Arguing across common ground
Reasonable People #3: When good reasons are a bad guide
Why listen to reasons? Why not just believe the people you trust (whether they give you reasons to believe or not), and ignore people you don't trust (whatever their reasons)? Even if individuals can reason in their best interests, it doesn't follow that other people won't selectively generate arguments which support their selfish ends, rather than the truth. Ignore other people's arguments and you also avoid any risk of being manipulated. If reason isn't necessary and can't be trusted, the argument goes, why bother?
Looking around the world it isn't hard to find spaces where the exchange of reasons seems missing, or merely performative; and plenty of thinkers who will claim that human interaction is conditioned by a stew of brute force, economic necessity and unconscious urges, reason and argument merely a comforting sop to human dignity but which exerts no causal force.
Mercier & Sperber (2011) have an explicit response to this challenge. Yes, how you feel about someone probably is the main guide we use when deciding whether to believe them. And, yes, arguments can mislead. But the value of argumentation is for precisely those situations where you do not trust someone. Reason, they claim, exists to produce arguments which can persuade across the chasm of distrust.
This social view of reasoning contrasts with the way we often think about reason (the solo, internal, activity of The Thinker). Their interactionist view explains why reason can sometimes be such a poor guide - it doesn't matter if we fool ourselves with reasoning, since it matters more that we stop ourselves being fooled by other people. We are evolved, Mercer & Sperber claim, to reason in an asymmetric way - slapdash generation of reasons, merrily justifying what we must or want to do, but skeptical evaluation of reasons when they are provided by other people.
"Would I lie to you?", bootstraps a claim on the assumption that I wouldn't, on interpersonal trust. The exchange of reasons can work even if you believe the opposite: "Yes, you would lie to me". Argumentation bootstraps a claim on the shared knowledge that exists between two people, even in the absence of trust between them. When you advance an argument, the first heuristic I apply is to see if I trust you, but the second (according to Mercier & Sperber) is to coherence check that claim, asking if it makes sense on its own terms, and in relation to everything else I know.
Even if you don't trust Owl's judgement, when he says (A) "Look you know Pooh Bear is a honey fiend" and (B) "You know the honey pot was full before Pooh came to visit" and (C) "You can see the pot is empty now" then this allows you to accept Owl's claim (D) "Pooh must have snaffled all the honey". A, B and C are part of the common ground, and suggest D independently of this being something Owl also wants you to believe.
By using epistemic common ground, arguments allow people to send and receive more and better information, and so benefit from communicating using reasons.
I like this line of thought a lot, but even if we accept the central claim - exchange of reasons could evolve because it is on average beneficial to individuals - it is interesting to ask when exceptions would occur. Why, exactly, does believing due to reasons work, and when does it fail?
The idea that coherence checking can be a guide to truth seems to rely on a basic alignment between good reasons and reality. You provide reasons which help me recognise the truth. But when might true reasons misdirect?
A simple case might be then I draw the wrong inferences: We establish common ground, but I draw different - and erroneous - conclusions from you due to some peculiarity of my own reasoning.
Other cases might be when I make a false assumption, or am missing a crucial piece of information. For example, in a duel of wits, you offer me two glasses of wine, telling me that one is poisoned: I must choose and we both shall drink. I end up drinking poison, because I assume you haven't poisoned both, leading me to discount your option to put poison in both glasses since you have spent the last few years developing an immunity to the poison. If I had known this, my reasoning would have drawn a different conclusion.
The interesting thing about this kind of case is that it isn't clear when I can be sure I have the right set of background knowledge to verify an argument. Classical logic is monotonic - established truths are never overturned - but reality follows a non-monotonic logic - additional facts can supersede the inferences granted by other facts.
"If it rains you'll get wet" can be true. But it isn't true if you have an umbrella. But if your umbrella has a hole in then the original claim holds (admittedly for finessed reasons). And if your umbrella hole is small, or patched, or you're wearing a coat, the original claim switches back to being false. No scaffold of facts is robust enough to survive a revision that turns the conclusions upside down (something, incidentally, that we obviously actually enjoy experiencing via a whodunnit movie or mystery story).
What this all means for Mercier & Sperber's argument is that although arguments may well be effective when trust is lacking (and indeed have evolved precisely for low-trust communication), they cannot be effective when there isn't common ground between argument-maker and argument-receiver. When we lack common assumptions and, crucially, knowledge that we share common assumptions, I cannot be sure that I will draw the correct inferences from your arguments, even if they coherence check against aspects of reality which are in our common ground. I can never be sure there isn't some knowledge I'm missing, or false assumption I've made, which will break the basic alignment between reality and reasons. As a consequence, I must distrust the arguments you make (note, this is second order trust - it isn't distrust of the person, but of the arguments they make).
To bring this home, I think this is why arguments are so hard on the internet. The internet brings more and more diverse people together, and so necessarily the area of overlap between their backgrounds, experience and knowledge will be smaller. This diminished common ground challenges the power of reasons. Arguments that seem good to one group don't translate for another, to the perpetual affront of a certain kind of rationalist, who thinks that arguments should stand isolated from social context (and for, whom, the social context of the early, more homogeneous, internet, was particularly suited to).
Head food
PAPER: Why do humans reason? Arguments for an argumentative theory.
