Although worries about the fragility and death of democracy are probably as old as democracy itself, they have, once again, become pressing and fashionable.Footnote ¹ While not wanting to downplay the dangers of the rise of authoritarianism at home and abroad, in this chapter I try to call into question the familiar story that locates the end of democracy in the breakdown of democratic institutions and their replacement by authoritarian ones. My goal is not to convince you that democracy is more robust than it currently appears, or that there is nothing to worry about, but to offer an alternative approach to thinking about democracy that shifts how we understand what makes democracy fragile and what that tells us about the end of democracy, as well as its futures.Footnote ²

The bulk of the chapter contrasts two pictures of democracy: one that depicts democracy as closed, and one that depicts it as open.Footnote ³ The first picture focuses on democracy as an institutional form that enables collectives to legitimately rule themselves. The second picture starts from the idea of democracy as a social form in which people work out together the rules under which they live together. Shifting from the picture of democracy as closed to the picture of democracy as open changes how we think about the relationship of democracy to its possible end. Exploring that space generates three thoughts about democratic fragility, as suggested by my title. First, from the point of view afforded by the open, social picture, the closed, institutional picture is wrong about what constitutes the death of democracy. Even when democratic institutions are subverted or overthrown, these events need not constitute the death of a democracy. This is not how democracy ends. Second, democracies need not end this way because even when democratic institutions break down (or when they never get fully up and running in the first place), democracy does not end as long as people remain committed to continuing to work out together how to live together. So, the demise of institutions is not how democracy ends. Third, once we begin to think of democracy as a way of living together, we will also see that democratic politics is an activity that is in principle ongoing: it is not the sort of action that can be completed or finished. If we picture democracy as a way of living together, then our work as democratic citizens is never over and done with. Thus, democracy doesn’t end.

Though the questions and circumstances driving this chapter are practical and political, the chapter itself is a work of philosophy. It thus deals primarily with ideas, and how they might be described, fit together, and be contrasted with one another. Nevertheless, my approach to philosophy is broadly pragmatist in the following sense: I do not take myself to be involved in a theoretical or metaphysical investigation into the true nature of democracy. I think of concepts as tools we use to make sense of the world around us. The value of a tool comes in what it allows us to do: here, how it allows us to think about some part of the world or our lives. The concepts with which we think are useful when they illuminate features or possibilities we might otherwise overlook, or when they show their connections to other, seemingly unrelated, ideas or phenomena. Altering the shape of those concepts can thus reveal features of our world that would otherwise remain obscure. My aim in laying out the two pictures and bringing out how they shape our thoughts about the end of democracy is to help us see both where our vision is blinkered if we insist on one picture and what possibilities for action emerge when we think differently. By seeing how familiar thoughts about the end of democracy draw strength and plausibility from the first picture, and seeing how a different picture refocuses our attention, we can begin to see the possibilities hinted at in my title that the first picture obscures.

Democracy as Closed

Democracy, like any social, political, or governmental form, offers a solution to a problem. We can thus begin to describe each picture of democracy by laying out the problem it takes democracy to solve and the particular features that make its picture of democracy a solution to that problem. What I am calling the picture of democracy as closed takes democracy to be a solution to a problem about collective action and decision. In particular, it starts from the question of how a large group of people can make and enact truly collective decisions in ways that give those decisions authority and thus make the actions that follow from them legitimate. Among the many things that make the problem of collective decision-making hard is the problem of dissent. A genuinely collective decision should be one that even those members of the collective who disagree with it can nevertheless regard as legitimately theirs. The picture of democracy as closed offers a solution to this picture by, centrally, describing a set of procedures and rules protected and enacted by institutions that serve the function of legitimate decision-making. At the heart of this picture of democracy, then, are things such as free and fair elections, representative legislative and executive institutions, and the rule of law. Among the features of these institutions, rules, and procedures that make them democratic (apart from enabling collective self-government) is that they treat citizens as both free and equal. Citizens are equal because the procedures for collective decision-making give them (in principle) equal say in the decisions. Citizens are free because, by giving them the capacity to issue authoritative commands to themselves, democratic institutions allow them to be self-governing, which is a form of freedom. It offers a solution to the problem of dissent and disagreement insofar as citizens can accept the authority and legitimacy of the procedure, and thus its results, even if they otherwise disagree with those results.

