenous regions of Chiapas have been among the most dynamic religious marketplaces in Mexico. Although Pentecostals, Jehovah's Witnesses, and Mormons are among the fastest-growing denominations in Chiapas today, for most of the twentieth century mainline U.S. Protestant churches were the leading challengers to the Catholic monopoly (Meyer Reference Meyer, Ríos and Anaya2000). Presbyterianism once was the state's second largest Christian denomination, and some of the most fundamental transformations that took place in the diocese of SCLC were a direct response to half a century of successful Presbyterian proselytizing.

U.S. mainline Protestant churches began their international missionary activity in the first quarter of the twentieth century. Discouraged from working in Asia by political instability (Gill Reference Gill1998), they opted to concentrate their missionary activity in Latin America. The U.S. Northern Presbyterian Church was charged with proselytizing in southern Mexico (Esponda Reference Esponda1986). Presbyterian missionaries first unsuccessfully targeted Catholics in Mexico's urban centers, but soon shifted their strategy to evangelize the most marginalized and underserved Catholic group, namely, rural indigenous populations. Without much missionary experience, the Presbyterians turned to other churches and nongovernmental organizations (NGOs) with expertise in the evangelization of indigenous peoples, particularly the Reformed Church in America (with Native Americans) and the Summer Institute of Linguistics (SIL; with Guatemalan Mayans).

The SIL was particularly influential in shaping the evangelizing strategies of U.S. Protestant missionaries in Latin America. The SIL is the largest NGO specializing in the study of minority languages in the world today. Founded in 1934 by William Townsend, it was created as a network of U.S. Protestant missionaries and professional linguists who traveled around the world to study minority languages, translate Bibles into minority languages, and establish Protestant missionary fields (Hartch, Reference Hartch2006; Rus and Wasserstrom Reference Rus, Wasserstrom, Hvalkof and Aaby1984). The SIL first flourished in Mexico under the sponsorship and political protection of the Mexican presidents of the day, who actively invited Protestant and SIL missionaries in order to counter the power of the Catholic Church—particularly after the Cristero War between Catholic militias and the Mexican government in the 1920s.

John Kempers, a Reformed missionary from Iowa, led the first U.S. Protestant missionary groups into the state in the 1930s. Protestant missionaries entered Mexico in the midst of the Cristero War and the violent attempt of the Mexican postrevolutionary elites to enforce the anticlerical 1917 constitution. The timing of Kempers’ missionary expansion into the Mexican religious marketplace left important long-term legacies. Because the constitution banned foreign-born clergy from clerical activity, U.S. Protestant missionaries were institutionally limited to training native clergy and setting up predominantly local churches. As the constitution suppressed all civil and political rights of clerics, U.S. missionaries and their native trainees were compelled to adopt an unusually low political profile. Moreover, the strategic partnership established by William Townsend and the SIL with every president from Lázaro Cárdenas (1934–1940) to Carlos Salinas (1988–1994) discouraged Protestant missionaries from becoming promoters of an independent critical civil society as Protestant missionaries did elsewhere.

The pastoral conversion strategy that Kempers and other mainline Protestant missionaries followed in Chiapas and Mexico's indigenous regions rested on a scheme of decentralized evangelization based on the study of the Bible (Esponda Reference Esponda1986). Missionaries first learned and translated Bibles to indigenous languages, then established Bible study and literacy groups, trained a local indigenous clergy, built houses of worship and local health clinics, and established workshops for training agricultural labor. After a decade of missionary work, they would move on to another community. Their work was aimed at creating Protestant native focal points—groups of indigenous ministers and pastors who then became local missionaries and evangelized others of their own ethnolinguistic group.

Young Indian males living in poor, overpopulated areas, facing shortages of land, and largely unattended by Catholic authorities were the first to join Protestantism (Garma Reference Garma1994). On material grounds, conversion to Protestantism entailed attractive economic rewards: access to informal education (literacy) and basic health care, and economic savings that came along with the social norms and lifestyle changes of Protestant converts (e.g., the ban on alcohol consumption and their refusal to contribute to financing the annual festival honoring the municipality's Catholic patron saint). On cultural grounds, the use of native languages and translation of the Bible contributed to creating a new linguistic pride among natives whose languages had been stigmatized as backward “dialects” doomed to disappear. On social grounds, by encouraging the creation of a native clergy, Protestantism introduced a powerful source of status differentiation, simply because Indians had been de facto excluded from the Catholic priesthood.

