Towards a culture of reason
Reasonable People #25: short summaries of some disparate writing on reasoning. A "find your own patterns" edition of the newsletter
Facts: necessary but not sufficient
Contrary to some portrayals of the vaccine hesitant, they are far from unpersuadable. We’ve some work on this coming out soon, but this Danish study shows that a 85 word, 15 second, information intervention can significantly shift people’s intention to get vaccinated. Here’s that intervention in full:
Did you know that when you get vaccinated, you’re protecting yourself and your community? This concept is called herd immunity. Herd immunity denotes the effect that occurs when acquired immunity against a pathogen within a population (the “herd”), generated through infection or vaccination, has reached such a level that non- immune individuals in the population are also protected because the pathogen can no longer be transmitted. Thus, if you get vaccinated, then you can protect others who are not vaccinated
The paper also discusses an empathy-inducing intervention, which had an independent effect. It is a truism that facts are not enough and we should not assume that people with different beliefs merely have an information deficit, but the dangerous opposite of this is facts are not necessary, which can shade into people are irrational and unpersuadable, so it is nice to see some evidence that information and arguments have a non-zero effect, even if they aren’t the whole story
Pfattheicher, S., Petersen, M. B., & Böhm, R. (2021). Information about herd immunity through vaccination and empathy promote COVID-19 vaccination intentions. Health Psychology.
arguments from authority are a significant rational resource for the lay person
Lists of logical fallacies occupy a special place for teachers of critical thinking, but this project is undermined when you realise most logical fallacies can be reframed as serving rational ends. There’s no better illustration than argument from authority. Sure, something isn’t true just because someone with authority says so, but I challenge anyone to rid their argumentation of authority claims. The real challenge is identifying unwarranted authority, which begs the question.
Here’s Louise Cummings:
So trust, it is suggested, has a heuristic function in public health. Trust and a range of other cognitive and emotional processes fill a void that is created by a lack of knowledge and uncertainty in public health. By bridging gaps in one’s knowledge, trust can facilitate judgment making and rational action. Skepticism and inaction are the inevitable and pernicious consequences of a cognitive policy to abandon trust.
the results of a study of public health reasoning in 879 members of the public are examined. These results confirm that lay people,like experts, are reasonably adept at recognizing the logical and epistemic conditions under which these arguments are rationally warranted
perceived competence, objectivity, and honesty, provide essential grounding of the premises of authority arguments
If public health can learn to trust the rational resources of the public, then perhaps the public will reciprocate by increasing its trust in public health.
Cummings, L. (2014). The “trust” heuristic: Arguments from authority in public health. Health Communication, 29(10), 1043-1056.
’
Language as powdered thought
I enjoyed this analogy from a Douglas Hofstader chapter on analogy:
The usual goal of communication is, of course, to set up “the same thought” in the receiver’s brain as is currently taking place in the sender’s brain. The mode by which such replication is attempted is essentially a drastic compression of the complex symbolic dance occurring in the sender’s brain into a temporal chain of sounds or a string of visual signs, which are then absorbed by the receiver’s brain, where, by something like the reverse of said compression — a process that I will here term “just adding water” — a new symbolic dance is launched in the second brain. The human brain at one end drains the water out to produce “powdered food for thought,” and the one at the other end adds the water back, to produce full-fledged food for thought.
Hofstader, D. (2001). Analogy as the core of cognition. In The Analogical Mind: Perspectives from Cognitive Science, Dedre Gentner, Keith J. Holyoak, and Boicho N. Kokinov (eds.). Cambridge MA: The MIT Press/Bradford Book, 2001, pp. 499-538.
