Why Proposers and Disposers need to be kept distinct: The Argumentative Theory of Reasoning

The ‘argumentative theory of reasoning’ (conceptualized by Dan Sperber and developed with the evolutionary psychologist Hugo Mercier) hypothesizes that reasoning serves two distinct survival-related functions: a) convincing people and b) evaluating the arguments of others – ‘thereby allowing communication to proceed even when trust is limited’ (Landemore, 2013, p. 126). The theory – developed as an evolutionarily-plausible alternative to the classical (Cartesian) model of reasoning as a way of updating and correcting one’s own beliefs – is based on the distinction between performing speech acts and evaluating the performative utterances of others:

According to this theory, individual reasoning works best when used to [a] produce and [b] evaluate arguments during a public deliberation. It predicts that when diverse opinions are discussed, group reasoning will outperform individual reasoning. (Mercier & Landemore, 2012, p. 243)

‘Exposing people to disagreement and debates increases their ability to entertain different opinions . . . either by witnessing a debate or by being part of one’ (ibid., p. 252). The important factor is not participation in speech acts so much as ‘the presence or expression of dissenting opinions in deliberative settings’ (ibid., p. 254, my emphasis). (Mercier & Sperber, 2017) also point out that as a species we are much better at evaluating reasons than producing them – I may not be able to see the beam in my own eye, but I can find the mote in yours. Laboratory studies indicate that we are better able to find the flaws in our own reasons when we believe those reasons to have been produced by someone else, thereby supporting the case for the division of labour between persuaders and evaluators, as it’s very hard to change one’s mind about one’s ‘own’ (i.e. indigenous) convictions.

Given that the argumentative theory of reasoning suggests two distinct mechanisms this would also suggest two categories of persons in a deliberative context – proposers and disposers. The former category

should only be concerned with  . . . [arguments] that increase the plausibility of a conclusion one is trying to defend, or those that decrease the plausibility of a conclusion one is trying to rebut. Representations that have the opposite effect are of no direct value if one wants to convince one’s interlocutor. Reasoning should therefore be directed towards these valuable representations, and, as a result, it should display a strong confirmation bias (Mercier and Landemore, 2012, p. 241, emphasis in original)

The confirmation bias, however, ‘mostly affects the production, and not the evaluation of arguments’ and can be ‘checked, compensated by the confirmation bias of individuals who defend another opinion’, hence the necessity for ‘the presence or expression of dissenting opinions in deliberative settings’ (ibid., pp. 251, 253, 254). The requisite cognitive tool for those whose task is to evaluate communicated information (i.e. disposers) is ‘epistemic vigilence’, the polar opposite of the confirmation bias (Mercier, 2013; Sperber et al., 2010).

Mixing the two cognitive styles in the same persons, as most deliberative democrats advocate, is sub-optimal as while ‘reasoning is mostly truth oriented when it come to evaluating arguments, it is still biased by the function of producing arguments’ (ibid., p. 127, my emphasis).  Whilst it is possible for the same agents to oscillate between the two different cognitive styles . . .

After reviewing and weighing for ourselves the reasons for and against a given action, we come to a conclusion. We then take a position. However, when we speak in public in the course of deliberation, we share only the part of information that supports our position. (Manin, 2005, p. 16)

. . . nevertheless, cognitive dissonance theory would suggest that it is more difficult to modify one’s evaluation of an argument once having expressed an opinion. E.M. Forster’s dictum ‘how do I know what I think until I [hear] what I say’ could be taken as contributing to the confirmation bias – it’s much harder to change your mind once you know what you think. Speech acts intended to persuade (whether oneself or others) are inherently partisan and Landemore makes a good case for the confirmation bias as a valuable cognitive tool for proposers and other rhetorical persuaders (but not for disposers). This would also suggest a dialectical approach to truth-proposing, whereby different parties institute ‘an exchange of arguments for and against something’ – as in a judicial trial (Landemore, 2013, p.140).

Habermas’s focus on the force of the argument (albeit an ‘unforced’ force) illustrates the relevance of speech-act theory to the proposing function (and also the evolution from war-war to jaw-jaw). Inverting Clausewitz, why go to the risk and expense of fighting when the same result can be achieved using purely rhetorical ordinance? Just as battles tend to have two sides, this tends to be equally true of argumentative exchanges and debates – legislative or otherwise – as most proposals can be reduced to a binary format – which has the added advantage of a simple resolution mechanism via up/down majority vote (Risse, 2004). By contrast, deliberative and epistemic democracy – which makes no distinction between proposers and disposers – would appear to adhere to the classical (Cartesian) model of reasoning as the revision and updating of one’s own beliefs. The fact that human agents do not appear to be endowed with cognitive mechanisms designed for the impartial updating/correction of their own beliefs would suggest that this is largely a set of ideal procedures – ‘a normative theory of democracy initially developed at a highly abstract level by philosophers and political theorists’ (Mercier and Landemore, 2012, p. 243), rather than a form of reasoning that comes naturally to human beings, and explains why attempts to implement deliberative democracy are heavily reliant on the use of procedural rules, along with trained moderators and facilitators.


