Integrative Democracy : Mary Parker Follett ’ s Integration and Deliberative Democracy

In recent years there has been a resurgence of interest in the work of Mary Parker Follett by scholars of management and public administration, but the acute relevance of Follett’s work to deliberative democracy has yet to be fully appreciated. In her 1918 work The New State Follett articulates a normative political theory that I refer to as Integrative democracy, which can be seen as an alternative formulation of deliberative democracy that is based on an activity that Follett refers to as integration rather than deliberation. In this paper I first present two contemporary challenges faced by deliberative democrats: how deliberation itself ought to be defined, and whether or not deliberation produces epistemic benefits in comparison to non-deliberative voting. I then show how Follett’s theory is able to respond to both of these criticisms. Finally, I discuss how Follett’s theory may need to be extended or modified to deal with challenges highlighted by the recent systemic turn in deliberative theory. Author Biography Dr. Graham Wright is an Associate Research Scientist at the Steinhardt Social Research Institute at Brandeis University

Although widely known as a philosopher and management theorist during her lifetime, Mary Parker Follett's work fell into relative obscurity in the decades following her death (Tonn, 2003, pp. 491-493).More recently there has been a resurgence of interest in her ideas among scholars of management and public administration (e.g., Bartels, 2015;Berman & Van Buren III, 2015;Nelson, 2017;Stout, in press;Stout & Love, 2017).However, Follett's work also has acute, and thus far unappreciated, relevance for the field of deliberative democracy, and may help the field respond to a number of contemporary challenges.
In her 1918 work, The New State, Follett articulates a normative political theory that I refer to as integrative democracy, which is closely connected to contemporary notions of deliberative democracy but posits an activity that Follett refers to as integration, rather than deliberation, as the essential activity of democracy.I argue that that integrative democracy can be seen as an alternative formulation of deliberative democracy, and that there are a number of theoretical benefits to treating integration, as opposed to deliberation, as a basis for democratic legitimacy.At the same time, I also argue the dearth of scholarly engagement with Follett's political writings means that there are areas where additional work may be required to update or modify Follett's theory to address current challenges and criticisms.
In this paper I first present two contemporary theoretical challenges faced by deliberative democrats.I then introduce Follett's (1918) theory of integration and discuss the distinction between integration and modern conceptions of deliberation.I then show how attending to this distinction helps to resolve both of these controversies.Finally, I discuss how Follett's theory may need to be extended or modified to deal with challenges highlighted by the recent systemic turn in deliberative theory.

Two Challenges in Deliberative Democracy
Deliberative democracy has become one of the most popular, and most discussed, theoretical conceptions of democratic governance.Yet despite, or perhaps because of this popularity, it is also the subject of intense debate among democratic theorists and practitioners.Most deliberative democrats would likely agree with Bohman and Rehg's (1997) overarching view that "deliberative democracy refers to the idea the legitimate lawmaking issues from the public deliberation of citizens (p.ix)."There is far less agreement, however, over the particulars of what this general definition means in practice.For example, there is still intense debate over what sorts of activities the term "deliberation" is supposed to include (Steiner, 2008).
An even more complicated question is why we should consider deliberation the foundation of democratic legitimacy.Although deliberative democracy is often justified on intrinsic, deontological grounds (Gutmann & Thompson, 1996;Habermas, 1996), a number of theorists have argued that it also requires an instrumental justification, showing that it tends to produce beneficial outcomes (Kuyper, 2018;Martí, 2006).However, the claim that deliberation produces beneficial outcomes raises a new question, which has perhaps been insufficiently articulated in earlier work, and which has no single answer: beneficial relative to what alternative?
In many debates, deliberation is treated as a model for productive civic engagement.In this context the question of instrumental benefit concerns the extent to which the act of participation in deliberation benefits participants, or society at large, in comparison to alternative modes of political engagement, such as activism (Young, 2001), or simply non-deliberative apathy (Pincock, 2012).At the same time, some of the most contentious debates surrounding deliberation involve its value as a method of democratic decision making (Cohen, 1989;Fearon, 1998).In this context the question of instrumental benefit tends to concern the extent to which decisions that are made through deliberation are, in some sense or another, "better" than those that arise from non-deliberative voting.Although "better" in this context might mean any number of things, including "more legitimate" (or "more likely to be perceived as legitimate"), there has been a longstanding debate-going as far back as Aristotle-over whether decisions produced through deliberation are epistemically superior (i.e., more likely to be correct, relative to those made through voting; Bohman, 2006).Although many have defended such an "epistemic conception" of deliberative democracy (Estlund, 2008;Martí, 2006) others have pushed back, arguing that deliberation does not tend to produce epistemic gains in comparison to voting (Chappell, 2011;Solomon, 2006;Sunstein, 2006).Follett's (1918) conception of integration may have relevance for a great number of these contemporary discussions.In this paper, however, I will focus on just two of these debates, and demonstrate the utility of Follett's conception of integrative democracy by showing how a shift from deliberation to integration might help to resolve each of them.The first debate questions how deliberation itself should be conceptualized, and how it can balance its twin goals of inclusion and effectiveness.The second debate concerns whether, and how, deliberative democracy produces instrumental epistemic benefits relative to non-deliberative voting What Should "Deliberation" Mean?