Starting from this basic outline, a number of familiar features of democratic institutions follow naturally. First, for democratic institutions and procedures to be mechanisms of legitimate collective decision-making, they must be fixed and settled before the decision in question is made. Consider the design of elections in this regard: elections are able to bestow legitimacy on their winners only if, among other requirements, it is not open to officials or others to change, after the fact, how votes are counted or what decision follows from the votes cast. What renders the decisions and actions taken through these procedures democratic is precisely that they result from following these procedures and working within these institutions. This is why violations of election law, whether through voter fraud, ballot tampering, voter suppression, or post-election reinterpretations of what counts as a valid vote, are thought to strike at the heart of the democratic character of a society. But notice that it also lends force to judgments that are dismissive of protests, marches, and other extra-electoral activities in a well-functioning democracy that aim to change policy or demand that duly elected government officials step down. Although such actions can be understood as attempts to change the views of elected officials or the voting public, they are also always the action of a small minority of that public. Since it is only by following established rules and procedures for decision-making that the entire public can make legitimate decisions, acting to change such decisions by other means will appear to be democratically suspect.Footnote ⁴

Of course, the actual procedures and institutions adopted by a given society need not be perfect. So, this picture will accept that democratic procedures and their outcomes can be criticized at any time for being neither free nor fair. But such criticism, on this picture, will only be legitimate if it points out how the procedure and institutions fail to yield legitimate collective decisions and acts of self-governance, and its “proper” use will be to reform how the next election is run, not to “overturn” the results of the previous one.

This is, I hope, a familiar line of thought. It sketches out, for instance, the terrain on which a number of central debates in democratic theory take place: between aggregative and deliberative conceptions of democracy, among various theories of deliberative democracy, and among institutional approaches over the place of representation in democratic institutions. In fact, I suspect that for some readers this characterization of democracy appears not to be a particular picture of democracy at all, but merely a basic description of what democracy is.Footnote ⁵

Note, however, how starting from this basic picture highlights some issues and obscures others. First, it leads us to focus, as we assess the democratic nature and health of a society, primarily on its procedures, laws, and institutions, rather than on the actions of its citizens. We need not take this point too starkly. A focus on institutions need not deny or ignore that the well-functioning of institutions depends on the proper behavior of those who run, maintain, interact with, and inhabit them, just as a focus on the behavior of citizens need not deny or ignore that citizens interact in large part via various institutions. The difference, rather, shows itself in two ways. The first is the order of priority we assign to the well-functioning of institutions in contrast to various good civic behaviors. On this picture, we see the value of good civic behavior as allowing for the properly democratic institutions to continue to function, rather than seeing the value of democratic institutions as enabling and easing certain forms of civic interaction. The second is whether we look to elites and office holders or ordinary citizens as the source of democratic health or fragility. On the picture of democracy as closed, the health of democracy lies primarily with elites and officeholders and, to the extent that the actions of the rest of us matter, insofar as we hold the office of citizen (primarily as voters).Footnote ⁶

A perhaps less obvious but for my purposes more important feature of the picture of democracy as closed is that it supports an attitude of what I will call “gatekeeping.” If the health of a democracy lies in the health of its institutions, rules, and procedures, then those merit protection from forces that might undermine them. On this picture, those forces interfere with proper democratic principles governing who is allowed to participate in collective decision-making and on what grounds. Protecting democracy then involves making sure that the various boundaries laid out by our democratic principles are respected and protected. This follows from the thought that our procedures, rules, and institutions must be fixed ahead of time in order to properly render and legitimately generate truly collective decisions. In addition, breaches of well-drawn boundaries compromise and corrupt the procedures that the boundaries safeguard by allowing those procedures to be hijacked or turned away from their basic purpose of generating legitimate collective decisions of those properly understood to constitute the demos.Footnote ⁷ Setting up those boundaries incorrectly or allowing them to be porous can allow undemocratic elements into or exclude certain legitimate voices from our politics. In either case, we risk threats to democracy. Thus, working within the picture of democracy as closed leads us to understand the work of protecting democracy, keeping it from coming to an end, in terms of defending those boundaries.

