I am deep into coding focus groups, along with my colleagues at CIRCLE. We have convened working-class, urban youth in several American cities. We listen to audio recordings of their discussions with the software package called NVivo and, in addition to making open-ended notes, we attempt to categorize individuals’ statements into one of several hundred codes that we have constructed.

Often, what you hear is not a belief, a preference, or a principle. It is the sound of someone thinking about and around a topic that he or she may never have considered before. Asked whether voting makes a difference, for example, an individual may give a short monologue that drifts between yes and no and then back again, passing by way of such ideas as “no, but you should do it anyway,” and “yes, but only if other people do it, too.”

This reminds me of Nina Eliasoph’s comments from Avoiding Politics (p. 18):

Research on inner beliefs, ideologies, and values is usually based on surveys, which ask people questions about which they may never have thought, and most likely have never discussed. … The researcher analyzing survey responses must then read political motives and understandings back into the responses, trying to reconstruct the private mental processes the interviewee ‘must have’ undergone to reach a response. That type of research would more aptly be called private opinion research, since it attempts to bypass the social nature of opinions, and tries to wrench the personally embodied, sociable display of opinions away from the opinions themselves. But in everyday life, opinions always come in a form: flippant, ironic, anxious, determined, abstractly distant, earnest, engaged, effortful. And they always come in a context–a bar, a charity group, a family, a picket–that implicitly invites or discourages debate.

That’s why the qualitative research we are doing now is interesting. And yet, there is a different way of thinking about people’s mental states and the relationship to their actions. It turns out (from a study of ethics rather than our topic, politics) that people “have a hard time offering an account of their moral reasoning that contains consistent substantive content.” They are “largely incapable of articulating their moral decision-making process in substantive, propositional terms.” Often, their responses to open-ended questions are rationalizations of what they have done, not reasons that will guide what they do.*

A cynic would conclude that people are just not very reasonable; our principles and reasons do not affect our behavior. But it turns out that individuals answer multiple-choice questions in ways that are consistent with their own responses; distinctive, when compared to other respondents; and strongly predictive of their own behavior. In other words, we are guided by something that’s in our heads, and it differs from person to person, but it is not linguistic or explicit. It is more like an unconscious network of associations. That is why fixed-response or “multiple choice” surveys often predict behavior better than open-ended questions do. They may work better for prediction because an actual decision (such as whether or not to vote) is more like checking a box than explaining a personal philosophy. So answering the forced choices on a survey resembles our ordinary decision-making process.*

Yet I remain interested in people’s explicit, verbalized, public thinking. We ought to give good reasons to justify (or criticize) our own actions. We should be interested in other people’s reasons and their reactions to ours. The act of interpreting the public thoughts of working-class urban youth thus has a moral motivation, even if those reasons are not strongly influential in their own lives. I don’t think that current psychological research precludes the hope that good arguments can change people’s implicit stances or premises, which then affect their behaviors.

In short, we should strive to understand other people’s arguments in case they are right and to decide how to respond effectively if they are not.

*Stephen Vaisey, “Motivation and Justification: A Dual-Process Model of Culture in Action,” The American Journal of Sociology, vol. 114, no 6 (may 2009), pp. 1675-1714.

In Coming of Age as a Poet (Harvard, 2003), Helen Vendler interprets the earliest mature verse of four major poets: Milton, Keats, Eliot, and Plath. She argues that great poets reach maturity when they develop consistent diction and formal styles; favored physical and historical milieux; major symbolic referents; characters or types of characters whom they include in their verse; and some sort of (at least implicit) cosmology. They often retain these combinations to the ends of their careers.