As mentioned, defining what forms of communication should be considered "deliberative" has been a longstanding challenge for deliberative democracy.Bächtiger, Niemeyer, Neblo, Steenbergen, and Steiner (2010) classify the diversity of opinions on this question in reference to two ideal types that they refer to as Type I and Type II deliberation.Type I deliberation more closely hews to the Habermasian ideal of rational discourse, and requires that arguments adhere to standards of rational justification, rather than appeals to force or naked self-interest.Participants are assumed to be truthful, to not act strategically, and to be willing to update their positions to accommodate the weight of argument.In contrast, what Bächtiger et al. refer to as Type II deliberation includes additional forms of communication and arguments beyond rational justification, such as storytelling, rhetoric and "testifying."Emotional appeals, or even insincerity may be permissible, as long as they are presented in an accessible manner, and the parties are willing to listen and respond to each other's arguments (Thompson, 2008).Bächtiger et al. argue that both of these conceptions of deliberations have a number of empirical and normative "blind spots." A key blind spot of Type I deliberation is that limiting the scope of discourse to rational justification has the potential exclude or marginalize certain groups, and may end up reinforcing pre-existing inequalities in power and access (Lyon, 2013;Sanders, 1997;Young, 2000Young, , 2001)).Such concerns were in fact a key impetus for the development of Type II conceptions of deliberation.More broadly, Type I deliberation is often criticized for insufficient attention to diversity, pluralism, and structural inequalities.Additionally, the rarity with which consensus is achieved in real-world deliberation suggests that Type I deliberation may be too idealized to be feasible in practice.
At the same time, Bächtiger et al. (2010) warns that broader Type II conceptions of deliberation may stretch the concept too far, making it difficult to distinguish deliberation from other forms of communication.It is precisely by forbidding certain modes of discourse (e.g.lying, intimidation, rhetorical manipulation) that deliberation is supposed to produce normative and empirical benefits relative to communication in general.Rolling back these restrictions therefore risks sacrificing the very gains deliberation is supposed to produce.For example, permitting rhetoric and appeals to emotion may allow demagogues to manipulate and deceive participants into agreeing to an epistemically or morally inferior option (Goodin, 1980).Bächtiger et al. (2010) argue that each approach has something to teach the other: Type I scholars should attend more closely to real-world deliberative processes, while Type II scholars ought to investigate how expanded standards of deliberation might produce beneficial outcomes.Yet, a tension remains between the two approaches.On the one hand, the more restrictions we place on which modes of communication count as deliberation, the more we risk excluding or marginalizing citizens who lack the resources to achieve proficiency in the permitted modes.On the other hand, the more we relax the restrictions imposed by Type 1 deliberation, the more we risk sacrificing the very benefits that those restrictions were supposed to provide.

Does Deliberative Democracy Really Produce Better Decisions?
The second debate discussed here concerns the extent to which deliberation is an epistemically superior method of democratic decision-making relative to nondeliberative voting.This claim underlies the epistemic conception of deliberative democracy, which sees the epistemic value of deliberation as grounding a justification of deliberative democracy in general.Yet, it seems clear that although deliberation does have the potential to improve the epistemic quality of democracy by increasing individual competence, it also has the potential to erode epistemic quality by reducing independence and diversity.
To explain why this is the case it is helpful to discuss a general phenomenon, which I will refer to here as error canceling, that tends to be invoked, in some form or another, by epistemic justifications for both deliberative democracy and nondeliberative voting.The basic idea of error canceling is that in a sufficiently large and sufficiently diverse group individual errors will cancel out in aggregation.The collective judgment of such a group will therefore be more accurate than the judgment of even the most competent individual member.This general concept underlies popular conceptions about the "wisdom of crowds" (Surowiecki, 2004) and is supported by a wide variety of mathematical, empirical, and theoretical results.
The most robust mathematical justification of the error canceling power of large and diverse groups is the central limit theorem, which shows how random (i.e., independent) disturbances cancel out when averaged over a large enough sample (Page, 2011, pp. 168-173).The central limit theorem itself underlies much of modern statistics and probability sampling methods.Related mathematical results have been used in the context of democratic and decision theory to argue that, under particular circumstances, large and diverse groups tend to make epistemically better decisions than smaller groups, or individual actors.The most well-known of these is Condorcet's Jury Theorem (CJT), which shows that, assuming a correct outcome exists, the likelihood of a majority of reasonably competent and more-or-less independent voters choosing the correct outcome increases as the size of the voting body increases (Berg, 1993).A similar result is obtained by the Diversity Trumps Ability (DTA) theorem, developed by Hong and Page (2004).This theorem shows that, under certain assumptions, a random sample of problem solvers with diverse perspectives can outperform a similarly sized collection of the best-performing problem solvers.The intuition behind the DTA model is that whenever one agent gets stuck the problem can be passed off to other agents who can try to solve the problem from a different perspective.
