杏吧原创

TODAY, many people value their own judgement on everything from climate change to mass vaccination more highly than that of specialists. A reaction to the reverence for science that typified much of the 20th century was, perhaps, inevitable: the idea of an infallible science is dangerous 鈥� like glass, one chip and it cracks. And when those cracks come into contact with the complexities of climate, health, pollution, economics and social life, just watch that glass shatter.

Even so, because reaching a scientific consensus takes time and practical decisions must be made, we need something more like concrete than glass, perhaps the rougher commodities of expertise and experience.

Little is known of either in any systematic way, so we decided to build a 鈥減eriodic table鈥� of expertise, though we still have many squares to fill in. Our table has an internal structure that makes it more than a hierarchy, with a strong core idea: the idea of tacit knowledge.

The table has 20 categories, divided into five rows, with the first based on 鈥渦biquitous tacit knowledge鈥�, that hard-to-pick-up knowledge you need to live in any society, acquired through practice and shared among us. This includes knowing how to talk fluently, how close to walk to other people, and when it鈥檚 OK to blow your nose with one finger (only on the football pitch).

The table gets more complex by the third row, when we separate 鈥渋nformation鈥�, which you can garner using ubiquitous tacit knowledge, from expertise proper. The first step on this row is 鈥渂eer mat knowledge鈥�. Fine as far as it goes, but you can鈥檛 do much with it except score points in pub or bar quizzes.

The next step is better: the 鈥減opular understanding鈥� gleaned from simple accounts of science and technology. The most complex category of specialist information we call 鈥減rimary source knowledge鈥�: this is acquired by reading research journals and, sometimes, the internet.

But this knowledge is still a long way from real technical expertise. And specialist peer-reviewed literature can contain a lot that is wrong, trivial, or misleading. Even a full-blown peer review system is not discriminating enough to work in the absence of a true specialist 鈥渙ral culture鈥�. Ask any doctor faced by patients armed with the latest 鈥� often meaningless 鈥� medical research findings from the internet.

This is equally true of science, where what might look like a ground-breaking paper is ignored by those with the inside track on what is really going on in the lab, or who know the team has a history of making 鈥渇laky鈥� claims.

鈥淲hat looks like a ground-breaking paper may be ignored by those with the inside track on the lab鈥�

If you spend a lot of time around specialists, you begin to share their oral culture and acquire 鈥渋nteractional expertise鈥�, the first category of genuine technical expertise on the specialist expertise row of our table, grouped under the generic heading of specialist tacit knowledge.

Interactional expertise is the ability to 鈥渟peak鈥� the technical language fluently enough to talk easily with the experts. A lot of our research is into the relationship between interactional expertise and 鈥渃ontributory expertise鈥� 鈥� what you need to know to do practical things as well as talk about them.

We used a forerunner of the Turing test to investigate the two, using colour-blindness and perfect pitch as our case studies. Using a remote computer, a judge asked the volunteers questions and had to then work out who was pretending to be a full-blown expert.

The judges were matched to the subjects: a colour-blind person trying to pass as someone who sees colour would get a colour-perceiving judge, while someone who sees colours who was trying to pass as colour-blind would get judges who were colour-blind. The same went for the perfect-pitch experiment.

The bottom line was that colour-blind people managed to fool the judges relatively easily. But the people pretending to be colour-blind failed to convince those who actually were. When it came to music, people without perfect pitch could not pass as people with it, but those with the ability could pass as people without it.

What the experiment shows is that you can acquire the expertise to speak fluently in a technical area so long as you have been surrounded by people who speak fluently; the colour-blind are surrounded by 鈥渃olour-speakers鈥� whereas the 鈥減itch-blind鈥� are not surrounded by 鈥減erfect pitch-speakers鈥� 鈥� there are just too few of them. The same logic works the other way when the 鈥渢arget鈥� fluency is 鈥渃olour-blind鈥� talk and 鈥減itch-blind鈥� talk.

That interactional expertise should be so powerful is less surprising once we recognise that it is everywhere in science. There could be no big projects if members of one specialism could not talk fluently and with technical proficiency to colleagues in other specialisms, even though they couldn鈥檛 do their work.

The next row of our table concerns meta-expertises, which are used to judge other experts and expertises. There are five kinds, from the least reliable ones, which require minimal immersion in a specialist area, to those requiring most immersion. The fifth and final row deals with qualifications, the best of which measure experience rather than training.

The table has some way to go. So far we have concentrated mostly on interactional expertise, but we have also written papers on primary source knowledge and referred expertise. The first paper is based on a study of Thabo Mbeki鈥檚 intervention on AIDS treatment in South Africa, while the second turns on the skills needed by the managers of big scientific projects, such as the 30-metre telescope or LIGO (the Laser Interferometer Gravitional wave Observatory).

We also want to explain why expertise should be trusted and what kind of experts can be trusted. Crucially, if expertise is based on specialist tacit knowledge and that can only be obtained by long experience in the company of experts, then by definition it is not something the general public can possess.

The Hungarian physical chemist and philosopher of science Michael Polanyi defined tacit knowledge as knowing more than you can tell. The true expert knows not only what is in textbooks and scientific papers but how to distinguish between sound information and the trivial or untrustworthy. That is wisdom born of experience and it can鈥檛 be explained. But what can be explained is who is likely to have how much of it 鈥� and that is what we are trying to work out.