an autodidact meets a dilettante…

Confirmation bias (and the benefits of social reasoning) in a nutshell:

How can you say to your brother, ‘Brother, let me take the speck out of your eye,’ when you yourself fail to see the plank in your own eye? Luke 6: 42 New International Version

The argumentative theory of reason

The recent New Scientist collection, Being Human, includes an essay, ‘The argumentative ape’, by Dan Jones, which is worth reading and contemplating for any teacher involved in encouraging her students to think richly about current ethical or political issues. In my college, NESB students study ‘English for academic purposes’, which involves a lot of basic grammar and vocabulary at the lower levels, and academic presentations and essays at the higher levels. In these ‘discussion’ or ‘argument’ essays and presentations, students are required to examine the pros and cons of some chosen activity or decision, such as the proper driving age, the consumption of GM food, or even whether humanity has benefitted or blighted our planet.

However, there seems to be a contradiction in asking students to write, and be examined on, individually written ‘discussion essays’, when discussions and arguments are group rather than individual activities. More importantly, if we want to improve our students’ understanding of current issues, perhaps we should be placing more emphasis on group discussion than on individual analysis.

This is hardly a new idea. The ancient Athenians, founders of democracy – decision-making by the people – built their city around the agora, a gathering place for public talk and argument. This design was quite deliberate, as all Athenian citizens were required to contribute to discussions from which civic decisions were made.

‘The argumentative ape’, however, provides contemporary evidence about the evolutionary importance of argument in human society. It describes a thesis put forward by European researchers Hugo Mercier and Dan Sperber, that human reason evolved not so much to assist us in more clearly understanding our world, but to argue, to persuade, to convince others of our position, our right-ness. So, it evolved socially. And there appears to be some evidence for the more general ‘social brain’ hypothesis, in that a clear correlation has been found between the number of individuals in a primate group and the average brain size of that particular species.

Now one essential problem here should be obvious, as it was to Socrates in his battle with the sophists. The most persuasive arguments aren’t necessarily the best. So it’s natural that along with persuasiveness, skepticism would have developed, as humans sought to evaluate competing arguments.

Through scepticism we’ve identified many types of fallacious reasoning, and ways we have of convincing ourselves in the process of trying to convince others. Confirmation bias, or motivated reasoning, probably tops this list, as it is extremely pervasive if not universal. As Mercier points out, using confirmation bias seems counter-productive if you wish to arrive at correct results, for example in scientific research, but it can be highly effective in argument, as your bias commits you to garnering a multitude of arguments for your position while ignoring, and thus rendering insignificant, all arguments against. If we accept an argumentative theory of the evolution of reason, then, we will see confirmation bias not as a flaw, but as a device to strengthen our own arguments, and the ability to detect such biases would in turn be a device to undo or diminish the arguments of others.

how individual reasoning is affected by the larger group

Experimental psychologists have found many ways in which our reasoning can be affected or manipulated. Take, for example, the framing effect. It has been found, and regularly confirmed, that how the same problem is worded will affect our decision. Jones presents the scenario, used by psychologists, of a small village of 600 people threatened by a deadly disease. In scenario one, if Plan A is adopted, exactly 200 people will survive. If plan B is adopted, there will be a 1 in 3 chance that all will survive, and a 2 in 3 chance that none will survive. When this scenario is presented to subjects, the majority invariably choose Plan A. However when, in scenario two, Plan A is framed with the slight difference that exactly 400 people will die (with no change to Plan B), this is enough for the majority to flip over to Plan B. This consistent result has been explained in terms of ‘loss aversion’ – we prefer to avoid the explicit loss of life as expressed in the change to Plan A in scenario two. Significantly though for the argumentative ape hypothesis, this loss aversion bias is strengthened when we have to justify our decision to a larger group. We have a ready-made justification as expressed in the framing. It’s probable that we always have in mind what the larger group, or ‘society’ will think of our decision, but when this need to justify ourselves is made explicit, the ready rationalisation is more likely to be adopted.