Independent of these formal mathematical results, the error canceling process also has a strong intuitive and theoretical justification.There is broad, interdisciplinary agreement that the observations of individual actors are in some way limited or biased by the actors' particular physical, social, psychological, or rhetorical perspective.Such a view is obviously reflected in the post-Kantian philosophical chestnut that "perceptions are theory laden," and in many critical and postmodernist epistemologies, but is also supported by empirical results on motivated reasoning and bounded rationality (Bartels, 2002;Kahn, Peters, Dawson, & Slovic, 2017;Simon, 1972).The intuition behind error-canceling is that because these limitations are tied to each actor's particular perspective they can be made to cancel out by aggregating across different perspectives, because an aggregation of diverse perspectives is not itself situated in any one perspective.A colloquial version of this principle can be found in the parable of the blind men and the elephant.Each man is limited to a single perspective (e.g., near the trunk, the side, or the tail) that limits his available information and biases his observations: the man who feels the trunk believes it is a snake, the man who feels the leg believes it is a tree, etc.Yet, when the men coordinate their observations, these biases cancel out, producing an accurate description of the object.
Error canceling of one form or another provides the foundation for most claims about the epistemic benefits of democratic decision making processes (e.g., Anderson, 2006;Bohman, 2006;Estlund, 2008;List & Goodin, 2001;Myers, 2018;Schwartzberg, 2015).For an epistemic justification of deliberative democracy, the key question is therefore whether the addition of deliberation increases or decreases the potential for error canceling, relative to non-deliberative voting.One way that deliberation might increase error canceling is through information transfer that increases the competence of individual actors (Fearon, 1998;Myers, 2018).From a statistical perspective, information transfer could decrease the magnitude of individual errors, which, all else being equal, will be decreased even further when aggregated.In the context of the CJT, deliberation could therefore increase the average competence of voters, which, net of all other changes, would increase the likelihood that the group will come to the correct choice.Indeed, recent empirical work suggests that, at least under certain conditions, debate can produce substantial epistemic gains in comparison to an aggregation of individual opinions (Navajas, Niella, Garbulsky, Bahrami, & Sigman, 2018).
Deliberation, or at least communication, may also be necessary for error canceling to happen at all.The DTA theorem explicitly assumes that problem-solvers work together, building off of each other's work, rather than merely voting on the best solution.Similarly, as Estlund (2008) notes, the parable of the blind men and the elephant plainly relies on the men communicating about their observations; the men would not arrive at the correct answer if they were forced to vote without first pooling their knowledge.
However, deliberation may also decrease the potential for error canceling by reducing the independence and diversity of the group, and it may do this even as it increases overall competence.In statistical parlance, deliberation might increase the correlation of error terms, preventing them from canceling out in aggregation, even if it decreases the total sum of errors.This potential loss of independence has long been seen as precluding the CJT as the basis for an epistemic conception of deliberative democracy (Chappell, 2011;Estlund, 2008), and critics have specifically argued that deliberation weakens the error canceling "wisdom of crowds" because it increases the potential for "groupthink" by reducing independence (Solomon, 2006;Sunstein, 2006;Surowiecki, 2004).
Deliberation may also reduce the diversity of perspectives precisely by producing consensus through the weight of rational augment.In such cases it is precisely the experts-i.e., those who poses the greatest arsenal of rational justifications and therefore the most convincing arguments on a given topic, who would be expected to -and perhaps ought to-prevail.Yet, insofar as experts are able to use rational justification to persuade others to simply adopt the expert's own opinion, rather than pooling their diverse perspectives and working together, the experts have robbed the group of the functional diversity that theorems like DTA require.
Loss of independence and diversity is perhaps the greatest threat to an epistemic conception of deliberative democracy.This threat is most obvious when deliberation allows bad ideas to spread, reducing group competence (e.g., in groupthink), but the examples above show how deliberation could reduce the likelihood of error canceling precisely by rejecting "bad" ideas (i.e., ideas which lack rational justification) and producing consensus on a pre-existing idea from a single perspective.These criticisms cast doubt on the idea that deliberation will tend to improve the quality of decisions, undermining efforts to justify deliberative democracy on epistemic terms.

Follett's Theory of Integrative Democracy
Although deliberative democracy is an intensely modern theory, it shares many fundamental similarities with the theory laid out by Follett (1918) over 100 years ago in The New State, to the point where Follett's theory can be productively analyzed as an alternative formulation of deliberative democracy, albeit with some radical differences.Indeed, scholars have already brought Follett's ideas to bear in current debates about deliberative democracy (Mansbridge et al., 2010).However, scholarship has yet to fully appreciate the key differences between Follett's notion of integration and modern conceptions of deliberation, or how these differences might inform the two challenges discussed above.
