Individual decision making can often be wrong due to misinformation, impulses, or biases. Collective decision making, on the other hand, can be surprisingly accurate. In Democratic Reason, Hélène Landemore demonstrates that the very factors behind the superiority of collective decision making add up to a strong case for democracy. She shows that the processes and procedures of democratic decision making form a cognitive system that ensures that decisions taken by the many are more likely to be right than decisions taken by the few. Democracy as a form of government is therefore valuable not only because it is legitimate and just, but also because it is smart.Landemore considers how the argument plays out with respect to two main mechanisms of democratic inclusive deliberation and majority rule. In deliberative settings, the truth-tracking properties of deliberation are enhanced more by inclusiveness than by individual competence. Landemore explores this idea in the contexts of representative democracy and the selection of representatives. She also discusses several models for the "wisdom of crowds" channeled by majority rule, examining the trade-offs between inclusiveness and individual competence in voting. When inclusive deliberation and majority rule are combined, they beat less inclusive methods, in which one person or a small group decide. Democratic Reason thus establishes the superiority of democracy as a way of making decisions for the common good.
This book follows the recent trend in democratic theory termed "epistemic democracy" in a novel way. Rather than rely on liberal philosophy or an analogy with science, it begins with results in mathematics, decision theory, psychology and cognitive science. It also mentions an evolutionary basis for the superiority of group decision making.
Landemore traces different nuances of epistemic arguments for democracy from Aristotle to Mill to Dewey to Hayek, then sets out on her own path resting on the Condorcet Jury Theorem, the Law of Large Numbers, and recent results in decision theory such as the Diversity Prediction Theorem. At its core it is a sophisticated application of crowd wisdom theory into the political realm.
Democratic Reason assumes that individuals bring good faith and (on average) better-than-random intelligence to their evaluation of the proper course of action or predictions about the future. She cites empirical work not only on the “wisdom of the many” but also the “averageness of the few,” such as Phillip Tetlock’s “Expert Political Judgment.” Tetlock, like Kahneman and Tversky before him, used extensive data to show that political and business elites are about as accurate as laypersons in predicting outcomes of political and strategic decisions.
The positive side of the argument (for inclusive deliberation) relies on Lu Hong and Scott Page’s work to the effect that diversity trumps average ability when it comes to group problem solving and decision making. To that end, Landemore recommends deliberation (in assemblies selected by lot) followed by voting, because large, diverse groups make better decisions. Her ideal democratic mechanism would begin with deliberation to bring out information and points of views, followed by a vote taken among options two at a time. She sets aside the issue of a “stop procedure” for the deliberation phase. While most deliberative democrats would recommend some version of consensus here, Landemore (in my view correctly) leaves this for more empirical work. It may very well be that stoppage at a point of having a satisfactory number of proposals is rarely an issue in deliberation; but it is not part of her theoretical argument.
Individual fallibility and bias, she says, citing for example Hugo Mercier’s work, may actually be an evolutionary adaption that helps us make better group decisions. If each person presents her best possible argument to a particular point of view (even blind to her own fallacies and biases) while others have the natural ability to find fault with her arguments, all important information will emerge before the collective tribunal.
In other words, evolutionary psychology could be telling us that we were “built to parley” when it comes to social decisions. We make horrible decisions (see Kahneman and Tversky) as individuals and surprisingly good ones in groups (Surowiecki, Page) at least when the group is free of systemic bias. Diversity is essential not only because it brings in different points of views, heuristics, and information, but also because it allows individual errors to cancel.
What first appeared to be a missing link between deliberation and voting turned out not entirely crucial. Landemore, like many others, assumes it is possible to end a deliberation with yes/no (binary) votes between possible political choices. She states that current parliamentary procedure already functions in this way, and that even if the order of voting could affect outcomes, the agenda could be set in a random order so that all options are given equal chances. Some might find it unconvincing; some might find it a detail to be worked out later.
An issue that most would find minor, but that in my own opinion ought to be further explored by political scientists, is majority vote. She, like most theorists, assumes that in so far as voting is concerned, we must go with majority vote, because any other way of counting amounts to minority rule. Granted consensus would be next to impossible in any but tiny groups, but the choice is not binary. The option of supermajority should be given fair consideration.
Landemore makes it clear that more data are needed on what actually happens in deliberation both in current representative assemblies and among non-experts in other contexts. If nothing else, Democratic Reason begins to undo the distrust of “the masses” that has plagued political science since the time of James Madison. While remaining humble about its own claims and leaving many questions open for more evidence, this book shifts the burden of proof (in my non-expert opinion) with respect to quality of political decisions onto those who oppose democracy.
It also makes perhaps a greater achievement; it beats rational choice theory, usually employed in political science to bemoan the meaningless of voting or belittle the ignorance of the average voter, at its own game. Democratic Reason, like much recent literature on collective intelligence, argues that (pre-deliberation) ignorance of the average citizen is irrelevant when thinking about systemic properties. Democracy works (absent systemic bias) because what happens at the group level overcomes the shortcomings of individuals. This is, after all, why we (more or less universally) work in groups and teams in the first place.
Yet another interesting aspect of Landemore’s book, like epistemic democracy in general, is side-stepping the “communitarian” versus “liberal individualist” conundrum. Democratic Reason proceeds neither from “aggregation of preferences” nor an assumption of a mythical “general will” (at leas as often interpreted). It is founded on something entirely different: “general wisdom.” Epistemic democrats, whether Joshua Cohen, David Estlund, Robert Goodin, or Hélène Landemore, do not normally claim they are inventing anything new but rather trace their roots as far back as Aristotle, Machiavelli, Rousseau, Mill, Dewey, Pierce, or Hayek.
The book tables the issues of fairness (that democracy is inherently desirable independent of outcomes) as well as good faith and corruptibility. But in so far as the tasks it sets out to do, provide a purely epistemic justification for democracy over aristocracy and monarchy, it thoroughly succeeds.
In defense of epistemic democrats, they do not claim that only correctness of decisions (however measured) is important but that correctness must be part of why democracy in the first place. Democracy, they would say, like anything one comes to love has more than one thing going for. New or simply a return to ancient wisdom, it is exciting work! Democratic Reason is an especially readable manifestation of epistemic democratic theory.
Hélène Landemore's Democratic Reason is an important book in a time when democratic cynicism and illiberal attitudes are waxing. Largely avoiding fundamental questions of democratic legitimacy, she focuses on the epistemic dimensions of democracy, arguing that any reasonable case for democracy must rest at least in part on its epistemic credentials. That is, it's ability to get certain questions and problems right.
Landemore draws heavily on the work of Lu Hong and Scott Page, who show that diversity enhances problem solving and prediction because diverse groups leverage different perspectives with different insights, heuristics, or just ways of seeing problems. Democracies of wide or universal suffrage are necessarily more diverse than monarchies/dictatorships or oligarchies/epistocracies. Importantly, this dynamic works even when group members aren't intellectual heavyweights, so Hong's and Page's "diversity trumps ability" theorem shows.
Landemore also effectively responds to the problem of rational irrationality posed by Bryan Caplan. She concedes that democracy is fully consistent with assigning certain especially technocratic questions to independent, appointed authorities (e.g., the Federal Reserve or Supreme Court). But beyond this she argues that Caplan and similar economists conflate questions of economics with political questions with economic dimensions. Hong-Page style reasoning about diverse perspectives apply. Economists, for all their obvious expertise, reason from perspectives that are limited in various ways, some straight from the axioms (e.g., rational utility maximizing individuals).
These arguments are the strongest in the book. One complaint I have is that some of the other arguments seem like Landemore is just throwing everything at the wall to see what sticks. Especially the section defending the relevance of the the Condorcet Jury Theorem is weak and unnecessary, given its dubious conditions (>50% likelihood of individuals getting problems right, their independence from one another, and the limited applicability to binary choices).
One of the most interesting ideas in the book is also unfortunately not developed, but ostensibly saved for a future work: the idea of epistemic democracy evolving through time and learning from past mistakes. Constitutional safeguards, independent apolitical technocratic bodies, and technological changes to the democratic process all may be fruitfully understood as democratic learning. I hope Landemore explores this in a future work.
Finally, as an awkward side note, I also found the chapter on the genealogy of epistemic democracy to be especially edifying.
On the one hand she keeps talking about "procedure independent truth". On the other, she says her account tries to understand non-ideal conditions in which people form their opinions mixed with emotions. When truth is "procedure independent", how do we know that it is truth when we arrive at it? If we are able to recognize it when we arrive at it, why the hell are we searching for it? If people like Landemore are the ones who are able to say, "now you arrived at truth", how different is this account from Neo-Kantians -who secure the outcome of the deliberation a priori- that she criticizes that much? A huge pile of empty concepts. Bullshit.
Everything but the central argument in the book is excellent. Unfortunately, the central argument is an obvious nonstarter given the assumptions Hong and Page's theorem needs to work do not even get within the same solar system as the conditions of real-world politics. I continue to puzzle how someone with all the abilities demonstrated in the other chapters of the book thought they had shown anything by using Hong and Page.