The Protestant provision of economic, cultural, and ecclesiastic rewards to communities that had been historically marginalized both by the Mexican state and the Catholic Church had dramatic consequences. Between 1940 and 1990, the percentage increase of Protestant membership grew in double and triple digits across several indigenous municipalities. Protestant rates of conversion reached their highest level in Tzeltal and Chol municipalities in eastern and northern Chiapas, where U.S. missionaries had developed the most successful missionary sites; conversions were also sizeable in Tzotzil, Mam, and Tojolabal municipalities, whereas the Zoque region experienced few conversions.

The Catholic Church Reacts. Although initially Bishop Ruiz had deemed communism and the “uncivilized” ways of the indigenous lifestyle as the main challenges to his church, after visiting every single township in the state he realized membership retention was his greatest challenge. Not only was Protestantism rapidly spreading in indigenous regions, but also la costumbre was more entrenched than he had initially realized. Ruiz's personal radical transformation and the reconstitution of his diocese's theological, pastoral, and social practices began here (Ríos Reference Ríos2005).

Ruiz reacted to the Protestant expansion like a gifted entrepreneur. Like his Presbyterian challengers, he favored specialization and, in 1963, persuaded Vatican authorities to let him redraw Chiapas’ ecclesiastic geography so that the state's indigenous population would be served under one diocese (SCLC), whereas the mestizo population would be served under two dioceses (Tuxtla Gutiérrez and Tapachula). By institutional design, Ruiz forced himself to serve the Indians, who represented two thirds of his new diocese and the single most important population segment under Protestant “threat.”

Like his Presbyterian challengers, who had subcontracted SIL and Reformed missionaries for proselytizing, Ruiz subcontracted three religious orders to help him defend the Catholic hegemony and strategically placed them in areas of rapid Protestant growth: Jesuits (who had been initially attracted by Ruiz's predecessor) in the Tzeltal/Chol region; Dominicans in Tzotzil and Tzeltal municipalities; and Marists in Tojolabal, Tzeltal, and Mam municipalities. Ruiz's diocesan team was multinational, including missionaries from Argentina, Belgium, France, Spain, and the U.S. (Meyer Reference Meyer, Ríos and Anaya2000).

Like his Presbyterian challengers, Bishop Ruiz introduced a series of important pastoral transformations to provide material and cultural rewards to Catholic indigenous parishioners in the most religiously contested regions of his diocese. Following the dictums of Vatican II, Ruiz quickly dropped Latin as the liturgical language of his diocese and, following the Presbyterian example, began to learn and encourage his pastoral team members to use indigenous languages. He also initiated the Catholic translation of the Bible to indigenous languages. Following a long tradition of communalism that dates back to the early years of Christianity, and also mirroring the competition, Ruiz encouraged the creation of communal Bible study groups. Thousands of lay workers from these communal study groups were trained as catechists and soon became the leading lay members of the diocesan pastoral team.

Like the Presbyterian missionaries who emphasized social service provision, throughout the 1960s and early 1970s, Ruiz actively encouraged the creation of hundreds of economic and social cooperatives for the production and marketing of agricultural commodities, transportation, education, and health care. Bible study groups served as the organizational basis on which these cooperatives were built, and catechists led this process. Unlike U.S. Protestant missionaries, however, Ruiz did not encourage the institutional decentralization of his diocese's ecclesiastic structures. Although by the 1960s U.S. missionaries had established Tzeltal, Chol, and Tzotzil presbyteries under native control, in the early 1970s there were no equivalent autochthonous Catholic churches. And even though Tzeltal, Chol, or Tzotzil Indians had been ordained Protestant ministers and were part of a rising clerical indigenous elite, communal Catholic indigenous leaders could only aspire to be catechists—lay assistants to the mestizo and white clergy.

Catholic Parishioners Demand a New Religious Contract. A decade after Samuel Ruiz had been ordained bishop of SCLC, catechists and indigenous Catholic loyalists in religiously competitive areas were not entirely satisfied with the bishop's initial wave of ambitious theological and pastoral transformations. Throughout the 1970s, catechists voiced their concerns to the bishop and demanded more radical and credible transformations, threatening otherwise to leave the Catholic Church.