Other effects of apparently faulty reasoning, such as the attraction effect and the sunk-cost fallacy, have been detected in psychological studies, and all have been shown to be enhanced when there is an explicit need for justification. The ‘argumentative’ thesis claims that we tend to choose the most easily justified option rather than what might be best.

Confirmation bias for me, scepticism towards you, and how it pans out

While this may seem a pessimistic outlook on our use of reason, the counterbalance lies in our ability, from clear evolutionary need, to identify and so counter the faulty arguments of others. This pattern follows a familiar evolutionary trajectory, in which a predator evolves a means to capture its prey, leading the prey to develop a defence mechanism to protect itself against the predator. Scepticism helps us to avoid being sucked in and ‘devoured’.

The result for group reasoning is that bias and the scepticism can balance each other out, leading to a greater recognition of the weaknesses in our own opinions and the strengths in those of others. And experimental evidence backs up this result. To quote from Jones’ article:

In one convincing study, psychologists David Moshman and Molly Geil at the University of Nebraska-Lincoln looked at performance in the Wason selection test – a simple card game based on logical deduction. When thinking about this task on their own, less than 10 per cent of people got the right answer. When groups of 5 or 6 people tackled it, however, 75 per cent of the groups eventually succeeded. Crucially for the argumentative theory, this was not simply down to smart people imposing the correct answer on the rest of the group: even groups whose members had all previously failed the test were able to come to the correct solution by formulating ideas and revising them in light of criticism (Thinking and Reasoning, vol 4, p 231).

He also points to research indicating that groups are more creative in their thinking than individuals (see sources below).

Implications for teaching, or how to best facilitate the best group thinking

Evidence from a series of studies by Anita Williams Woolley of Carnegie Mellon University in Pennsylvania suggests that a group’s individual skills are not the best predictor of the group’s overall performance in problem-solving. These studies were designed to measure the ‘collective intelligence’ of the group, in something like the manner of IQ tests for individuals. She found that those groups who scored highest were the most inclusive, allowing maximal participation within the group. Sensitivity to the moods and feelings of others helped groups to score highly, and the best groups were those with the greater number of female members, presumably because females have a greater social sensitivity.

Group thinking can, of course, backfire. Groupthink in fact has long been seen negatively, but this is because people with the same cognitive biases often congregate together, as with political parties and religious organisations, or gravitate towards similar professions, such as the police or the military. In such groupings it’s often the case that the group moves collectively to quite extreme positions. Where group thinking would be expected to work most effectively is precisely in a college for NESB students from different cultures and backgrounds, in which individuals are challenged by widely different but (hopefully!) cogent opinions.

As educators, we need to consider the best outcomes for our students. Clearly there is pressure, in an individualised results-based system, to push for individual skill in argumentation, with the resultant high test scores. However, the evidence for group interaction in improving students’ understanding of the many issues focused on in essays and seminars at the higher levels is clear. Of course the situation is complicated by the fact that many students at EAP2 and EAP3 levels still don’t have the  grammatical and lexical skills to present cogent arguments in English, so that it’s often hard to determine whether their difficulties are those of reasoning or of language. Even so, I believe it is vital to take advantage of the cultural diversity of students’ experience and knowledge (even within identical language groups) to encourage interaction that will challenge biases and create awareness of a variety of perspectives. Hopefully this will enliven their thinking both within the college and in their studies beyond Eynesbury.

Some sources are found in the links. Here are others.

D Sperber and H Mercier,”Why do humans reason? Arguments for an argumentative theory”, Behavioural and brain sciences: Published online March 2011. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1698090

http://edge.org/conversation/hugo_mercier-the-argumentative-theory. Mercer elaborates on the theory very interestingly in a video on this website

Williams Woolley, Anita, ‘Collective intelligence in human groups’, April 2012: Center for Collective Intelligence: http://cci.mit.edu/ci2012/plenaries/speaker%20slides%20ci%202012/Woolleyslidesci2012.pdf

D Moshman & M Geil, 1998 ‘Collaborative reasoning: evidence for collective rationality’. Thinking and reasoning V4 issue 3: http://www.tandfonline.com/doi/abs/10.1080/135467898394148