The focus on boundaries is not merely a question of geographical borders and immigration, although immigration is one terrain on which this gatekeeping orientation manifests itself. Nor is such a focus merely the position of those who want to keep others out or draw the boundaries narrowly. Many advocates of greater democratic inclusion are also arguing about where the gates and boundaries should go: they just want them further out. They are no less interested in and concerned with patrolling the boundaries once they are properly drawn. The orientation to gatekeeping shows itself not in the wish to draw the boundaries narrowly, but in the thought that the basic questions to be answered in working out a theory of democracy are where to erect those boundaries and how to protect them.

The first and most prominent set of boundaries separates the members of the demos from those outside of it: it determines who gets to participate in the collective decision-making. Debates about this boundary include debates about immigration, but also about the extent of the suffrage within a given territory. Historically, these have included arguments about expanding suffrage to the poor, women, and formerly conquered or enslaved peoples. In political theory these days, a more common debate concerns whether fair principles of inclusion should extend to various resident noncitizens as well as both citizen and noncitizen nonresidents. Thus, debates between advocates of an all-affected or an all-subjected principle for determining the demos are debates about this kind of boundary.Footnote ⁸ Although these recent debates appear to go beyond where and how to draw geographical or demographic boundaries, they nevertheless rely on the same picture. They assume that it is only once we have properly established the membership criteria for the demos, and thus properly drawn the boundaries between those who constitute the demos and those who are outside of it, that the procedures that allow the demos to make collective decisions can be properly legitimate and authoritative. They merely acknowledge that, in an age of mass migration and global interaction, the demos need not form a geographically cohesive set of individuals.

One of the more perverse effects of taking the question about the constitution of the demos as fundamental in these ways is how it shapes discussions in settler colonial states, such as the United States, Canada, Australia, and New Zealand, about how to secure justice for Indigenous peoples. Its model of treating others as free and equal involves their inclusion in the demos, and thus their subjection to the principles and institutions of a given democratic state as citizens. Treating colonized people who wish to maintain their own sovereignty as full members of the colonial demos does not, however, undo the injustice of colonialism. It finalizes it.Footnote ⁹

The second boundary concerns the inputs to the democratic procedures: the types of speech or action that can contribute to the collective decision-making process. Many debates over the proper definition of “reasonable” or over the criteria of public reason aim to settle the proper place of this boundary. Thus, both those who draw those concepts narrowly and those who argue for a more capacious understanding of appropriate methods of civic discourse and action are oriented toward gatekeeping along this boundary.Footnote ¹⁰ In both cases, the underlying assumption is that, in order for democratic procedures and institutions to serve their purpose, we need to work out ahead of time a set of criteria to determine their acceptable inputs, and thus distinguish the inputs that are necessary for the procedure to be democratic from those which would pose a threat.

Finally, there is the boundary that sets the legitimate scope of the outputs of democratic procedure: the scope and limits of democratic authority. Debates about where to locate the boundary between private and public, or the extent of certain basic, fundamental, and inalienable rights, often take this form. In each case, we are working out and trying to institutionally establish the terms under which our collective decision can be legitimate, and where the authority of that process runs out. From this perspective, we can see one role of the individual private rights often described as the liberties of the moderns as placing a gate beyond which democratic action cannot proceed.Footnote ¹¹

It is this orientation toward gatekeeping and boundary-drawing questions that leads me to call this a picture of democracy as “closed.” It also supports the familiar picture about how democracies end: democracies die when they can no longer maintain their boundaries. Depending on which boundary is breached, we get a different form of concern about the fragility or end of democracy. Breaches of the membership boundary are the particular concern of nationalists, who worry that hitherto foreign people may enter the decision-making process and change its results (“undermining our way of life”). Breaches of the input barrier tend to concern institutionalists, who worry that democracies end when prominent agents within democratic institutions (again including citizens qua voters) fail to safeguard them against antidemocratic ideas or behaviors.Footnote ¹² Finally, breaches of the output barrier tend to concern libertarians, who worry about state overreach: democracy ends when the state starts to meddle in the lives of individuals or the market.