Belgian historian Jan De Vos—a former Jesuit missionary who served in the Bachajón mission in the Tzeltal region in the 1970s—provides a detailed account of the intense negotiation processes that took place between catechists and Bishop Ruiz, as parishioners tried to renegotiate a new “religious contract” with the local Catholic Church. The content of these negotiations was profoundly influenced by the proselytizing strategy and worldly rewards Presbyterian and other mainline Protestant missionaries offered to converts to their creed in such places as Oxchuc and Chilón (see quantitative analysis).

Ecclesiastic decentralization was at the top of the agenda. Catechists sensed that Protestant converts were getting a better deal than they were. Presbyterian converts were not religious instructors (catechists), but ministers (pastors); they were no longer part of the laity, but heads of churches. A Presbytery had been founded as a Tzeltal institution, and it was run by Tzeltal Indians. In a heated debate chronicled by De Vos (Reference De Vos2002, 216), catechists confronted Bishop Ruiz and demanded both full ecclesiastic decentralization and the transformation of the Catholic gospel to the Tzeltal culture: “If the Church does not turn Tzeltal here,” they challenged him, “we don't understand how it can be Catholic at all.”

Material redistribution and the diocese's long-term commitment to the Indian cause were the two other major issues at stake. During the 1970s, catechists became convinced that their material situation could not be dramatically improved in the absence of more radical action beyond the promotion of economic cooperatives for self-help. Thus, they confronted Bishop Ruiz again: “The Church and the gospel have told us how to save our souls, but we don't know how to save our flesh. While we work for the salvation of our souls [. . .] we are victims of hunger, illness, poverty and death” (De Vos Reference De Vos2002, 220–1). In the closing plenary session of an Indian Congress convened in 1974 by Bishop Ruiz and state authorities to celebrate the five hundredth anniversary of the birth of Dominican Friar Bartolomé de Las Casas, a passionate defender of the Indians in the early years of Spanish Colonial rule, a participant hinted at the need to politicize their communal organizations: “Who will be the new Bartolomé [de Las Casas] now? Will there be one? No, brothers, it is us now the ones who must become the new [de Las Casas]; it is our communities coming together that will make us strong” (Morales Reference Morales1992, 262).

Faced with growing threats that community leaders might exercise their exit option and defect to the Presbyterian missions, Ruiz introduced a series of radical reforms to improve the Catholic ecclesiastic and material rewards to loyalists in competitive regions and reassure them that this was not a short-term opportunistic response to avoid a defection en masse, but a genuine long-term commitment to their cause.

On the ecclesiastic front, Ruiz initiated an ambitious program to promote hundreds of catechists to permanent deacons, that is, low-ranking members of the Catholic clergy with some sacramental rights, including the right to celebrate baptisms and marriages. He also began a radical process of “inculturation” of the gospel that entailed the reinterpretation of the Bible through the lens of indigenous peoples’ own cultures and histories and the nominal creation of what Ruiz termed the Tzeltal, Tzotzil, Chol, and Tojolabal “churches.” For all practical purposes, however, permanent deacons only enjoyed limited ecclesiastical authority and, despite the nominal existence of local indigenous churches, decision-making power remained in the hands of the bishop.

If Bishop Ruiz was unable (or unwilling) to match the ecclesiastic rewards dispersed by the competition, he was left with the possibility of responding with major innovations to the material and cultural claims the catechists were making. Unable to tax rich landowners to finance a long-term program of material redistribution to the poor, and with no access to public resources, Ruiz responded to the unified demand expressed at the Indigenous Congress to encourage the politicization of communal associational networks and, from the late 1970s on, became a major promoter of peasant indigenous movements, which first fought for material redistribution (land reform) and later for indigenous rights (ethnic autonomy). Bible study groups and economic cooperatives provided the networks and mobilizing structures for the creation of some of the most powerful peasant and indigenous movements in Chiapas. By reinterpreting the Bible in terms of economic exploitation and cultural alienation of indigenous peoples, both the clergy and the Bible study groups gave rise to Land and Indigenous Theologies. These theological developments contributed in important ways to the creation of a militant opposition consciousness against capitalism and authoritarian rule and for indigenous rights (Norget Reference Norget, Cleary and Steigenga2004) and to the development of a dual public consciousness of Mayans as Indians and peasants (Mattiace Reference Mattiace2003).