The paper which switched me on to the area of argumentation theory, and the precursor to their landmark book on the topic “The Enigma of Reason”. I revisited it because the journal where it is published, Behavioral and Brain Sciences, has a very special feature where they invite commentary, and then publish the commentary alongside the original paper and a response by the authors. I’ve always loved how this format makes public (and puts in one place) the disagreements around a theory, so I went back to it to see what people didn’t like about The Argumentative Theory when it was first proposed. A couple of commentaries (Sternberg, Opfer & Sloutsky) claim that the theory fails because arguments compete with (or are irrelvant compared to) other persuasive factors such as trust and emotion. In response Mercier & Sperber write:
“But if detecting the trustworthiness of communicators were the only heuristic used, then receivers would end up rejecting a good amount of genuine and relevant information when they lack sufficient ground to accept it on trust. ... Coherence checking, we argue, is the second major heuristic used in filtering communicated information, and is at the basis of reasoning proper…Arguing consists in displaying coherence-based reasons for the acceptance of a given message. It is, in essence, an “honest display” strategy opened to evaluation and aimed at the audience’s epistemic concerns”
which got me thinking about the issue of when reasons are and are not a good guide to truth (see above).
Full reference: Mercier, H., & Sperber, D. (2011). Why do humans reason? Arguments for an argumentative theory. Behavioral and Brain Sciences, 34, 57-111.
LINK: ‘Eugenics is possible’ is not the same as ‘eugenics is good’: Rows involving Richard Dawkins and Andrew Sabisky reveal how difficult it is to 'decouple' controversial concepts in our heads by Tom Chivers. Tom Chivers uses Dawkin’s tweet about eugenics to talk about uncoupling, which I like as an idea, but I think only goes so far in the Dawkin’s case. As a professor of both biology and public understanding for him to use the word “Eugenics” and then claim he just meant selective breeding was somewhere between disingenous, inflamatory and culpably ignorant. Simon Witten is good, here, on some of the baggage the idea of Eugenics comes with (this doesn’t mean I’m a low decoupler, just that I believe that words have specific meanings).
LINK: Cory Doctorow: Inaction is a Form of Action. An argument about how big tech should be regulated to protect free expression and a public sphere online (and absence of government censorship isn’t the end of what we need to do create “freedom of expression”)
BOOK: Michael Tomasello, Origins of Human Communication (2010, MIT Press)
Tomasello develops an argument about the evolution and development of language which, he argues, is based on the common ground of knowledge between people and - crucially - referentially indexed by gesture before words. Imagine, he says, communicating with someone with whom you share no language (easy, he says, you gesture). Now imagine communicating with the same person only with words, without gesture (very hard, he says, you’ll never solve the “gavagai” problem).
What this means is that many of the especially powerful properties that people often attribute to language— including referring others to perspectives on things and to absent referents—are actually present more fundamentally in human cooperative communication with very simple gestures. This is possible because of—and only because of—various types of common conceptual ground and joint attention between communicators
Language, he says, can only be understood in the context of human hyper-sociality, a hyper-sociality which is nascent in great apes but has undergone a step-change in human evolution.
PODCAST: Government vs The Robots
Jonathan Tanner interviews a guest about how technology will change politics. The episode with James Ball is very good: Fake news, post-truth and all that jazz, and I enjoyed this one about the psychology of politics and political campaigning with Alex Evans: Trigger Warning. Twitter: @govt_vs_robots
LINK: We we need to disagree. From the always-worth-reading Tim Harford, with links to some classic studies and new books on the role of dissent in group decision making. For what it is worth, the two classic studies which are alluded to in the column, but not explicitly referenced are
Stasser, G., & Titus, W. (1985). Pooling of unshared information in group decision making: Biased information sampling during discussion. Journal of personality and social psychology, 48(6), 1467.
Phillips, K. W., Liljenquist, K. A., & Neale, M. A. (2009). Is the pain worth the gain? The advantages and liabilities of agreeing with socially distinct newcomers. Personality and Social Psychology Bulletin, 35(3), 336-350.
REVIEW: Enlightenment Later. Will reason survive rationalism?
Kent Anhari reviews "Irrationality: A History of the Dark Side of Reason" by Justin E. H. Smith in The New Atlantis.
To argue that reason entails its opposite, as Smith does, is to suggest that both reason and unreason have fixed and stable identities, such that we can track the motion from one to the other. At his least ambitious, Smith seems to forget himself and to take for granted that what “reason” means is obvious
LINK: Reasoning Is More Intuitive than We Think
Short blog for Psychology Today in which Hugo Mercier argues that the distinction between intuition and reason is not a good one
When people reason on their own, reasoning can indeed by slow and effortful. But when they argue, finding and evaluating reasons comes very spontaneously-sometimes, all too spontaneously: we can all think of cases in which we kept arguing long after we should have given up. This difference is no accident. If reasoning comes much more easily in the context of a discussion, it's simply because it is designed to work in such a context, and not when we engage in private ratiocination
OFF-TOPIC: Dead Wood by Nick Hunt in Emergence magazine
Nick weaves together the story of his visit to Białowieża, Europe’s largest surviving primeval forest, with the history of the appearance and disappearance of the forests across old Europe, with myths of clearing and cover, life and death. Disclaimer: I know Nick and got extra pleasure hearing his voice reading this essay (follow the same link for the podcast version), so your milage may vary.
END, cheers