A second broad feature of this picture of democracy is the sharp distinction it draws between the civic action of democratic citizens and the background structure of institutions and laws in which those activities take place, and thus also between what might be called constitutional and ordinary politics. The point of basic democratic institutions and laws is not only to identify the class of citizens, but also to enable them to engage in action that counts as legitimately political and thus democratic. My actions count as legitimately political and democratic as long as they are carried out within the established boundaries and via the various institutions and procedures that have been established for that action, since it is through such institutions that my individual action can contribute to legitimate collective decisions and actions. And while those procedures and institutions can be challenged and changed, this picture leads us to hold them fixed in our thoughts when we are thinking about what might be called ordinary democratic politics. The model at work here is of a stable container that is sharply distinguished from what it contains. The actions we take and the speech we engage in within the boundaries of these institutions make a difference to what those institutions do but not to what those institutions are. We do not, on this picture, shape the institutions that contain our ordinary politics (electoral campaigns, legislative debates, regulatory hearings, etc.) through our ordinary politics. On this picture, it is, after all, precisely the ability of democratic institutions to contain our politics that renders our political actions democratic and thus capable of being legitimate.

This sharp distinction is what leads those working with this picture of democracy to think that the health or death of a democracy is to be read from its institutions, laws, and procedures – its constitutional structure – and not in the behavior of its citizens. On this picture, democracies die when their boundaries no longer hold and their institutions collapse or are corrupted and transformed into nondemocratic ones, not when their citizens stop acting like democratic citizens. Note, however, that this means that if we are trying to figure out whether a democracy is healthy or coming to or at an end, our attention will be drawn toward institutional and constitutional features, and not the manner in which we conduct our ordinary political lives.

This has two consequences that I note briefly here but return to when discussing the contrasting picture of democracy. The first is a reformulation of a point I made earlier: on this picture, the death of democracy is primarily an institutional and elite affair. It happens when elections are subverted or ignored by officials, when leaders put themselves above the law, find ways to change the law without following established democratic procedures, or use their authority beyond its established limits, and no one else in authority (including, of course, citizens in their office of voter) rises up to stop them. When these things happen, it is somewhat irrelevant what ordinary citizens do outside of the voting booth. Their main role is through their participation as gatekeepers in prescribed institutional procedures.Footnote ¹³ The second concerns what this picture obscures: it makes it hard to see how the manner in which we conduct ourselves politically as ordinary citizens can itself mark the end of democracy, as well as how it can work to preserve a democracy even as its institutional structure breaks down. As we turn to the picture of democracy as open, I hope to bring into our vision how such actions can change how we think about how democracies don’t end.

Democracy as Open

What I call the picture of democracy as open sees democracy as a solution to a different problem than the picture of democracy as closed. Here, we start with the problem not of collective decision-making but of living together. Specifically, how can a group of people live together under conditions of pluralism in a manner that treats them all as free and equal? The rough democratic solution to this problem is that we can do this if we also work out together the terms on which we live together.

By focusing on several aspects of this formulation, we can see why it generates a different picture of democracy – as open. First, the emphasis is on living and doing things together. I mean to signal here a more robust form of interaction than mere coordination or a procedure to which each has an input. We can begin to see the force of the idea of “acting together” if we contrast it, as I have done elsewhere, with “acting side-by-side.”Footnote ¹⁴ Acting and living side-by-side requires us to coordinate our actions to avoid running into each other or getting into irresolvable conflicts, and thus requires that each be aware of others and what they are doing. But that coordination can be achieved without there being anything that we see as our action by, for instance, a procedure for collective decision-making that pools our individual choices in a fair manner. In contrast, when we act and live together, we undertake a more robust form of sharing, where we not only coordinate our actions but understand those actions as ours, as what we are doing (together) that is not reducible to what each of us does. We act together when we act in a way that is governed by shared norms, rules, or goals that don’t merely coordinate our behavior (lay out what each of us is to do) but make our action intelligible to us as our action (as what we are doing).