Throughout the 1980s and 1990s, some of the most influential peasant indigenous movements in Chiapas emerged in the Tzeltal, Chol, Tzotzil, Mam, and Tojolabal regions, particularly in municipalities facing intense levels of religious competition. These movements became a major source of recruitment for the EZLN in the 1990s and played a major role when the state's leading peasant indigenous movements and the EZLN strategically shifted from class (peasant) to indigenous (ethnic) identities and embraced a radical discourse of rights to autonomy and self-determination for indigenous peoples.

Competition forced Catholic clergy to transform Catholicism in radical ways. Although they did not give up ecclesiastic powers to indigenous communities, they introduced radical social transformations that provided catechists and loyalists in competitive districts with incentives to remain in the Catholic fold. Materially, although Protestants offered converts basic education and health services and economic savings, Catholics received similar benefits but were also organizationally empowered to fight for land redistribution and indigenous rights. Culturally, while the Protestant God had learned to speak Tzeltal, the Catholic God had become Tzeltal. In the social sphere, although Catholic catechists and deacons had not been ordained as regular members of the clergy, as their Protestant peers had, over the course of two decades they became major social leaders in state and national politics, tilting the balance of communal status differentiation in their favor. Finally, any suspicions of opportunistic behavior by the Catholic authorities were dissolved when Bishop Ruiz and his pastoral team became the most radical defenders of indigenous peoples’ human rights as indigenous dissidents confronted state repression and landowners’ violence.

Similar, though less intense dynamics of religious competition led Catholic clergy to become major institutional promoters and sponsors of the mobilizing structures and ideologies on which peasant indigenous movements were partly launched in a number of Mexican indigenous regions, including Oaxaca, Veracruz, Tabasco, and, to a lesser extent, Guerrero, Jalisco, and Hidalgo (Trejo Reference Trejo2008).

Cardinal Corripio in Puebla: No Religious Competition and No Ethnic Mobilization

Puebla has been a Catholic stronghold since the arrival of Catholicism to the New World. Although SIL missionaries penetrated the northern indigenous highlands of the state and succeeded in converting a few Catholic indigenous villagers, the Catholic monopoly has never been seriously challenged in Puebla's indigenous regions (Garma Reference Garma1994). When Corripio was transferred to Puebla in 1976, the resilience of the state's monopolistic structure provided him with no incentives to introduce the incipient progressive indigenous practices he had previously developed in Oaxaca. This monopolistic structure turned Corripio and his successors into conservative traditionalist bishops and provided them with no motives to create Bible study groups, cooperatives, or peasant indigenous movements, or to embrace indigenous cultures. The absence of these mobilizing structures and identities, together with the strong corporatist grip of PRI-affiliated corporatist unions, and the unsuccessful development of leftist parties, explains why no major peasant indigenous movement emerged in Puebla and why ethnicity was never politicized (Trejo Reference Trejo2008).

Archbishop Castro in Yucatán: Pentecostal Competition and Catholic Charismatic Renewal

Despite relatively high levels of religious competition, Catholic Mayan peasants in Yucatán did not renegotiate their own religious contract in ways similar to Mayan peasants in Chiapas. Unlike Bishop Ruiz, who had moved his theological and pastoral practices into the arena of worldly rewards and developed Land and Indigenous Theologies, Castro focused on otherworldly spiritual rewards and favored the development of Catholic charismatic practices. Castro's focus on charismatic renewal and spiritual rewards reflected the practices of the most dynamic competitor in the state's religious marketplace: U.S. Pentecostalism. Although Presbyterian missionaries had been the first non-Catholic Christian church to proselytize in Yucatán, by the time the local Catholic clergy began to respond to competitive challenges in the 1980s, Pentecostalism was viewed as the leading competitor. In a long interview held in the 1980s, Castro and his pastoral team made this point clear (Guzmán Reference Guzmán1989). While in Chiapas mainline Protestants had set the agenda of renegotiation of a new religious contract between Catholic parishioners and clergy centered on worldly rewards, in Yucatán Pentecostals shifted the arena of competition to the spiritual realm. Concerns about charismatic renewal and spirituality outweighed material concerns for land redistribution or indigenous rights. No major independent peasant or indigenous movement emerged in Yucatán, partly because Pentecostalism shifted the arena of competition to spiritual concerns and partly because PRI rural corporatist structures were effective in deterring independent mobilization (Mattiace Reference Mattiace2008) and because no leftist party emerged in the state (Trejo Reference Trejo2008).