Ariel Armony's work provides a useful corrective to some of the more extravagant claims made on behalf of civil society, associationism, social capital, and “the third sector.” In particular, the crude claim that is sometimes made—namely, that high membership in voluntary associations leads uniquely to a higher quality of democracy—is roundly and justly criticized. “Civil society may or may not lead to democracy,” the author claims, “because what matters is the context in which people associate, not because association is inherently and universally positive for democracy” (p. 2).

In defense of this assertion, Armony introduces evidence from Weimar Germany, the United States, and Argentina, and he shows how organized civil society groups in these contexts have often acted not in support of democracy but in distinctly antidemocratic ways, supporting the rise of the Nazi Party in the first case, putting up barriers to desegregation in the second case, and upholding the activities of a dictatorial state in the third case. Whether civil society groups work for or against democracy depends, the author claims, on the context within which civil society is given birth within any particular country.

Two aspects of context are particularly important for him: the rule of law and socioeconomic inequalities, respectively. These two aspects are related. “Socioeconomic inequalities impinge on the rule of law,” he asserts, “which affects civic engagement and, in turn, its impact on democracy” (p. 4). The implication is that the higher the degree of inequality in any country, the less civil society groups will support ends that are democratic in nature.

Armony turns on its head the usual connection between civil society and the macrocontext. The arrows point in his case from macrocontext to civil society: Context, in particular the extent of inequality in a country, matters for what civil society organizations do. Civil society, in turn, does not matter so much for the macrocontext.

The trouble with this thesis, as with much of the civil society and social capital literature, in general, lies with establishing convincingly the precise nature of the micro–macro links. Proponents of the social capital and civil society hypotheses are not entirely clear about how patterns of social relationship at the microlevel travel upward to affect results at the macronational level.

Armony, who argues the opposite viewpoint, is equally unconvincing about the mediating links. His thesis about socioeconomic inequalities and the rule of law can explain why groups of white Americans organized patently anti-democratic activities in order to forestall desegregation. But it does not explain why Martin Luther King and the Civil Rights movement organized at the same time and in the same context to seek democratic objectives in more democratic ways.

“The capacity of political institutions to process diverse demands and the overall effectiveness of the rule of law shape attitudes in society,” Armony states, asserting a unique link between macrocontext and civil society (p. 207). However, in the same society, experiencing the same set of political institutions, different civil society groups can thrive, simultaneously pursuing both pro-democracy and anti-democracy ends. The Mafia exists in Italy alongside all of those other, pleasanter groups identified by Robert Putnam. White racist groups worked in 1960s America but also Freedom Riders. Mahatma Gandhi and Nelson Mandela organized civil society groups for democracy and against repressive regimes, but other, less democratic and more violent groups also worked at the same time in the same countries.

In reversing the “positive and universal link between civil society and democracy” (p. 200), the author proposes an equally unviable thesis. He is correct to assert that no one-to-one association can run from the micro- to the macrolevel but incorrect to claim that such a one-to-one association does, in fact, run in the opposite direction.

A more relevant question to examine concerns why within the same macrocontext both pro-democracy and anti-democracy organizations are active at the same time. Armony presents some hints in this respect when he refers to diverse channels, links, and interaction mechanisms that variously connect different civil society groups with the state. In any given country, different types of such mediating links can exist—and they can channel civic groups' actions in different directions, strengthening democracy in some cases while undermining it in others. Political parties and other forms of mediating agency can provide these middle-level links between organized citizens and the state, but so can local governments, nonprofit organizations, ethnic associations, and religious congregations. The goals that civil society groups choose for their collective actions depend to a large extent upon the information, access, and interpretations that mediating agency makes available to them.

Neither civil society organizations nor political institutions are individually all-important. What matters in practice is the nature of interactions between them. Focusing on the character of such mediating links helps us get away from making ultimately untenable statements proposing deterministic links of either a top-down or bottom-up kind (e.g., see my Active Social Capital, 2002).

Armony's concern with inequality is refreshing, and his definition of democracy is more encompassing than most. Democracy is, for him, not simply equivalent to some particular set of political institutions; it is as well “a system in which significant segments of the population are [not] de facto excluded from the full benefits of democratic citizenship” (p. 12). He is justifiably concerned that even as democracy has been formally established, in many countries “the capacity of individuals (largely the poor and marginalized) to exercise many of these rights has remained unchanged or has even diminished” (p. 55).

In seeking to rectify this situation, however, Armony is guided by some incautious analysis. He asserts, for example, that the poorer any individual is, the less trusting he or she is likely to be (p. 212). The only evidence he offers in support of this surprising assertion is composed of differences in answers provided by “lower-income” and “higher-income” respondents in 28 countries to two very broad questions about generalized trust. The average difference observed is negative in all countries, which is proof enough to the author that lower-income individuals are less trusting everywhere. But what is this evidence worth exactly? Look more carefully at these numbers, and you find that the opposite conclusion is also supported: The difference in trust between lower-income and higher-income respondents is highest in more equal societies, including Norway and Denmark, and it is lowest in more unequal societies, including Brazil and Mexico. Increasing inequality should lead, if these numbers are to be believed, to spreading trust more evenly within a society.

In general, the cross-national data analysis (in Chapter 5) is not of the same standard as the carefully compiled case studies presented in other empirical chapters. For example, from observing that social trust and inequality are closely associated in regression analysis conducted for 26 countries, the author infers that inequality is a cause of low trust. However, cause and effect could as well go the other way; for example, high social trust could lead citizens to support measures that reduce inequality.

Armony presents an important corrective to expansionist views about civil society and social capital. His reversal of the causal connection is not, however, entirely convincing. To the extent that it helps restore more balance between civil society and state institutions, The Dubious Link is an important book to read.