Rationality is a property of institutions, not individuals
Here’s Jonathan Haidt on the Rationality Speaking podcast
[quoting himself from Chapter 4 of the Righteous Mind]
“Anyone who values truth should stop worshiping reason.” And what I mean by that is once you see all the psychological evidence on how biased we are, on how people with higher IQs aren't less biased, they're just better at coming up with post hoc reasoning... You look at all the research on how self-interest, or group interests or reputation, biases our reasoning, and how nobody has found a way to de-bias judgements... Nobody's found a way to teach critical thinking that makes people basically rational and unbiased... And so just being realistic, I say that it is a delusion to believe that humans have, or could have this faculty, which New Atheists think that we do, or think that this should be the basis of society. So, I think it's a delusion to think about reasoning that way. But here's the key thing which will bring me back probably to agreement with you: “I'm not saying we should all stop reasoning and going with our gut feelings, I'm saying we must be wary of any individual's ability to reason.” We should see that each individual is like a neuron, very limited, very good at this doing one thing, which is, it sums up the stimulation from the dendrites. And if it reaches a certain threshold at fires and you get a signal down the axon. Each neuron is kind of dumb, it just does one thing. But if you put them together in the right way, you get a brain which can do amazing things. And in the same way, if you put biased, post hoc social creatures like us together in the right way, so that we cancel out each other's confirmation biases, then you get rationality. But you're not going to get it from an individual, or from teaching individuals to be rational. You're going to get it from setting up the right institutions -- a university, or a jury, or an intelligence agency is that sort of institution.
Rules for Reasoning
In Rule for Reasoning (1993) Richard Nisbett argues that American psychology was founded on the assumption that there were no general principles of reasoning which could be taught or trained. From Thorndike to Newell, researchers believed that training on one task wouldn’t generalise to another. The legacy of behaviourism was the belief that training ‘was going to be a bottom-up affair, consisting of little more than slogging though countless stimulus-response associations’ (Nisbett, 1993, p2).
The Europeans, on the other hand, didn’t have such strong antiabstractionist prejudices. And they had Piaget, whose formal operations posited acquiring something akin to abstract principles as core to cognitive development (although Piaget also believed that these principles couldn’t be taught - they were acquired when the child reached the right developmental stage).
The judgement and decision making programme appeared to confirm anti-abstractionism. The typical person failed to apply domain general principles of propositional logic to things like Wason’s selection task (RP #24), and Tversky, Kahneman and others developed numerous other tasks of logic and statistical reasoning which highlighted a tendency for heuristic, rather than principled, problem solving.
One such task from the JDM programme was the maternity ward problem (Tversky & Kahneman, 1972):
There is town with two hospitals, one large and one small. At the large hospital about 60 babies a day are born. At the small hospital about 15 babies are born. At which hospital will there be more days of the year in which 60% or more of the babies born are boys?
The answer is obvious if you understand the (domain general) law of large numbers, but the original report found that 1/3rd of subjects answered “the large hospital”, 1/3rd “the small hospital”, and 1/3rd said “no difference”.
Nisbett, however, reports that when he tried to replicate this result with his students he found that most got it right, and could give descriptions of the law of large numbers which they used to determine the answers. The key factor was whether individual students had taken courses in statistics or not.
This led to the research programme described in Rule for Reasoning, which has the explicitly abstractionist commitment to believing that general principles can be taught, and are taught in disciplines like statistics and economics. While these rules may not be perfectly general, they can applied across domains when people are able to recall a specific instance of the application of the rule or make a domain-bridging analogy. For example, the Wason task is famously difficult - individuals fail to apply the appropriate propositional logic - but can be made more easily solvable if framed in contractual terms of obligation and permission (Cheng & Holyoak, 1985). Similarly, people don’t seem to apply knowledge of statistical processes, such as the law of large numbers, unless they recognise a mechanism in the problem domain as being an example of a random-generator. Randomness is everywhere - from dice to diplomacy - but we have to be culturally trained to recognise it, says Nisbett.
The bulk of the book is experimental reports which back up the claims of the first chapter, looking at some general principles and evidence that training them can produce generalised performance boosts on reasoning tasks.
What general principles? I’ve already mentioned the law of large numbers and social contract schemas (contractual reasoning for the Wason task). Nisbett also points to ‘the confounded variable principle’, causal schemas (e.g. the natural logic of understand an ‘if..then..’ rule) and cost-benefit rules of choice (which includes recognising the sunk cost fallacy).
