Groupthink is no match for individual genius

Groupthink is no match for individual genius

By Christopher Caldwell

字号

背景

中文 评论 打印 电邮 收藏

 

If you lock a bunch of high-IQ people in a room and tell them to get on with a simple task, what will they emerge with? Lower IQs, for one thing. A study done by Virginia Tech and a few other institutions, written up over the winter in a publication of the Royal Society, tried to replicate how people think under social pressure. Subjects with an average IQ of 126 were clustered into problem-solving groups and exposed to judgments about their work. A pecking order formed. The low performers showed high responses in the part of the brain that regulates fear. Most of the men became “high performers”, most of the women “low performers”, but no one blossomed. The scientists concluded that “individuals express diminished cognitive capacity in small groups, an effect that is exacerbated by perceived lower status”. In other words, they get dumber.

This confirms common sense. Maybe you can communicate with a slower person by turning off brainpower you have, but you can’t communicate with a cleverer person by creating brainpower you don’t have. Yet this is the first ill word any scientist has had for the way groups think in a very long time. Group intelligence is in vogue. Over the past decade or two, story after story has spoken glowingly of “bandwagon effects” and the “hive mind”, of “memes” and the “wisdom of crowds”. Are these profound new insights or are they a cognitive-science trend on which the tide is now receding?

They are both. There is certainly something measurable that can be called collective intelligence. A fascinating study of its operation was carried out by scientists at Carnegie-Mellon university, MIT and other universities and published in the magazine Science two years ago. The authors started by describing the concept of “g”, or “general intelligence”. The English psychologist Charles Spearman discovered g in 1904, showing that practically all mental tasks are positively correlated. If you’re good at maths, you’re more likely to be a good poet. And since there is an intellectual component to a lot of things we don’t think of as “brainwork”, if you’re a good poet you’re more likely to be a good soldier or a good athlete, too.

The idea that mental talents should be so unfairly meted out in society was disheartening to people’s sense of fairness in Spearman’s age and it is repugnant to egalitarians today. People have spent a century trying to debunk the idea of g, and they have failed. So this unpopular concept has become “arguably, the most replicated result in all of psychology”, as the scientists put it. They therefore had the idea that if they could find a collective equivalent of g – group intelligence correlated across all tasks – they would probably have found group intelligence. They asked small groups to do a variety of mental tests and then play a computer in a game of draughts.

A collective equivalent of g is just what they found. Moreover, it was not just an artefact of the individual intelligences that made up the groups. The correlation of group thinking with the average intelligence of the group, or with the intelligence of the group’s smartest member, was weak. Strong correlations were with the “average social sensitivity of group members, the equality in distribution of conversation turn-taking and the proportion of females”. In plainer English: listening helps. Office bullies and alpha-types who can’t shut up drive down productivity. And there is a benefit to gender diversity. One article commenting on the study speculated that this might be due to women’s readiness to admit when they don’t know something.

These two findings – that there is such a thing as collective intelligence and that working in groups makes individuals a bit duller – are not necessarily contradictory. A human being probably loses a bit of thinking capacity in subordinating himself to a group, no matter what feats the collective is able to carry out. Whether this trade-off is worthwhile depends on what the groups are doing. If western culture as it existed until two decades ago stood for any one thing, it was the defence of the individual against the herd. Individuals produced King Lear and the Discourse on the Method. The “wisdom of crowds” produces a few retail fads at best, book-burnings and pogroms at worst.

Our own time thinks itself different. It is marked by integration of markets and innovations in networking and sales. Crowd-sourced Wikipedia (flawed, quick and free) helped drive Britannica (authoritative, labour-intensive and dear) out of the paper encyclopedia business. No one has the time to read King Lear, let alone write it. Anybody who can spark a retail fad is acclaimed a genius. The wisdom of crowds, in fact, may be just an updated version of the age-old wisdom of retail: when it comes to what the crowd wants, the crowd is omniscient.

The writer is a senior editor at The Weekly Standard