Robert Lowell provides an example (mine, not Vendler’s). From the 1940s until his death, his characteristic milieu is New England–specifically the coastal region from Boston to Nantucket–over the centuries from the Puritan settlement to the present. His diction mimics the diverse voices of that region’s history, from Jonathan Edwards to Irish Catholics, but he brings them into harmony through his own regular rhymes and rhythms. His major symbolic references include gardens, graveyards, wars of aggression, the Book of Revelation, and the cruel ocean. He avoids presenting a literal cosmology, but he describes several worldviews in conflict. Sometimes, the physical and human worlds are cursed or damned and we are estranged from an angry, masculine God. Other times, the world is a garden: organic, fecund, and pervasively feminine. (See my reading of The Indian Killer’s Grave for detail.)

A combination of diction, favored characters, milieux, subjects of interest, value-judgments, and a cosmology could be called a “personality.” I don’t mean that it necessarily results from something internal to the author (a self, soul, or nature-plus-nurture). Personality could be a function of the author’s immediate setting. For instance, if Robert Lowell had been forceably moved from Massachusetts to Mumbai, his verse would have changed. Then again, we often choose our settings or choose not to change them.

A personality is not the same thing as a moral character. We say that people are good or virtuous if they do or say the right things. Their diction and favorite characters seem morally irrelevant. For example, regardless of who was a better poet, Lowell was a better man (in his writing) than T.S. Eliot was, because Eliot’s verse propounded anti-Semitism and other forms of prejudice, whereas Lowell’s is full of sympathy and love.

So we might say that moral character is a matter of holding the right general principles and then acting (which includes speaking and writing) consistently with those principles. Lowell’s abstract, general values included pacifism, anti-racism, and some form of Catholic faith. Eliot’s principles included reactionary Anglicanism and anti-Semitism–as well as more defensible views. The ethical question is: Whose abstract principles were right? That matter can be separated from the issue of aesthetic merit.

I resist this way of thinking about virtue because I believe that it’s a prejudice to presume that abstract and general ideas are foundational, and all concrete opinions, interests, and behaviors should follow from them. One kind of mind does treat general principles as primary and puts a heavy emphasis on being able to derive particular judgments from them. Consistency is a central concern (I am tempted to write, a hobgoblin) for this kind of mind. But others do not organize their thoughts that way, and I would defend their refusal to do so. What moral thinking must be is a network of implications that link various principles, judgments, commitments, and interests. There is no reason to assume that the network must look like an organizational flowchart, with every concrete judgment able to report via a chain of command to more general principles. The hierarchy can be flatter.

To return to Lowell, one way of interpreting his personality would be to try to force it into a structure that flows from the most abstract to the most concrete. Perhaps he believed that there is an omnipotent and good deity who founded the Catholic church when He gave the keys of heaven to Peter. Peter’s successors have rightly propounded doctrines of grace and nature that are anathema to Puritans. Puritans massacred medieval Catholics and Native Americans who loved nature and peace. Therefore, Lowell despises Puritans and admires both medieval Catholics and Wampanoags. In his diction, he mocks Puritans and waxes mournful over their victims. His poetic style follows, via a long chain of entailments, from his metaphysics.

But I think not. It is not even clear to me that Lowell, despite his conversion to Catholicism, even believed in a literal deity. (Letter to Elizabeth Hardwick, April 7, 1959: “I feel very Montaigne-like about faith now. It’s true as a possible vision such as War and Peace or Saint Antony–no more though.”) The point is, literal monotheism did not have to be the basis or ground of all his other opinions, such as his love for and interest in Saint Bernard or his deep ambivalence toward Jonathan Edwards. Those opinions could come first and could reasonably persuade him to join the Catholic Church. By mimicking the diction of specific Puritans in poems like “Mr Edwards and the Spider,” Lowell could form and refine opinions of Puritanism that would then imply attitudes toward other issues, from industrial development to monasticism.

Poets are evidently unusual people, more self-conscious and aesthetically-oriented than most of their peers, and more concerned with language and concrete details than some of us are. As a “sample” of human beings, poets would be biased.

But they are a useful sample because they leave evidence of their mental wrestling. Poetry is a relatively free medium; the author is not constrained by historical records, empirical data, or legal frameworks. Poets say what they want to say (although it need not be what they sincerely believe), and they say it with precision.