This feature of acting and living together generates a particular problem under conditions of pluralism, given that pluralism involves precisely not agreeing about particular values, norms, and meanings. If we are united by a single faith, worldview, or mission, acting together may be psychologically difficult, but it is more or less clear what it would entail. The problem that democracy aims to solve is how to act and live together, given that we are not so united. It does so by giving us a task to do together that turns out to be possible under conditions of pluralism: working out together the terms of our living together.

To genuinely work out together those terms, we need to treat one another as free and equal: we cannot impose those terms on others. And this, in turn, generates a surprising result. My continual acceptance of what we do as done in my name means that I need to always have a way of challenging and criticizing the terms on which we act together. If I am prohibited from raising concerns about or criticisms of what we do, or if these concerns and criticisms are not taken normatively seriously,Footnote ¹⁵ then I am no longer working out with others how we live together or what we do, and so I am no longer interacting democratically with them. But, of course, this also means that if I am not open to hearing and taking seriously the criticisms and concerns of others, then I am not engaging democratically with them. So, on this picture, the activity of working out together the terms on which we live together requires continual openness to criticism, challenge, and contestation. In fact, it is this constant remaining open to criticism, challenge, and contestation that comprises, in large part, the activity of working out together the terms on which we live together. Moreover, since among the things we need to keep open to challenge are the very institutional forms through which we engage in this activity, we cannot preserve or sustain this activity by locking it into a fixed institutional form. Instead, we preserve this openness by sustaining an ever-shifting pluricentric conversation, wherein we engage with different people in different situations and for different purposes, but in which from any of them we can raise challenges to and criticisms of those different people. This, then, is the basic outline of the picture of democracy as open. Rather than being built around a set of fixed, fair procedures, it is modeled on a set of ongoing conversations. And so, preserving the health of a democratic society will not be a matter of patrolling its boundaries, but of widening the scope and enlivening the quality of its various conversations.

In fact, on this picture, establishing and patrolling fixed boundaries will serve to undermine rather than protect the democratic character of our interactions insofar as it cuts off certain avenues of criticism and contestation from democratic legitimacy. Giving up on the gatekeeping function of boundaries also dramatically changes how we think of the demos. Rather than thinking of this as a group whose membership is determined ahead of time and then given a certain status within various institutions, we can think of it as one whose membership is always open: my being a citizen is a matter of whether I engage with others about how to live together in this open fashion.Footnote ¹⁶

That a democracy is not marked by firm boundaries also gives us a way to rethink what democracy might look like in colonial societies. Challenges from Indigenous people to settler colonial societies’ practices of occupation and colonization do not, generally, come in the form of demands to integrate more fully into the colonial society. They more often take the form of wanting the colonizer to withdraw and recognize the sovereignty and dignity of the colonized society to run its own affairs in its own ways and relate to its land in a manner it might not share with the colonial society. Such demands often include insisting on borders and erecting barriers to entry, and so seem to involve a rejection of democratic relationships between settler and Indigenous peoples. However, if we think of these demands and these borders from the perspective of the picture of democracy as open, we can make, hear, and respond to these demands differently. One way to think about how to do so is to use an image from early attempts to work out such relationships between Indigenous peoples in North America and European settlers: we can hear them as a demand that each side paddle its own canoe while acknowledging that we share the same river. That is, we can interact democratically without all sitting in the same canoe (sharing the same institutions) so long as we can continue talking with and listening to one another as we work out where we are vis-à-vis each other.Footnote ¹⁷ Because democracy on this picture need not be contained within and protected by fixed and solid boundaries, we can develop means of democratic interaction that take place across borders. A demand to establish or respect a border, then, need not involve a rejection of continued democratic relations across it.

