There will be many inputs in this series with a title like changing the way we think about outcomes. Pinker’s work is just one of many reasons to rip up the rule book on measuring impact.

A lot of my life has been spent counting up 1s and 0s. A child measures positive for a conduct disorder. Mark it 1. A child is dyslexic. Another 1. An addict. 1. A homeless person. A lonely person. Each a 1.

Then there is that part of my life spent trying to turn 1s into 0s. A student with poor social and emotional regulation participates in a programme and moves from 1 to 0. The good stories end with more 0s than 1s.

Some people are stuck with a 1. A treatment can help manage the symptoms of schizophrenia but it will never turn a 1 into a 0. An addict is told he is a 1 for life, even if he never takes another drink or drug.

Limitations on what science could do encouraged us to measure whether people were a 1 or a 0 before an intervention, and then again afterwards. (And ideally repeating the process with a control group).

But what if the idea of better angels of our nature is correct, that for the majority of challenges to health and development our ability to cope is contextual. Could it be that we are 1 today, and maybe tomorrow, but on the day after we are zero, and we stay like that for a few days, before becoming a 1 again? That we are mostly a continuous series of 1s and 0s over time?

That would require us to think about outcomes and measurement differently. And to do that, we would need to be unshackled from the limitations of science.

The ongoing revolution that has seen machines take a stronger role in learning may make that possible.