The programme is optimistic, and progressive. Progressive, because it is the new disciplines of economics and statistics which provide the principles of reasoning, not the maths, latin or logic of classical education. Optimistic, because it says that better reasoning can be taught, and can even be very brief (because abstract instruction can work for conveying abstract rules).
In the thirty years since Nisbett was writing, the strain of anti-abstractionism still seems strong in psychology. The JDM programme has reinforced the view that individual cognition is often fallible, heuristic and unprincipled. General transfer of training effects continue to be alluring, but empirically hard to verify (witness the decade long hype-then-disppointment around working memory training). Nisbett had a vision that we might refine the set of ‘pragmatically useful rule systems’, understand and teach them better and more widely, and so create cultural evolution which raises the reasoning level of all adults. Work related to this vision still occurs across psychology, but the explicit commitment to the vision seems to get more attention outside of psychology, in places like the rationalist/less wrong communities.
Nisbett, R. E. (1993). Reasoning, abstraction, and the prejudices of 20th-century psychology. In Nisbett, R. E. (Ed.), Rules for reasoning (pp. 1–14). Hillsdale, NJ: Lawrence Erlbaum.
The enduring enigma of reason
Catarina Dutilh Novaes’ response to Mercier and Sperber’s Enigma of Reason resonates with some of the theme of Rules for Reasoning. M&S make the case that reasoning is biologically evolved. Novaes wants to emphasise the role of culture in reasoning :
the idea that much of human cognition is a result of enculturation processes and cultural learning has been gaining traction in recent years (Henrich, 2015; Heyes, 2012) and, in fact, has venerable origins in the work of the soviet psychologist Vygotsky (1931). There is much work still to be done on the hypothesis that reasoning is significantly shaped by social interaction and sociocultural environments—that is, that it might be closer to a cultural phenomenon rather than to a biological adaptation
(but no mention of Nisbett)
She uses the biological and cultural forms of language as an analogy:
But what exactly would it mean for reasoning to be tuned by social interaction? What would it mean for reasoning to be a cultural product? To address these questions, it may prove useful to consider one paradigmatic case of a biologically evolved trait in humans and one paradigmatic case of a culturally evolved trait. Spoken language is (most likely) a feature in humans emerging from our genetic endowment as a species, a result of biological evolutionary processes. As such, it is universal (every known human population has a signaling system that we do not hesitate to call a language), and the emergence of the trait in individuals occurs without the need for specific training (though exposure is crucial). Written languages, by contrast, are clearly a cultural product with a very recent history, having emerged independently in just a handful of circumstances (times/places) in world history. Writing skills, unlike speaking, typically require a fair amount of formal training to be mastered, as noted by M&S themselves (pp. 72–73). But of course, any culturally evolved skill, including writing, must latch on to cognitive and physical possibilities determined by our biological endowment, which may then be recycled and reused for novel purposes (as per the ‘neural reuse’ hypothesis (Anderson, 2010; Dehaene, 2009, referred to by M&S).
I submit that reason is closer to written languages than to spoken languages as a cognitive skill: sociocultural environments, including a certain amount of training and exposure to specific linguistic practices, play a significant role in how reasoning skills develop in individuals and how these skills emerged in the history of the human species.
Dutilh Novaes, C. (2018). The enduring enigma of reason. Mind & Language, 33(5), 513-524.
Apologies for the lack of coherence in today’s newsletter. I guess it is a ‘find your own patterns’ edition. Here are some good tweets to finish off with
Which gives me an excuse to link to a recent, and more narratively engaging, newsletter, RP#23 on domestication, trust, trickster spirits and inter-species communication.
And finally, this “plate vending machine” piece of art seems to capture something important about human innovation:
Hi Tom, I am so very pleased to have come across you here. Your writings have been both valuable and inspirational for me. Your argument for the possibility of good reasoning in spite of conditions and qualities that mitigate so heavily against it is for me like a light twinkling at the end of a dark tunnel. Thank you for starting this. I hope you still plan to keep it active for a while, I’m trying to drum up a little more interest among my friends and colleagues.