Teresa Bejan's Mere Civility is a deeply admirable book: original, persuasive, witty, and eloquent. It is also admirably, bracingly, skeptical, in the best sense: the kind of liberal skepticism that we associate in political theory with Judith Shklar, Bernard Williams, and George Kateb. I do not think there can be much doubt that Bejan means this as a kind of tongue-in-cheek praise of the “obnoxious” (69) Roger Williams, the hero of the book: “By the end of his life he worshipped in a congregation of only two, him and his wife— and he may not have been entirely sure about her” (54). Bejan seems to hold out Williams as something of an exemplar of the willingness to bluntly argue for the truth as one sees it, and to go on arguing.

My chief complaint about the book is its modesty. Bejan's treatment of Williams, Hobbes, and Locke makes clear that, for them and in their era, debates about freedom of speech, religious liberty, and civility reached fundamental questions of political life. But in her introduction, conclusion, and epilogue she connects those debates mainly to much shallower cognates in contemporary political philosophy and political life: about the regulation of hate speech, about niceness in political debate, about trigger warnings on university campuses. What she has to say about these questions is clear-headed and helpful. Here as elsewhere Bejan demonstrates those Williamsian virtues: a considerable willingness to strike out on her own, to confront even the most familiar and comfortable of arguments, and to deflate (as it were) pieties of all sorts.

But this focus on the debates over what we now call civility understates the importance of what she has shown us about the early modern debates. I think she hints at the broader lessons briefly, in what I take to be the most important passage in the book: “Williams's great insight, derived from his experience of founding a tolerant society under conditions lacking precisely the stability and ‘assurance’ modern liberals argue is essential for toleration, was that while social life requires common ground, it requires much less than we think. This is because our judgments of in/civility are inevitably partial—to ourselves and to our sect” (153). This reaches those fundamental questions about political belonging, membership, and community that the early moderns were concerned with, those that Bejan shows Locke wrongly answered with an “emphasis on the need for unifying, affective ‘bonds’ in a tolerant society—of mutual charity, trust, and good will—… [that] recalls nothing more than the bonds of church communion shattered by the Reformation” (139). It calls for a rethinking, not only of rhetoric in democratic debate or of on-campus speech regulation, but of basic themes in political theory then and now.

The idea that political life must rest on some moment of initial harmony or concord, that disagreement is only safe if it is preceded by some moment of more fundamental agreement, is pervasive in political theory and political life. It runs from liberal social contract theory through both civic- and ethnic-nationalist conceptions of the unified democratic “people.” It shapes political theory's ongoing discomfort with contestatory and partisan democracy, and the desire to show somehow that taking part in democratic politics implies and entails a whole host of substantive commitments that we can then pretend everyone agrees on. It is explicit in ideal theory— Rawls uses not only universal compliance with the principles of justice, but also consensus about what they are, as a device in the ideal-theoretic stage of his theory—but also present in, for example, a great deal of constitutional theory about constitutional moments and self-binding. Theorists of many stripes, nervous about whether majorities or officials can be counted on to support what they should, seek theoretical guarantees by imagining ex ante unanimity, as if unanimity now is somehow easier to come by than majorities later. We imagine fellow residents of the same polity as being extended kin, or civic friends, or copartisans of the same theory of justice, or anything else that will allow us to impute a fictive concord that offers some theoretical guarantees.

The guarantees are only ever theoretical, however, and it is a category mistake to take them as underpinning such political peace and stability as we ever have in the world. Bejan argues that we need neither prior religious agreement on fundamentals nor prior discursive agreement on decorum to ground merely civil ongoing religious disagreement. It seems to me that the lesson of Williams's political experiment is even more radical than that, however. He showed the possibility of substantial local peace and ongoing civil argument among those who were barely in agreement that they shared a polity—or even, as between the settlers and the Narragansett, those who did not do so at all. And this was true amid institutional novelty and fragility, not against the background of a well-established state that had an institutional imperative to keep the peace. What Williams managed was a kind of founding, but it was a very cobbled-together founding, very unlike the capital-f Foundings that populate the historical and theoretical imagination from Machiavelli through Plymouth Rock to Philadelphia.

Bejan's crucial insight, offered but sped past in that passage quoted above, is that there is a general reason why theorists (and constitutional framers, and so on) overestimate the amount of ex ante social unity that is needed to underpin ongoing civility. Civil politics is our coexistence in the face of disagreement and difference. But we who imagine the conditions of such coexistence are also among the participants in the disagreements. Even when we are trying to think about coming to terms with our disagreeable and disagreeing neighbors, and so trying to bracket part of our own views, we are still prone to smuggle them in at the level of fundamentals and preconditions.

This introduces a genuine instability into the political-theoretical project. It is not just one view or another—contractarianism or constitutional patriotism or ethnic nationalism—that is likely to fall into this mistake. We are all vulnerable to it. While social life requires common ground, it requires much less than we think, even when we try to build that thought into our thinking. And so (here is that bracing skepticism) we do not necessarily need guarantees in theory in order to find civility in practice, and we will not know where the boundary lies until we try out some politics in practice that looks doomed to fail in theory.