Both consolidating and disruptive work are needed for scientific progress, of course, but science now seems to favour the former over the latter in a potentially unhealthy way. Mr Park and Drs Leahey and Funk found that the average cd score for papers has fallen by between 92% and 100% since 1945 (see chart), and for patents between 79% and 92%. These declines are not mere artefacts of changing publication, citation or authorship practices; the researchers controlled for that. Why, then, has science become less disruptive?

One hypothesis is the low-hanging-fruit theory—that all the easy findings have been plucked from the branches of the tree of knowledge. If true, this would predict different fields would have different rates of decline in disruption, given that they are at different stages of maturity. But that is not the case. The decline the researchers found was comparable in all big fields of science and technology.

Another idea is that the decline in disruptiveness stems from one in the quality of published work. To test this, the researchers looked at two specific categories: papers in premier publications and Nobel-prizewinning discoveries. “If there were a pocket of science where the quality might have declined less, or hasn’t declined,” said Mr Park, “it would be in those places.” But the downward trend persisted there, too.

A more likely reason for the change, the researchers argue, is that scientists and inventors are producing work based on narrower foundations. They found that citing older work, citing one’s own work, and citing less diverse work all correlate with less disruption. As the amount of published science grows, the effort required to master a pool of knowledge that is both deepening and narrowing as the years roll by may inhibit the ability to form creative connections between disparate fields. Here is an argument for the rebirth of the renaissance human.