Like deliberative democrats today, Follett (1918) argues vociferously against conceptions of democracy based on majority vote, writing that, "democracy means the will of the whole, but the will of the whole is not necessary represented by the majority … nor even by a unanimous vote" (p.142).Instead, Follett argues that democracy should be seen as a participatory group process where individuals with different desires and perspectives comingle and interact through the dialectical activity Follett calls integration.The result of integration represents a creative synthesis of the individual desires of each member of the group, arrived at through dialogue and reflection.For Follett, integration is not merely a method for resolving particular conflicts, but produces a qualitative shift in thinking, with the interpenetrating of understanding achieved during an integration laying the groundwork for future integrations, and more collaborative modes of association in general (Follett, 1924). 1ollett's vision of genuine democratic government begins with integration occurring first at the local level, in what she refers to as "neighborhood groups" (Follett, 1918, pp. 189-257).These small groups of neighbors and community members would meet regularly, not merely when debating specific issues, and always with a goal of integration.Individual neighborhood groups would then choose representatives, who would meet in similar "intermediate groups" at the city level, with the process repeating at the state, national, and potentially even international levels.Follett therefore sees integration at the in-person levelmediated through representation-as the foundation of authentically democratic government.Majority vote, by contrast is a "clumsy makeshift until we devise ways of getting at the genuine collective thought" (Follett, 1918, p. 144).Although Follett herself never used the term, I will refer to her theory as integrative democracy, to emphasize its basis in integration, as opposed to deliberation. 2

Integration
Because of the centrality of integration to Follett's thought, it is essential to discuss the concept in some detail.Follett sees integration as one of three possible methods for resolving conflict.She describes domination as the victory of one side over the other, and compromise as a stalemate where both sides "gives up a little in order to have peace" (Follett, 1942a, p. 31).In contrast, integration involves both sides working together to create a new solution that satisfies the core desires of each.Mansbridge et al. (2010) refer to integrative solutions as "win-win," and although Follett did occasionally describe integration in this way, this characterization fails to emphasize the creative component of integration that Follett sees as its key feature, and what truly distinguishes it from domination and compromise.By definition, an integrative solution is co-created within the conflict itself, through the mutual interpenetration of desires within the group, and could not have been formulated by any one party beforehand.In Follett's typology any method of resolving conflict which does not involve the creation of a new idea would represent either a domination or compromise.
In her other work, Follett outlines a number of procedures that she argues increase the potential for integration, and which also help explicate its nature.For Follett, the first step towards integration is to bring differences into the open so that they can be clearly evaluated by both sides.This requires forgoing tactical obfuscation of one's own position, as well as an analysis of language and symbols to clarify differences and similarities which may be obscured by each side's idiosyncratic use of language.Follett then advocates a breaking up of wholes, whereby larger concepts and problems are disaggregated and dealt with separately.This also involves a separation between the substantive and the dramatic (or symbolic) features of the conflict, each of which may be important, but which often need to be addressed through different mechanisms.The goal is to move the discussion from the initial expressed positions of each side to the underlying desires driving those positions.Exposing these underlying desires can then prompt a revaluation of the original stated positions, which can then intertwine to create a new integrative solution that is orthogonal to the original axis of conflict.
As an example of what integration looks like in practice, Follett (1942a) describes a personal experience in the Harvard library: …in one of the smaller rooms, someone wanted the window open, I wanted it shut.We opened the window in the next room, where no one was sitting.This was not a compromise because there was no curtailing of desire; we both got what we really wanted.For I did not want a closed room, I simply did not want the north wind to blow directly on me; likewise the other occupant did not want that particular window open, he merely wanted more air in the room.( p. 32) To provide a more realistic example of integration in action, Follett discusses a conflict among members of a farming cooperative, who had all signed a five-year contract to exclusively market their crop through the association.When it was discovered that only a third of members were actually adhering to the contract a dispute arose within the cooperative's executive committee about whether to prosecute the offenders.One side argued that failing to prosecute the free-riders would undermine-and eventually bankrupt-the cooperative.The other side argued that the committee would be unable to adjudicate any extenuating circumstances that might affect individual growers, and that mass prosecution would turn the association against itself.In this case the integrative solution was: [for the headquarters office] …not to proceed with prosecution unless the specific cases were handled through a committee of the association located within the local community or the county in which the violators lived.That is, the initiation for prosecution was to come from the local community after it had thoroughly investigated each case.Thus both sides were satisfied: one because the policy of prosecution was to be continued; the other because the responsibility for prosecution was placed in the hands of the local group.(Follett, 1924, p. 159)

Integration versus Deliberation
As Follett defines it, integration has a number of obvious similarities to traditional definitions of deliberation.Although there are, as discussed above, substantial disagreements about what deliberation ought to mean in this context, the term is almost always used to describe a dialectical, non-coercive process whereby a group aims to achieve consensus on some question or issue through a cooperative search for mutual understanding.Here I wish to highlight one additional feature of deliberation that is particularly important in a comparison with integration: the activity of persuasion.
There are, of course, different types of persuasion, and not all of them are "deliberative."3Both Type I and Type II conceptions of deliberation, however, tend to see a particular form of dialectical, non-coercive persuasion as central to the distinction between deliberation and other forms of discourse.As Mansbridge et al.
(2012) put it, "deliberation is about genuine persuasion, not pressure" (p.18) 4 In a deliberation, I successfully persuade someone when I convince them to abandon their own position and adopt mine, not by any coercive means but purely through the "forceless force of the better argument" (Habermas, 1975, p. 108).As a participant in a deliberation I must be willing to let my own mind be changed in the face of superior arguments (or compelling storytelling, etc.) but I am still permitted to enter the deliberation with the expressed goal of winning the debate by persuading the other side to embrace my position, insofar as I honestly believe that it is a superior one.