I think the testimony of poets at least suffices to show that some admirable people begin with concrete admirations and aversions, forms of speech, milieux and referents, and rely much less on abstract generalizations to reach their moral conclusions. Their personalities and their moral characters are one.

In an essay entitled “Why Last Chapters Disappoint,” David Greenberg lists American books about politics and culture that are famous for their provocative diagnoses of serious problems but that conclude with strangely weak recommendations. These include, in his opinion, Upton Sinclair’s The Jungle (1906), Walter Lippman’s Public Opinion (1922), Daniel Boorstin’s The Image (1961), Allan Bloom’s Closing of the American Mind (1987), Robert Shiller’s Irrational Exuberance (2000), Eric Scholsser’s Fast Food Nation (2001), and Al Gore’s The Assault on Reason (2007). Greenberg asserts that practically every book in this list, “no matter how shrewd or rich its survey of the question at hand, finishes with an obligatory prescription that is utopian, banal, unhelpful or out of tune with the rest of the book.” The partial exceptions are works like Schlosser’s Fast Food Nation that provide fully satisfactory legislative agendas while acknowledging that the most important reforms have no chance of passing in Congress.

The gap between diagnosis and prescription is no accident. Many serious social problems could be solved if everyone chose to behave better: eating less fast food, investing more wisely, using less carbon, or studying the classics. But the readers of a given treatise are too few to make a difference, and even before they begin to read they are better motivated than the rest of the population. Therefore, books that conclude with personal exhortations seem inadequate.

Likewise, some serious social problems could be ameliorated by better legislation. But the readers of any given book are too few to apply sufficient political pressure to obtain the necessary laws. Therefore, books that end with legislative agendas disappoint just as badly.

The failure of books to change the world is not a problem that any single book can solve. But it is a problem that can be addressed, just as we address complex challenges of description, analysis, diagnosis, and interpretation that arise in the social sciences and humanities. Every work of empirical scholarship should contribute to a cumulative research enterprise and a robust debate. Every worthy political book should also contribute to our understanding of how ideas influence the world. That means asking questions such as: “Who will read this book, and what can they do?”

Who reads a book depends, in part, on the structure of the news media and the degree to which the public is already interested in the book’s topic. What readers can do depends, in part, on which organizations and networks are available for them to join and how responsive other institutions are to their groups.

These matters change over time. Consider, for example, a book that did affect democracy, John W. Gardner’s In Common Cause: Citizen Action and How It Works (1972). After diagnosing America’s social problems as the result of corrupt and undemocratic political processes and proposing a series of reforms, such as open-government laws and public financing for campaigns, Gardner encouraged his readers to join the organization Common Cause. He had founded this organization two years earlier by taking out advertisements in leading national newspapers, promising “to build a true ‘citizens” lobby—a lobby concerned not with the advancement of special interests but with the well-being of the nation. … We want public officials to have literally millions of American citizens looking over their shoulders at every move they make.” More than 100,000 readers quickly responded by joining Gardner’s organization and sending money. Common Cause was soon involved in passing the Twenty-Sixth Amendment (which lowered the voting age to 18), the Federal Election Campaign Act, the Freedom of Information Act, and the Ethics in Government Act of 1978. The book In Common Cause was an early part of the organization’s successful outreach efforts.

It helped that Gardner was personally famous and respected before he founded Common Cause. It also helped that a series of election-related scandals, culminating with Watergate, dominated the news between 1972 and 1976, making procedural reforms a high public priority. As a book, In Common Cause was well written, fact-based, and clear about which laws were needed.

But the broader context also helped. Watergate dominated the news because the news business was still monopolized by relatively few television networks, agenda-setting newspapers, and wire services whose professional reporters believed that a campaign-finance story involving the president was important. Everyone who followed the news at all had to follow the Watergate story, regardless of their ideological or partisan backgrounds. In contrast, in 2010, some Americans were appalled by the false but prevalent charge that President Obama’s visit to Indonesia was costing taxpayers $200 million per day. Many other Americans had no idea that this accusation had even been made, so fractured was the news market.