Landemore, H. (2013). Democratic Reason: Politics, Collective Intelligence, and the Rule of the Many. Princeton, NJ: Princeton University Press.

Manin, B. (2005). Deliberation: Why we should focus on debate rather than discussion. Paper presented at the Program in Ethics and Public Affairs seminar.

Mercier, H. (2013). Our pigheaded core: How we became smarter to be influenced by other people. In K. Stereiny, R. Joyce & B. Calcott (Eds.), Cooperation and Its Evolution (pp. 373-398). Cambridge, MA: MIT Press.

Mercier, H., & Landemore, H. (2012). Reasoning is for arguing: Understanding the successes and failures of deliberation. Political Psychology, 33(2), 243-258.

Mercier, H., & Sperber, D. (2017). The Enigma of Reason: A new theory of human understanding. London: Allen Lane.

Risse, M. (2004). Arguing for majority rule. Journal of Political Philosophy, 12(1), 41-64.

Sperber, D., Clement, F., Heintz, C., Mascaro, O., Mercier, H., Origgi, G., et al. (2010). Epistemic vigilance. Mind and Language, 25(359-393).

7 Responses

  1. Thank you, Keith, for this very interesting summary!

    Couldn’t one implication for citizens’ assemblies be that we ought to invite parties (and other stakeholders) to advocate their positions and then let the assembly assess them, rather than starting from the participant’s own opinions?


  2. Absolutely — that’s the conclusion we draw from the argumentative theory of reasoning. Alex’s Superminority Principle is an attempt to ensure the positions advocated map as closely as possible to the diversity of public preferences. But we do conclude that the old model of reasoning underlying current initiatives for citizens assemblies is misguided and we’re encouraged by the fact that Bernard Manin, who’s seminal paper kick-started the deliberative democracy movement, has come round to this view. The puzzling thing is that Helene Landemore, who is in most respects a traditional deliberative democrat, was the one who drew this to our attention.


  3. Could it be because she believes that participants in citizens’ assemblies don’t start with strong personal views and take up arguments presented by others (including experts) rather than trying to defend their own? That sounds plausible to me and it would weaken the confirmation bias challenge. Is anyone aware of empirical evidence about this?


  4. Yes, that’s certainly plausible. The question that I would want to ask is what then is the added value of deliberation within the disposing body (given the potential drawbacks)? A trial jury is charged with determining the facts of the matter and the ideal is unanimity, whereas a legislative jury is only interested in informed preferences. If you look at the evidence from the Bloomfield Track DP (reported in Bob Goodin’s book), most jurors made their mind up on the basis of the evidence supplied, and the deliberative phase added little value.


  5. It’s hard for me to believe that deliberations would have no added value. Merely listening to testimonies doesn’t mean participants have understood the relevant information. So, there may be a reduction of mistaken beliefs through informed deliberations. In addition to this, there is the diversity of social perspectives gathered through random selection that is beneficial only if members of majority groups come to integrate the legitimate demands of underrepresented or dominated groups. So, I don’t believe that being merely exposed to a diversity of facts and arguments without engaging with them and weighing them together is enough. But I’ll have to check the empirical literature more closely.


  6. Active deliberation is a double edged sword with benefits and harms to good decision-making. Ideally a good democratic system should be designed that captures the bulk of the benefits while avoiding the bulk of the harms. This can be done by having different groups of people perform separate tasks. Active deliberation by a highly diverse group can help discover win-win solutions. It can also share important information or perspectives that is initially only in the heads of a minority, or even a single member, that turns out to be very valuable to the group. It would be a terrible disservice to prohibit active deliberation by a random diverse group.

    However, this active deliberation makes these people no longer able to fairly judge their decisions. Even if the argumentation theory of reason is disproved, there are a raft of proven cognitive biases that support this. So the solution is to have ACTIVE deliberation in the phase where policies are being drafted, refined and finalized, but to have a separate group that only observes pro and con presentations, with only the option to ask clarifying questions make the final yes/no decision about adoption.


  7. The Athenian political juries did not deliberate — it was down to the advocates to ensure their arguments were comprehensible and persuasive. Alex did suggest small-group deliberation might be possible, but that would require an exponential increase in the size of the overall decision-making body. And Fishkin’s experiments have indicated a need for trained moderators and it’s hard to see how that would be compatible with a jury with statutory powers (perhaps that’s why the DP is only advisory).

    Terry, Alex and I agree with the Deliberative Systems approach (Parkinson and Mansbridge, 2012), whereby a democratic system does not require that all parties should have a deliberative role. In the Kovner-Sutherland approach, deliberation takes place within the advocacy groups (political parties) and wider civil society (in the Habermasian sense), whereas Terry wants all legislative bodies to be randomly selected. I haven’t read Helene Landemore’s new book (no need to now, given Yoram’s 10-part book review!), but her original work provides no examples of the epistemic potential of deliberation within randomly-selected groups. And her “selective genealogy of the talking [collective wisdom] approach” actually supports the opposing (wisdom of crowds) model. And she draws the wrong conclusions from her study of the argumentative theory of reason, as the two cognitive modules it posits underwrites the case for a formal separation of powers between persuaders and disposers, rather than the collaborative group deliberation that she advocates.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: