akovner.github.io

A blog about hybrid politics

What is a Majority?

Yes, I know, it seems like an obvious question: a majority is more than half of a group. But actually the term has taken on a loaded meaning in democratic theory: it is at once associated with democracy itself, in phrases like “majority rule”, while it is simultaneously raised as a danger to democracy, as in the “tyranny of the majority”. So which is it?

First of all, we should note that the meaning of a majority in the legislature depends on the structure of the chamber. Traditional legislative chambers handle all three major legislative tasks: agenda setting, proposal writing, and final passage. As a result, such chambers must have a pre-existing majority just to conduct ordinary business. The existence of a durable majority is the defining feature of legislatures in elective oligarchy everywhere. In fact, it is one of the key properties that makes elective oligarchy undemocratic.

There is another way that majorities can form, however: in response to particular proposals. This is closer to the vision of the framers of the U.S. Constitution, who did not envision a strong party system, and imagined legislators as voting for proposals on a case-by-case basis, rather than merely advancing the interests of their party. To the extent that the founders’ design depends on this property, the Constitution has not worked well at all.

There is an intuitively strong pull toward majorities that form in response to proposals, rather than durable majorities. The dangers of durable majorities in promoting extremism and veering away from a democratic course are, unfortunately, all too obvious in light of current political conditions. The question, then, is how do we set up the conditions for majorities to form in response to proposals, rather than to create and enact them?

A good starting point in this quest is the argumentative theory of reason, advanced by Hugo Mercier and Dan Sperber in 2011. This theory has been described as trite, but while it is a simple model, its explanatory power belies that description. Basically, it asserts that the primary evolutionary function of reason was to argue and convince others, and be convinced in turn—in other words, it is first and foremost a networking tool, not a tool for making decisions. The original paper makes clear that, far from being vacuous or obvious, this theory makes definite predictions that are consistent with observation, and that contradict alternative explanations.

One consequence of the argumentative theory is that confirmation bias is a feature of reason, not a bug. A durable majority is a kind of structural analogue to confirmation bias: as part of such a majority, there is a strong incentive to align one’s views with the other members of the majority, as well as a strong social component in that most of one’s discussions (argumentation) will take place in a polarized context.

The most obvious reason for durable majorities is simply motivation: durable majorities give their members power. I have other posts in my legislative series that discuss structural solutions to the problem of durable majorities, most notably separating agenda setting, proposing, and deciding. Here, however, I want to look at some cognitive aspects of this problem.

Probably the most important conclusion of the argumentative theory is that reasoning often lowers the quality of decision making. Sperber and Mercier state that, “Using similar paradigms in which some participants are asked for reasons, it was found that providing reasons led participants to choose items that they were later less satisfied with…or that were less in line with the ratings of experts…. Participants got worse at predicting the results of basketball games…. People who think too much are also less likely to understand other people’s behavior…”

Sperber and Mercier go through a list of mechanisms by which coming up with reasons can harm the quality of decision making. In politics, the main ones are licensing, in which decisions are justified that would otherwise be viewed as immoral or unethical, and accountability, in which decisions will be taken because it is easier to justify (to account for) rather than because it is better. It may seem counterintuitive that accountability is here viewed in a negative light, whereas the accountability of elections is ordinarily viewed as a strength. But I think we can all identify with “soundbite” culture, in which decisions that can be explained in few words win out over better decisions that are harder to explain.

This suggests that, in order to make good decisions, people actually need to be freed from the requirement of coming up with reasons for their choices. It turns out that there is a big difference between forming an intuition about something and reasoning about it. In the former, a person takes in and processes information in order to form a mental model of the objects to be considered. In the latter, one is primarily concerned with rationalizing a choice; that is, justifying or explaining it to others.

Sortition’s great promise isn’t to merely replace elected officials with allotted ones. It is to allow decision makers to consider the options without being bound to generate reasons for it. In that sense we want to remove any ties to the “specialness” of the decision makers. Elections are bad (for final decision makers) because they force them to justify themselves, both before election (election promises) and after. This is not just a matter of winning reelection, it also causes elected officials to say, “I was elected, I must be special. I must therefore show how special I am.”

Returning to the idea of a majority, how on earth could an elected official be part of a majority that forms in response to a proposal? There are, I’m sure, rare exceptions (e.g. Justin Amash) in history where this has happened, but generally, an elected official is cognitively bound to adhere to a durable majority by the nature of the electoral process, even apart from more pedestrian political considerations.

Using allotted legislators is therefore an important step. But it is by no means sufficient: since a traditional legislative chamber requires a durable majority just to conduct ordinary business, even allotted members will find themselves forced into decisions prior to proposing legislation. For an assembly that meets over years and considers multiple proposals, allotted members are public figures irrespective of their manner of selection. They will feel pressure to justify their actions, bringing us back to the state we wished to avoid in the first place.

Even worse, the production of proposals by its very nature requires reasoning, as opposed to merely forming an intuition about proposals created elsewhere. My view (not that of Sperber and Mercier) is that reasoning really exists for this purpose: it allows a group to form complex plans, consistent with the fact that reason is a networking phenomenon first. Those who form proposals should therefore be disqualified from deciding upon those proposals on purely cognitive grounds, apart from issues of self-dealing or corruption.

Final decisions should therefore be made by people who have the time and resources to learn about the various proposals, and can form strong intuitions about what those proposals will do in practice. But they should be as free as possible from the need to justify their decisions. The natural structure for this is a citizen jury with the following requirements:

  1. Randomly selected from among those who had no direct hand in writing any proposals. (There is a conflict here with the democratic requirement that every adult citizen be included in the jury pool. Here I am considering only the cognitive requirement.)

  2. The jury is single use. This prevents jurors from engaging in reasoning by “consistency”; that is, making poor decisions in order to meet some imaginary standard of having applied the same reasoning to two different, dissimilar, situations.

  3. The jury should be sheltered from direct scrutiny. There is no accountability mechanism for jurors, other than to prevent blatant corruption. This is perhaps the most controversial criterion, but it is necessary. But doesn’t this present a real danger that citizen juries will go rogue and make malicious decisions? I will address this in another post, but for now, it must be noted just how constrained such a jury is. It can only act on proposals given to it, only once, and each vote is only a small part of the whole.

  4. The jury should be sheltered from lobbying. Interactions within the jury should be allowed only to aid in forming a clear intuition about the proposals. Great pains should be taken to avoid argumentation among jurors, as opposed to merely educational discussion.

Returning to our question, what is a majority? Or more appropriately, what is a deciding majority, i.e. a majority for final disposition as opposed to a working majority for other purposes? For me, a majority should only be dispositive if it forms in response to proposals, and to the underlying issue that those proposals address. As I have argued here, the requirements for breaking up durable majorities are stringent, but they are quite feasible. If nothing else, this exercise should show us just how compromised our leaders really are, not just by money or corporate power (though certainly by those as well) but just by the fact of having to justify their actions at every step. It may feel good to us as citizens to demand such justifications, but they lead to poorer decisions.