John Gardner was able to reach a generation of joiners who were setting records for organizational membership.* Newspaper reading and joining groups were strongly correlated; and presumably people who read the news and joined groups also displayed relatively deep concern about public issues. Thus it was not surprising that more than 100,000 people should respond to Gardner’s newspaper advertisements about national political reform by joining his new group. By the 2000’s, the rate of newspaper reading had dropped in half, and the rate of group membership was also down significantly. The original membership of Common Cause aged and was never replaced in similar numbers after the 1970s. John Gardner’s strategy fit his time but did not outlive him.

Any analysis of social issues should take account of contextual changes like these. Considering how one’s thought relates to the world means making one’s scholarship “reflexive,” in the particular sense advocated by the Danish political theorist Bent Flyvbjerg. He notes that modern writers frequently distinguish between rationality and power. “The [modern scholarly] ideal prescribes that first we must know about a problem, then we can decide about it. … Power is brought to bear on the problem only after we have made ourselves knowledgeable about it.”** With this ideal in mind, authors write many chapters about social problems, followed by unsatisfactory codas about what should be done. As documents, their books evidently lack the capacity to improve the world. Their rationality is disconnected from power. And, in my experience, the more critical and radical the author is, the more disempowered he or she feels.

Truly “reflexive” writing and politics recognizes that even the facts used in the empirical or descriptive sections of any scholarly work come from institutions that have been shaped by power. For example, in my own writing, I frequently cite historical data about voting and volunteering in the United States. The federal government tracks both variables by fielding the Census Current Population Surveys and funding the American National Election Studies. Various influential individuals and groups have persuaded the government to measure these variables, for the same (somewhat diverse) reasons that they have pressed for changes in voting rules and investments in volunteer service. On the other hand, there are no reliable historical data on the prevalence of public engagement by government agencies. One cannot track the rate at which the police have consulted residents about crime-fighting strategies or the importance of parental voice in schools. That is because no influential groups and networks have successfully advocated for these variables to be measured. Thus the empirical basis of my work is affected by the main problem that I identify in my work: the lack of support for public engagement.

Reflexive scholarship also acknowledges that values motivate all empirical research. Our values–our beliefs about goals and principles–should be influenced and constrained by what we think can work in the world: “ought implies can.” Wise advice comes not from philosophical principles alone, but also from reflection on salient trends in society and successful experiments in the real world. An experiment can be a strong argument for doing more of the same: sometimes, “can implies ought.” If there were no recent successful experiments in civic engagement, my democratic values would be more modest and pessimistic. If recent experiments were more robust and radical than they are, I might adopt more ambitious positions. In short, my values rest on other people’s practical work, even as my goal is to support their work.

Finally, reflexive scholarship should address the question of what readers ought to do. A book is fully satisfactory only if it helps to persuade readers to do what it recommends and if their efforts actually improve the world. In that sense, the book offers a hypothesis that can be proved or disproved by its consequences. No author will be able to foresee clearly what readers will do, because they will contribute their own intelligence, and the situation will change. Nevertheless, the book and its readers can contribute to a cumulative intellectual enterprise that others will then take up and improve.

*In 1974, 80 percent of the “Greatest Generation” (people who had been born between 1925 and 1944) said that they were members of at least one club or organization. Among Baby Boomers at the same time, the rate of group membership was 66.8%. The Greatest Generation continued to belong at similar rates into the 1990s. The Boomers never caught up with them, their best year being 1994, when three quarters reported belonging to some kind of group. In 1974, 6.3% of the Greatest Generation said they were in political clubs. The Boomers have never reached that level: their highest rate of belonging to political clubs was 4.9% in 1989. (General Social Survey data analyzed by me.)

**Bent Flyvbjerg, Making Social Science Matter (Cambridge: Cambridge University Press, 2001), p. 143