Because this picture of democracy does not require a fixed set of institutions, rules, and procedures to contain the action of its citizens and render them democratic, it also need not insist on a sharp divide between constitutional and ordinary politics. Among the things we do in the course of democratic living together is working out the terms on which we live together (as well as, as we have seen, who we are). The terms of living together are not something that is, in principle, to be set up, worked out, and nailed down prior to our democratic interaction. These terms also require openness to challenge and contestation from within the activity of living together; the form of the container is shaped by the activity of what it contains. To turn that around, the mere fact that a group of people are challenging the very terms on which they live with others does not put it outside the boundaries of proper civic action. In fact, it is precisely that they are challenging those terms that makes it properly democratic civic action. This means that the democratic quality of our life together is in part a function of how we conduct that life and the ordinary politics that we undertake along the way. We can erode the democratic features of a society by erecting gates and failing to be open to other voices, criticisms, and contestations, and we can revive and bolster it by taking seriously those criticisms and contestations and taking each other normatively seriously.

This blurring of the line between ordinary life and politics and constitutional politics then changes the place of law and other democratic institutions on this picture. Although laws and institutions continue to provide a framework for our interaction, they are also the outcome of that interaction, and it is precisely their being vulnerable to the effects of that interaction that make them democratic, insofar as this vulnerability is what it means for them to be open to contestation and challenge. Because the role of laws and institutions is not only to enable legitimate collective decisions but to provide a framework for and an expression of our mode of living together, they cannot be thought of or justified by their gatekeeping function. Serving such a role would be a sign they were not fully democratic on this picture.

Adopting a picture of democracy as open has several implications for our thinking about democracy (for what we notice, see, and pay attention to) that are important for addressing the questions with which I began. First, the democratic character of society lies not merely in a set of fixed laws and institutions, but in how we live together or fail to, and thus in our ordinary interactions as well. A society with representative institutions in which citizens no longer engage with each other in the project of working out together how to live together, or are no longer invested in that project, is not merely a democratic society burdened with bad or apathetic behavior, but one whose democratic character has frayed. In contrast, a society in which people genuinely work together in an open fashion to determine the terms of how they live together but do so without the traditional institutions of representative democracy is one that displays signs of democratic health. A society in which we are concerned to delineate and enforce various boundaries, to ignore or silence certain voices, or to cease to interact in a way that counts as genuine engagement will, to that extent, be undemocratic, while one where we work to make ourselves intelligible to others and strive to understand them and their criticisms and concerns will be democratic, possibly independently of the form of the institutions in which we take these actions. This means that when we are assessing whether a society is democratic or whether its democratic character is imperiled or at an end, we need to look beyond the health of its formal institutions. Note that the focus of the open picture on civic practices does not deny the importance of institutions. Institutions play a central role in making it possible for groups of people to live together democratically, and some sorts of institutions do this better than others. Some institutions and other large-scale social dynamics obstruct or block attempts to work and live together. Adopting the open picture, then, does not entail an anti-institutional orientation or an exclusive focus on civic practices and virtues. Nevertheless, on the open picture, what constitutes a society as democratic is its civic practices, not its institutions, and so the institutions will be justified to the extent that they help to enable those practices and proper targets of criticisms if they erode or block those processes. So, for instance, it might be more important for state institutions to be trusted and trustworthy than for them to be formally democratic as defined by a set of fixed criteria.

Second, picturing democracy as open in this way shifts us from thinking of the ideal democratic citizen as one who faithfully patrols various boundaries toward one who displays attitudes and practices of hospitality, inclusion, and neighborliness. That is, it suggests that, as citizens, we should be less concerned with which people, behaviors, or topics are a threat to various democratic norms, institutions, or values and how to protect ourselves and our institutions from them. Rather, we should learn to see our democracy as supported and sustained when we strive to be open to everyone’s contribution to how we live together: when we treat others not as outsiders and threats, but as neighbors and potential civic friends. The idea of hospitality I want to invoke here is not one that makes a sharp distinction between residents and guests and works out a special set of norms for the treatment of those who are mere guests, but one which welcomes those who cross various boundaries and treats them not as outsiders at all, but as welcome members of society. That is, it is an attitude which approaches those who might be taken for outsiders and accepts them as full members whose voices, concerns, and needs are taken as seriously as anyone else’s, and which recognizes that each of us is also an outsider and guest to the extent that we are dependent on the hospitality of our neighbors and fellow citizens for our position within the demos. This contrasts with a view that delineates and protects boundaries by placing various burdens and conditions on those who find themselves on the other side of those boundaries before they can be admitted in good standing to democratic processes.