Like deliberation, integration must be non-coercive, requires participants to afford one another mutual respect, and aims for shared understanding and consensus.As with deliberation, integration is framed in contrast to voting as a method of resolving conflicts and producing democratic legitimacy.Indeed, integration can be productively analyzed as a subcategory of deliberation that also adheres to two additional requirements, which are not generally conceived of as necessary for deliberation in general.
The first of these requirements is that the outcome produced by integration be cocreative, i.e., formulated in the course of the debate itself, rather than by one of the participants beforehand.The usual goal of deliberation is consensus on some alternative or another, but the question of where the chosen alternative came from, and whether it was pre-existing or created during the deliberation itself, is rarely considered.In other words, deliberative theory tends not to distinguish between creative and non-creative outcomes.An integration, by contrast, is only considered successful when it produces an outcome which no individual actor could have formulated on their own.This requirement also clarifies that some form of consensus is a necessary, but not sufficient condition for integration.Although integration does aim to produce a decision through consensus, a group that achieves consensus on a pre-existing, non-creative solution has not achieved integration.
The second requirement of integration is a rejection of persuasion as a central activity.This is a corollary of the requirement that integration produces a creative outcome.If I persuade you to adopt my position and abandon yours then there has been no joint creation, or any creation at all, and there is no longer any opportunity for an integration of our two positions.Successful persuasion destroys the very difference that integration seeks to harness for creative purposes. 5This is true 4 Chambers (2009) likewise distinguishes between "deliberative rhetoric that focuses on engaging, persuading and informing citizens" (p.341) and "plebiscitary rhetoric" that is "concerned first and foremost with gaining support for a proposition and only secondarily with the merits of the arguments or persuasion for that matter (p.337)." 5 Individual instances of persuasion might be instrumentally acceptable in an integration insofar as they help to achieve mutual understanding on a particular point (e.g., that wholes should be broken regardless of what methods were used to achieve persuasion.In Follett's terminology successful persuasion-even persuasion by force of the superior argument-constitutes domination, rather than integration, since it results in the non-creative victory of one side over the other.By the same logic Follett sees even voluntary submission of one side to the other to be counter to the goals of integration, which is neither "to be passive and learn" nor "to push through something we have already decided we want.Each must discover and contribute that which distinguishes him from others, his difference" (Follett, 1918, p. 29).Thus, Follett believes that for integration to be possible, both sides must forfeit the possibility of domination, including domination via persuasion, and commit to a goal of joint creation.
These two restrictions may appear radical and idealistic.Joint creation may be unrealistic when positions seem diametrically opposed and animosity is high.A rejection of rational persuasion likewise seems problematic, given its traditional importance in democratic theory (e.g., Ackerman, 1980;Gutmann & Thompson, 2004;Lupia & McCubbins, 1998;Pratkanis & Turner, 1996).However, it is precisely these two restrictions which allows integrative democracy to more effectively respond to the two challenges faced by deliberative democracy discussed above. 6

Integration's Potential for Inclusive and Effective Communication
As described by Bächtiger et al. (2010) above, deliberation faces a tension between its twin goals of inclusion and effectiveness.More restrictive notions of deliberation jeopardize equity and inclusion, while less restrictive notions risk opening the door to demagogic coercion and threaten the prospects of actually achieving mutual understanding and consensus.Although integration also aims for inclusion and mutual understanding it is less subject to these tensions than deliberation due to its additional restrictions.Because it rejects persuasion and demands co-created solutions integration is more sensitive to diversity and inclusion than even Type II deliberation.At the same time, these restrictions also up in a particular way).But insofar as persuasion leads to domination by a single solution it is contrary to the goals of integration as Follett sees them. 6Some similarity can also be seen between integration and Bohmian conceptions of dialogue.Like Follett, David Bohm rejects the activity of persuasion and sees the goal of dialogue as an interpenetration of ideas, rather than the victory of one side over the other.However, Bohm (1996) also argues that it is "crucial" that "[i]In the dialogue group we are not going to decide what to do about anything" (p.17) and that "we are not trying to change anything, but just being aware of it (p.21)."This is in contrast to integration, which Follett explicitly frames as not merely passive learning, but a method of producing concrete, actionable resolutions to particular conflicts.As such, although Bohmian dialogue might serve as a helpful preparation for integration, it has a different fundamental goal.
prevent coercion and deception.Finally, although integration is clearly an idealized form of discourse, if may actually be more feasible than deliberation in some instances.
In contrast to the most expansive Type II definitions of deliberation, integration's restriction on persuasion and insistence on joint creation clearly distinguish it from "ordinary communication."At the same time, although integration does not require participants to restrict their communication to rational justification it still requires sincerity.In Follett's framework the first task in an integration is to bring differences into the open and jointly remove any linguistic obfuscationsintentional or otherwise.Participants in an integration may tell stories or testify but they must not be insincere with respect to their core desire and positions, or attempt to dishonestly manipulate one another.
