SOCIOLOGISTS OF RELIGION have long been perplexed by certain basic definitional questions. What gets considered religious, or even appropriately religious, and what does not? Who gets to decide whether something is a religion? What are the effects of Western scholars going to Japan or South Asia and deciding what they see is actually a “religion” when the word, not to mention the concept, is a foreign translation? What about going to a Greenpeace meeting or a Crossfit gym and saying that’s religion too, because it has a ritual, or a Buddhist meditation center, and saying it’s not a religion, because it doesn’t have a God? Is Scientology a religion or a cult? What about Marxism? Or Mormonism? What about Mormonism 10 years after it was founded? Or what some call the Jesus Movement, a strange apocalypticism that spread across the Roman Empire not long after the death of Augustus?
The stakes get bigger once the state gets involved. Modern governments usually accord “religions” certain privileges they don’t give to clubs or cults. When people ask these questions, they’re usually asking them not only in reference to the role of “religion” (whatever that means) but also to the role of secularism, whose definition is equally complicated. In A Secular Age, the philosopher Charles Taylor famously divided “secularity” into three divisions: the idea of something separate from the divine, the idea that there should be a neutral public life separate from the religious, and the particularly modern idea that a good life could be imagined entirely absent from the supernatural. Taylor goes back hundreds of years in describing how those “moral imaginaries” become possible, and while he nods at the role of institutions and coercion, his critics wish he had spent a bit more of his massive tome showing how material and institutional factors mattered at least as much as the many intellectuals he adroitly cites in their native languages.
One such institution — more or less completely ignored by Taylor — is the poll. How has polling, and survey research more broadly, changed not only how we think about religion but also how religious people themselves conceive of who they are and what they do? In sociologist Robert Wuthnow’s latest book, Inventing American Religion, he asks this very question, arguing that polling got Americans “thinking about religion using numbers and percentages and notions of what is typical and what is atypical,” a radical change from when “faith would have been a matter of conviction grounded in the age-old teachings of particular traditions.”
Wuthnow nods at the historian Sarah Igo’s The Averaged American, which shows how all of this polling and social science radically reshaped Americans politics and Americans’ conception of themselves. There is space here for an important and radical acknowledgment of the feedback loop from social science back into society. Just as scholars of race and gender have come back and formed what they might only have wanted to describe, it seems fair to say the same thing might have happened to American religion.
Wuthnow’s introduction implies he’ll make that very critique, alongside a lament that the many problems with polling show it should not be taken as seriously as Americans seem to take it. While he certainly achieves the second goal, providing a devastating critique of polls through the book, his first and more ambitious project, the one that forms his title, never really materializes. We learn a lot about polling on religion in Wuthnow’s six body chapters, the first five on its history, the sixth drawing on interviews with religious leaders to learn how polls affect their work. Yet we never encounter any arguments as ambitious as the book’s title promises. We learn a lot about measuring American religion, and while there are certainly important arguments to be made about how measuring necessarily bleeds into inventing, we don’t get that story here.
I feel somewhat uncomfortable making these criticisms, not least because Robert Wuthnow has been so generous to me both personally and professionally. (I didn’t go to Princeton for my PhD, but when I lived in the area for a semester he allowed me to join a writing group he facilitated.) Wuthnow is a shockingly prolific writer: Inventing American Religion is his 33 single authored book. That’s not counting the two books he’s co-written, the 11 he’s edited, and countless articles, supervised dissertations, lectures, and presentations. I’ve joked with other sociologists that his productivity is our speed of light, and that the rest of us should measure our output by multiplying the letter w by a number between 0 and 1. Even if we had a book every two years, we would still be at best a .5w.
To be clear, Wuthnow’s new book works quite well as a history of polling on religion in America. I’d recommend it to scholars of American religion or American studies more broadly: there are fascinating stories here about the history of polling, the mixture of polling and religious politics, and the way in which polls helped us understand what American religion really is. We learn all about George Gallup and his 1935 “first national poll about American religion.” Wuthnow follows Gallup and then his son throughout the century, joining them with other prominent pollsters and polling agencies. Along the way, he nods at much of the formative work in the American scientific study of religion, including W. E. B. Du Bois’s 1899 landmark, The Philadelphia Negro: A Social Study, the Jesuit Joe Fichter’s 1951 sociological study of Catholics in New Orleans, Will Herberg’s 1952 argument for an American common faith in Protestant, Catholic, Jew, and the prolific work of priest, novelist, and quantitative sociologist Andrew Greeley.
Polling, Wuthnow argues, not only measured but also helped define the shape of American religion in the 20th century. The book’s fifth chapter, “Pollsters as Pundits,” shows the historical developments that forced discussions of religion into the horse race betting that now constitutes most discussions of politics. Pollsters increasingly went on TV to talk about their findings, and in the 1980s polling on religion was often linked to a newly politicized religious right. “Had there been no culture wars,” Wuthnow writes,
polls that attracted wide publicity would not have portrayed religion as a deeply contentious cultural arena. Polling nevertheless did more than simply register the opinions that would have been there anyway. It added an aura of facticity to the conflicts, putting numbers to them and suggesting that they were characteristic of the general public and not merely the views of activist groups.
Here is where we begin to see something of Wuthnow’s argument about “inventing religion,” though I wish he had developed it more. He argues that religion as described by pollsters and pundits is not only an indication of a culture war: it actually reinscribes and, in some senses, helps to create the battles involved.