Third, the interactions that constitute our living together democratically on this picture are, in principle, ongoing. The actions that constitute democratic politics on this picture are not undertaken merely to achieve a fixed goal or end point, but are, in principle, such as can be continued indefinitely. Living together democratically, unlike passing this piece of legislation or electing that candidate or winning this argument, is not something we come to the end of even when we complete some particular action. There is thus no end point of democratic action: democracy does not, in this sense, end. However, actions that are in principle ongoing can only continue if the conditions for their continuation are met; these democratic actions are not eternal and their continuation is neither automatic nor guaranteed. Ongoing action must be sustained even as it is carried out. Furthermore, there is no guarantee that, in succeeding at our proximate aims in an ongoing activity, we thereby lay the conditions for continuing on beyond that point. It is thus part of the work of doing such actions well that we attend to and provide for the continued existence of those conditions. While in conversation with you, I can successfully tell a joke or argue a point in a way that nevertheless undermines the conditions which would allow us to keep conversing. Being good at conversing, and not merely telling jokes or making arguments, depends on my also attending to the conditions necessary for us to continue our conversation. Similarly, I can successfully work toward an institutional or legal reform that I regard as improving the justice of my society, but do so in a way that erodes the conditions under which we can continue living together democratically. Moreover, since the activity of working out together the terms of living together is pluricentric, both the mechanisms of sustenance and those of erosion may involve effects on other conversations and interactions. Making the ongoing nature of democratic action visible helps us see the value of acting in ways that are democratically sustaining, which support and sustain the conditions under which we can go on living together, and thus why it might be worth bearing their extra costs.

Finally, if democracy is to be thought of as open in this way, then it cannot have a fixed and settled institutional form. That is, we cannot set out ahead of time the essential institutional features of a democratic society and then ask of any given society or practice whether it conforms to that template. Since being democratic is being open to contestation, it must be that the shape of a democracy can change in response to criticism without it thereby becoming undemocratic. What will mark societies as democratic is not that they conform to a particular range of familiar shapes, but that they display a certain kind of self-preserving activity, a way of going on, and that the shapes they come to both arise out of and make possible the continuation of that form-preserving activity.

We can sum up the points noted here by saying that if we picture democracy as open, then we need to pay attention to the activities that might sustain or undermine the possibility of going on together.Footnote ¹⁸ We cannot assume, as we will if we approach the matter from within the closed picture, that the democratic character of our society inheres entirely in a set of fixed and stable institutions and laws that can persist indefinitely without any further upkeep even if they are also vulnerable to attack and subversion. Rather, on this picture, the lifeblood of democracies is how their citizens interact, and this is something to which they must both continually commit and whose conditions they must continually sustain going forward. This, then, gives us a way to understand how democracy doesn’t end as well as how it does.

Democracy doesn’t end as long as those living together continue to work out together the terms on which they live together, something they do by remaining open and responsive to the challenges and criticisms of the forms that living together takes, and do so in ways that preserve the conditions under which they can continue to do that. Since such activities and such conditions are not entirely dependent on particular institutional forms or policies, democracy need not end when democratic institutions break down or adopt antidemocratic policies and laws. Of course, acting this way can be made easier or harder by various institutions and material conditions, and so institutional break down can be a step on the way toward, and increase the likelihood of, a democracy coming to an end. Nevertheless, from the point of view of this picture of democracy, the end of democracy is not something that can just happen to us; it is something we must do to ourselves. That means, however, that it is also always within our power to forestall the end of democracy or even renew and sustain it. Actions that are in principle ongoing can be restarted even after they have been cut off or wound down if the conditions for their continuation can be regenerated.

On the other hand, it means that democracies do end when citizens stop acting and living together as democratic citizens, when we replace democratic engagement with forms of interaction that lack the features described herein, or when we neglect the conditions that make it possible for us to continue doing so. Democracies can die in this way with all of their institutions, laws, and constitutional structures intact. When that happens, although we can revive our democracy by developing and deploying new democratic habits, there is no one else, and no institution, law, or procedure, that can do it for us. In other words, democracy ends, or doesn’t, with us.