At the same time, integration is far more sensitive to inequality and diversity than Type I deliberation.By insisting on creative solutions and forbidding domination through persuasion, integration assures that whatever consensus results will incorporate multiple perspectives, including those which have been marginalized by pre-existing structural inequalities.Integration's categorical rejection of domination by persuasion also obviates concerns that those with greater familiarity with the permitted modes of discourse will illegitimately achieve such domination through ostensibly non-coercive rational persuasion.Indeed, one of the reasons Sanders (1997) argues that "the material prerequisites for deliberation are unequally distributed" is because "some Americans are more likely to be persuasive than others" (p.349).Because it rejects persuasion as a central activity integration is less impacted by such inequalities.
In fact, integration is perhaps even more sensitive to diversity and inclusion that Type II conceptions of deliberation.Although Type II deliberation aims to ensure that marginalized and disadvantaged perspectives are heard, integration goes further and requires that such perspectives be actively incorporated into the resulting solution.Integration explicitly sees difference not merely as something be preserved, but as the essential ingredient of progress.Because the goal of integration is joint creation the contribution of those with less power, or less familiarity with traditional modes of rational justification, is not just permitted or preserved, but demanded.
Like Type 1 deliberation, integration is vulnerable to charges that it is overly idealized or utopian, and thus unlikely to be realized in practice.However, it is difficult to compare the feasibility of deliberation and integration a priori, because each places different demands on participants.As mentioned, integration requires participants to preemptively forfeit the possibility of domination, and commit to a jointly created solution.Follett (1942a) acknowledges that, because we have been trained to enjoy domination, the difficulty of rejecting it is one of the chief obstacles to integration.Deliberation, however, makes a different demand on participants that may be equally unrealistic.In deliberation participants must be willing to accept domination, and admit "you were right, I was wrong" if they are unable to justify their position.In many cases such an admission could be so painful that participants would be willing to go to great lengths, including rejecting the norms of deliberation, to avoid it (e.g., Festinger, Riecken, & Schachter, 1965).In an integration, by contrast, participants can be assured that they will never have to accept domination.Thus, in situations where fear of domination is greater than the urge to dominate, integration may actually be more feasible than deliberation.
In sum, integration is clearly distinguished from ordinary communication, and prohibits dishonesty and domination by emotional manipulation, while being potentially even be more hospitable to a diversity of perspectives and marginalized populations than even more expansive Type II conceptions of deliberation.Although the demands it makes on participants may be infeasible in certain cases, they may, in other cases, actually be more feasible than those required by more traditional notions of deliberation.This suggests that a normative conception of democracy based on integration rather than deliberation may be better equipped to navigate these perennial challenges of inclusion and difference.

The Epistemic Benefits of Integration
I have shown that proponents of deliberation face a challenge in explaining why it is epistemically superior to non-deliberative voting.On one hand, information transfer among citizens during deliberation may increase overall competence.On the other hand, deliberation may erode the independence and diversity which are essential to the error canceling process that serves as a foundation for most claims about the epistemic power of democracy in general.Because Follett herself anticipated many of these challenges, her conception of integration was explicitly developed to promote information transfer while still preserving independence and diversity.Furthermore, integration's instance on creative solutions makes it especially well-suited to harness the epistemic power of error canceling.Follett (1918) clearly recognized the epistemic danger that "groupthink" and a loss of independence posed for deliberative conceptions of democracy.In the New State she articulates this danger by contrasting a "crowd," where individuals gravitate towards a single opinion, leading to "unthinking unanimity," with a "group," where differences are actually made more explicit, leading to "genuine collective thought" (Follett, 1918, p. 86). 7Follett explicitly notes the importance of this distinction for error canceling based justifications of democracy like the CJT, writing, "I have seen it stated in a sociological treatise that in any deliberative assembly there is a tendency for the wisest thought to prevail.This assumes that 'any deliberative assembly' is more like a group than a crowd" (Follett, 1918, p. 87).
Thus, Follett sees the tendency of a crowd towards unity and agreement as eroding the independence required for error canceling.Therefore, an integrative group, which not only preserves but highlights differences, is more likely to produce epistemic gains than an ostensibly deliberative crowd.
Follett is also sensitive to the potential for expert domination, which, as the DTA shows, can lessen the epistemic power of groups precisely by leading to consensus on an expert's pre-conceived opinion.For Follett (1924), experts do have an important function in integration: providing accurate information, which can help to distinguish productive differences from mere confusion and thus increase overall competence.But because experts are subject to the same biases and limitations as other individuals, she argues that "the expert must find his place within the social process; he can never be made a substitute for it" (Follett, 1924, p. 29).
Follett's skepticism of both the "crowd" and the "expert" help to justify, in epistemic terms, her rejection of domination and persuasion, and her insistence on co-creation.By rejecting domination, including domination through the voluntary submission to the crowd's pressure towards unanimity, integration guards against groupthink and preserves the diversity of opinions within the group.By rejecting persuasion through rational justification and requiring co-creation, integration prevents the expert from persuading the rest of the group from simply adopting his own pre-conceived "expert opinion," and ensures that, as the DTA recommends, the resulting solution reflects a diversity of perspectives and approaches.
More broadly, integration's insistence on co-created solutions makes it especially well equipped to harness the power of error canceling.As discussed above, the aggregation of diverse perspectives has the potential to cancel out the biases and limitations associated with any particular perspective because the aggregation itself is not formulated from within any one perspective.Yet, when deliberation produces consensus on a pre-existing alternative it fails to transcend the perspectives of individual actors.