The problem, however, is that these moments of “inventing religion” are rarely as fleshed out as the book’s introduction promises. In his chapter of interviews with religious leaders, we learn that polling is of only secondary importance to most religious leaders, and, anyway, they often think what they want to think and pick the poll that will support it. If the story of polls in America is really a story of inventing American religion, I would have expected a much longer and more intricate study of how the individualist and aggregating assumptions of polling have worked their way into theology and religious practice. There’s nothing like that here, and given Wuthnow’s incredibly important early work on just those sorts of historical processes, it’s certainly something he’s capable of doing.
Yet there’s another curious problem with Inventing American Religion, which is Wuthnow’s insistence that the problems of polling are somehow utterly separate from the broader problems of social science. That’s not entirely wrong, of course: there are important differences between academic surveys and polling agencies, and there are even bigger differences between the kinds of qualitative sociology Wuthnow does and the big quantitative projects putting out the “numbers and percentages” he laments in his introduction. However, while polling does have worse problems with response rates and instrument testing than standard academic surveys, these differences may be matters of degree rather than kind, especially considering the care put into certain polls and the sloppiness of certain surveys.
More broadly, some of the epistemic problems of polling that Wuthnow describes in his introduction (a focus on numbers and aggregates over group experience and tradition) are problems common to both surveys and polls, no matter how careful, which makes his efforts to distinguish the two later on in the book somewhat befuddling.
However, even more important than whether polls are really so much worse than academic surveys (and I think most sociologists are willing to acknowledge that, on average, they are), there are other, more troubling questions. How does description come back around and shape what it’s describing, surreptiously folding the watcher’s norms down into the watched? The more sophisticated sociologists of religion point to the implicit normativity in any description, alongside the implicit politics. That’s not wrong, of course, but if the fact/value distinction is not as clean as it might appear, neither is it as porous as its critics might claim. Religious critics of the sociology of religion complain, for example, about sociology’s methodological naturalism: you’d have a hard time finding any sociologist, no matter how personally devout, who would use the supernatural as part of a sociological explanation. It’s hard to deny that choice is as much normative as descriptive: the nature of the “neutrality” circumscribes what’s available as description. You could make a similar argument about gender or race, nationality or poverty. The problem of categories — and the epistemic violence contained within — is a burgeoning industry in cultural theory, postcolonial studies, and anthropology.
Sociology — especially the sociology of religion — is just beginning to catch up, and we’ve mostly considered ourselves inoculated by Weber’s defense of “ideal types”: these categories you’re so upset about just help us simplify the vast complexity of social life, we’ll tell you. Everybody knows that’s not really how things are. But the problem with that argument, as Edward Said famously pointed out, is that scholars’ descriptions can stop being descriptions. Sometimes those descriptions come back in to form the reality they’re describing.
So, for example, if the only way you study religion is by assuming that God had nothing to do with it, then it might not be surprising if the religious people you’re studying, upon reading your work, decide that religion isn’t exactly true. It’s not just that the normative can slip into the descriptive; the descriptive can also work its way back into the normative. Yet the sociology of religion’s implicit normativity can be much less overt than that, and most of the criticisms of the social scientific study of religion have been less about its implicit naturalism than its implicit Protestantism. “Social scientific” here doesn’t have to mean sociology, or even the social sciences per se: it’s more a reference to a scientific approach to social life, and, at least as far back as the Enlightenment, to the concept of religion. That science can be more or less positivist: some want hypotheses and falsifiability, others simply want research questions and the possibility of correction. Yet what they all have in common is an identity formed in opposition to the premises and assumptions of theology. It’s hackneyed to say it all goes back to Descartes, but in a sense it does: take nothing for granted, not even God.
The identity crisis in religious studies has many in the field questioning (or defending) whether something called religion exists at all. The critiques are diverse and multifaceted, but they come down to a few key points. First, religion is a category created in Western Europe out of a Christian (or at least Abrahamic) experience. Second, that category took on a uniquely Protestant valence, emphasizing transcendent theism over immanent supernaturalism and the elite’s well-ordered beliefs over the locals’ mixture of practices and prayers. Third, the category was applied to other groups around the world, shoehorning diverse philosophical and cultural experiences into something within the same category as a thing called Christianity. Fourth, this critique is always about how those in power have control over definitions, and can therefore be applied back to the West as well.
There are many important books that explicitly engage the construction of religion, among them Winnifred Sullivan’s The Impossibility of Religious Freedom, Jason Josephson’s The Invention of Religion in Japan, Talal Asad’s Genealogies of Religion, and, earlier, and perhaps most importantly, Wilfred Cantwell Smith’s The Meaning and End of Religion. Yet it’s worth pointing out that none of these books is by a sociologist. Part of the problem is that the sociology of religion has historically been concerned with many of the same questions as the pollsters Wuthnow describes: the beliefs, practices, and politics of American religion. Recent work in the sociology of religion has become much more interested in religion outside of congregations (“lived religion”), religion outside of the United States, and the spiritual and secular experiences that aren’t quite religion but aren’t quite not religion either.
Wuthnow has been an important intellectual hero to many (including myself) in that change. He has consistently insisted that the sociology of religion is just as capable of developing high theory as any of the more traditionally innovative disciplines, and his work has proved it: The Restructuring of American Religion remains the definitive work on postwar American politics and religion, while Poor Richard’s Principle and Meaning and Moral Order, among many others, provide important theoretical accounts of the structuring of moral life, often using religious cases as data. With Inventing American Religion, he might not have conclusively shown how polls remade American religion, but his broader body of work has laid the groundwork for other sociologists of religion to do it for him.