Between the lines

Failure is a beginning

Professionally speaking, biologist Stuart Firestein is a fan of failure. Done right, he argues, failures in science are much more than mere mistakes, and better even than the “painful but character-building” sort of failure described in self-help and business books. In Failure: Why Science is So Successful, Firestein explores the complex role that failure plays (or should play) in how science is done, taught, perceived and funded, focusing on problems that transcend disciplinary boundaries. The best example of his cross-disciplinary reach is probably the section on education. Here, the author decries what he calls “the bulimic model of science education”, in which teachers cram “gobs of facts into [students’] heads so that they can puke them back up on an exam, then move on to the next unit with no measurable gain”. Few teachers (or students) will dispute that stomach-churning critique, but Firestein isn’t finished. In his view, the source of the problem isn’t just the current vogue for targets and tests: it’s the scientific method, at least as it is commonly understood. The idea that science proceeds by making observations, formulating hypotheses and testing them sounds great, he writes, “except that no scientist that I know of actually follows this prescription”. Telling a science student to form a hypothesis, he explains, is “like giving an art student a brush and the direction to ‘do painting’ ”. Much better, he argues, to teach that science begins with curiosity and proceeds through a series of failures. Firestein’s perspective on the so-called “replication crisis” in science (see feature “No result, no problem?”, May 2016 print and digital magazine only) is similarly unorthodox. While he acknowledges that many published papers contain irreproducible results, in his view this is only natural in a discipline where failure is both common and necessary. Moreover, some of the measures proposed to remedy the situation would, he argues, be counter-productive. Deeply thought-provoking and (as this review shows) frequently quotable, Firestein’s book is nevertheless better at diagnosing problems than at offering solutions. The author’s remedy for overly cautious grant-funding procedures, for example, amounts to skewing application criteria in favour of “the merit of their science and the creativity of their approach”. A good idea, of course, but one does not need to be a bureaucrat to see that “merit” and “creativity” are extremely hard to define. In this sense, one could call Failure a failure – but only in the best possible sense of the word.

• 2016 Oxford University Press £14.99/$21.95hb 304pp

Seismic shifts in history

In 132 AD the Chinese court mathematician and astronomer Zhang Heng ordered the construction of a curious instrument. The device (now sadly lost) incorporated eight dragon heads, eight squatting toads and a set of heavy brass balls. Its purpose was earthquake detection: distant tremors would cause one or more balls to fall from the dragons into the open mouths of the toads, creating a noise that would alert the authorities and (in theory) indicate the direction of the shaking. The story of Heng’s seismometer is one of many gems in Andrew Robinson’s book Earth-Shattering Events: Earthquakes, Nations and Civilization. The book is organized as a series of case studies on specific earthquakes, from the 1755 shock that devastated Lisbon, Portugal to the 2011 Tōhoku earthquake off Japan’s north-east coast. In each case, Robinson – a veteran author equally at home with history and science – draws on previously published accounts to summarize what happened, and to analyse the earthquake’s effect on the country that experienced it. Individually, these case studies make vivid reading, and collectively they tell a fascinating story about how seismology developed as a science. However, the author’s broader goal is to explain how earthquakes have influenced the course of human history, and occasionally, this ambitious effort leads him – if you’ll pardon the pun – onto shakier ground. One of Robinson’s arguments is that historians and archaeologists tend to ignore or downplay the role that earthquakes play in initiating political change. Yet as his own case studies show, even when major tremors do spark significant changes (and they often don’t), the nature of those changes is highly dependent on long-term sociopolitical trends. With such weak evidence of causality and few clear rules, it’s hardly surprising that historians prefer to focus on other things (including, of course, long-term sociopolitical trends). The bottom line is that if predicting earthquakes is hard, predicting their effects on the grand sweep of history is nigh-on impossible – although as this very entertaining book shows, it can still be interesting to try.

• 2016 Thames & Hudson £18.95hb 256pp