To illustrate this point, consider once again the blind men and the elephant.Each man begins with a particular pre-existing proposal ("snake," "tree," "wall," etc.) for the identity of the object, which is limited and biased due to his particular perspective.Let us assume that the men are able to make use of a decision process (be it voting, deliberation, or fiat) that will lead them to consensus on the epistemically best of these pre-existing proposals with complete certainty.It is clear that while the chosen proposal will-by construction-be the most accurate of the original set, it will still be subject to all of the biases and informational limitations of its creator's own perspective.In order for the men to harness the error canceling power of their diversity, they must not only deliberate among their pre-existing proposals, but jointly create a new proposal, viz."elephant," which none among them could have formulated alone.It is thus integration, rather than a conception of deliberation centered on persuasion, that most closely resembles the error canceling process described in the DTA model, which assumes that agents collectively search for new ways of solving the problem that were not available to any one agent at the outset.
Integration's two requirements therefore give it the potential to exceed the epistemic benefits of both deliberation and voting.It has the same potential to increase competence though information transfer as deliberation, but with less risk of sacrificing diversity and independence.Unlike voting and deliberation it mandates joint-creation, enhancing the possibility of a solution which transcends the biases and social positions of any particular actor.These features provide a strong epistemic justification for integrative democracy, while simultaneously highlighting the challenges still faced by efforts to provide a similar justification for deliberative democracy.
The Systemic Turn: Can Integration be "Scaled-Up?"Integration may also offer other benefits over deliberation in other areas beyond the two I have explicitly discussed above.For example, decisions made through integration may be more likely to be perceived as legitimate by all sides, since they will never reflect the view of any single perspective.Integration may also be superior to deliberation as a method of civic engagement, insofar as it instills in participants a stronger sense of collaboration and collective identity than persuasion-based deliberation.These additional (potential) benefits of integration for democratic theory certainly deserve additional exploration.At the same time, there are also areas where Follett's theory of integrative democracy may need to be extended or updated in light of other challenges raised in contemporary debates about deliberative democracy.I will illustrate this general point by briefly discussing some challenges that the recent systemic turn in deliberative democracy may pose for Follett's theory.
Deliberative theorists have long wrestled with the question of how in-person, face to face deliberation might be "scaled-up" and implemented at the societal level.Such concerns are evident in Habermas' (1996) two-track model of deliberative democracy, and in the decades of debates surrounding deliberative minipublics (Dahl, 1989;Fung, 2007;Goodin & Dryzek, 2006;Mackenzie & Warren, 2012).The recent systemic turn within deliberative theory both highlights and formalizes these issues.As articulated by Mansbridge et al. (2012), a systemic approach to deliberative democracy is motivated by three key problems that emerged in earlier forms of deliberative theory.Firstly, it seems evident that no single fora (including minipublics) could possibly possess sufficient deliberative capacity for an entire democratic system.Secondly, deliberative theory must acknowledge the deep interrelatedness of social structures, and the difficulty of evaluating the deliberative character of a given institution without reference to the larger structures in which it is embedded.In particular, an institution may have low deliberative quality when viewed in isolation but increase the deliberative quality of the system as a wholeor vice versa.Finally, echoing the concerns of Type II deliberative theorists, Mansbridge et al. (2010) note the difficulties that deliberation has in dealing with structural inequalities-especially in regards to questions of access: who gets to participate in the deliberation in the first place?One promising but controversial idea which has emerged from a systemic conception of deliberative democracy is the possibility of what Goodin (2008) calls "distributed deliberation" (p.186), whereby different features of deliberation are not instantiated simultaneously but distributed across different democratic institutions.For example, the caucus room, parliamentary debate, electoral campaigning, and post-election bargaining display different deliberative virtues.Although none is classically deliberative in isolation, in combination they may be "good enough" to produce a deliberative society on the macro scale.Owen and Smith (2015) push back against this idea, and the broader systemic turn, arguing that conceiving of deliberation at the systemic level may neglect efforts to help individuals build the deliberative capacity necessary for achieving deliberation's normative benefits.
It is clear that many of the challenges for deliberation highlighted by the systemic turn are equally applicable, and perhaps even more applicable, to integration and integrative democracy.Like deliberation integration is conceived of as a face-toface activity taking place in small groups which cannot easily be scaled up to the societal level.The complex interrelatedness of different institutions and groups with vastly different characteristics likewise makes it difficult to evaluate the integrative character of a particular fora without an analysis of its place in the larger societal system.Finally, although integration's prohibition on persuasion make the actual process of integration less vulnerable to domination by those with greater resources and power, integration shares deliberation's challenge of ensuring that access to the dialogue in the first place is not restricted on the basis of pre-existing inequalities.One particularly salient challenge is whether those in positions of power who could achieve domination through coercive means would be willing to give up this possibility to participate in creative integration.Follett's conception of a hierarchy of integrative groups at the neighborhood, city, and state level clearly represents an attempt to deal with some of these issues.In some ways Follett's neighborhood groups anticipate modern conceptions of minipublics.Neighborhood groups and minipublics both represent an attempt to scale up a form of interpersonal dialectic to serve as a foundation for broader democratic government.In both cases ordinary citizens gather to discuss and debate societal issues, with the results of these discussions informing the broader policy process in some way.However, while minipublics are usually seen of as constituting a diverse "sample" of ordinary citizens who are brought together to debate a specific issue, Follett (1918) appears to see neighborhood groups as something in which all citizens participate regularly, not merely when debating specific issues.
One key question for integrative democracy is whether criticisms leveled against minipublics in the wake of the systemic turn are also applicable to neighborhood groups.Follett sees neighborhood groups as the central organizing principle of democratic government in general.Yet, proponents of the systemic turn argue that minipublics alone are unlikely to facilitate a deliberative society writ large, and may even be harmful to the deliberative legitimacy of society in certain contexts (Dryzek, 2010;Lafont, 2015).To the extent that this is true of minipublics, it is likely also applicable to Follett's vision of integrative democracy centered on a hierarchy of integrative groups.Follett's vision of the neighborhood group as the initial organizing unit of integrative democracy may also be undermined by partisan and ideological geographic sorting, which has become especially prevalent in the United States in recent years (Bishop, 2008;Lang & Pearson-Merkowitz, 2015).Because integration views difference as an essential resource, the increasing tendency of individuals to seek out like-minded communities may hamper the production of genuinely integrative solutions at the neighborhood, city, or even state level.Such tendencies would somehow have to be overcome before Follett's vision of a democratic system based on the hierarchical integration of differences could be fully instantiated.
At the same time, the systemic turn also suggests that a possible alternative for the totalizing vision of integrative democracy articulated by Follett (1918) in The New State could be found in adapting the concept of distributed deliberation.As with deliberation, it may be possible for interconnected parts of a larger system to work together to produce integrative outcomes without any one of them actually engaging in the process of integration as Follett defines it.For example, mass media communication might help to bring differences into the open, while public debate breaks apart wholes, and elected delegates collaborate to synthesize the revealed desires and positions into a creative integration.Echoing Owen and Smith's (2015) concerns, Follett (1918) herself would likely reject such a conception due to its failure to help individual citizens learn to exist as "part of a social whole," (p.180) which she sees as an essential component of integrative democracy.Yet, it might represent more realistic instantiation of integrative democracy that is still able to reap some of the benefits of integration described above.It may in fact be the case that integration proves most useful as just one component of a more traditionally deliberative democratic theory.That is, it could be that integration is more effective than deliberation in resolving some kinds of conflicts, but not others.Such an approach clearly has potential, especially if integration is found to be infeasible under many real-world circumstances.
The chief barrier to addressing all of these questions is the relative lack of empirical research on integration in the context of democratic decision making and civic engagement.Questions remain surrounding the viability of distributed deliberation and the effectiveness of minipublics despite substantial observational and experimental research on the feasibility and effectiveness of deliberation in small, interpersonal group settings (e.g., Dickson, Hafer, & Landa, 2008;Karjalainen & Rapeli, 2015;Muhlberger & Weber, 2006;Steenbergen, Bächtiger, & Steiner, 2003).Far less research exists exploring similar issues with respect to integration.Building on Follett's work, Walton and McKersie (1965) developed the concept of "integrative bargaining" in the context of labor negotiations, and some empirical investigations have explored the factors that lead participants in such negotiations to discover and achieve integrative solutions (Carnevale & Isen, 1986;Kirk, Oettingen, & Gollwitzer, 2013).However, such results may not be applicable when applying integration to political debates over controversial social or economic issues.Before integrative democracy can fully engage with the challenges raised by the systemic turn, more empirical research is needed on the determinants and overall feasibility of integration with respect to contentious political debates.

Conclusion
Although formulated a century ago, integrative democracy is still a young theory in developmental terms.Over the last fifty years or so deliberative democracy has gone through dozens of iterations and turns, and benefitted from the critiques and revisions of hundreds of scholars.By contrast, the development of integrative democracy as a political theory is virtually exhausted by Follett's own work in the early 1900s.Like any theory, integrative democracy requires the engagement of other scholars in order for its potential to be realized.This paper has demonstrated that integrative democracy, and integration in general, offer a number of theoretical benefits relative to traditional deliberative democracy.By prohibiting persuasion and insisting on joint creation, integration is even better equipped than more expansive notions of deliberation to effectively respond to questions of inclusion and fairness, without sacrificing sincerity or risking domination through rhetorical manipulation.Because it is well equipped to harness error canceling processes and preserves diversity and independence, integration also has the capacity to produce epistemic gains that exceed those of both deliberation and voting.At the same time, and unsurprisingly for such a "young" theory, integrative democracy requires additional development before these benefits can be fully realized.Insights from the systemic turn in deliberative democracy suggests that, even if it is feasible at the interpersonal level, integration must address a number of complex issues in order to be viable at the societal level.Integrative democracy requires both theoretical development and empirical verification before it can be fully assessed as a modern democratic theory alongside deliberative democracy, but the theoretical benefits of integrative democracy outlined in this paper suggest that such research is well worth pursuing.