Evolution made us the ultimate learning machines, and the ultimate learning machines need to be oiled by curiosity.
We humans have a deeply curious nature, and more often than not it is about the minor tittle-tattle in our lives. Our curiosity has us doing utterly unproductive things like reading news about people we will never meet, learning topics we will never have use for, or exploring places we will never come back to. We just love to know the answers to things, even if there’s no obvious benefit.
From the perspective of evolution this appears to be something of a mystery. We associate evolution with ‘survival-of-the-fittest’ traits that support the essentials of day-to-day survival and reproduction. So why did we evolve to waste so much time? Shouldn’t evolution have selected for a species which was – you know – a bit more focussed?
The roots of our peculiar curiosity can be linked to a trait of the human species call neoteny. This is a term from evolutionary theory that means the “retention of juvenile characteristics”. It means that as a species we are more child-like than other mammals. Being relatively hairless is one physical example. A large brain relative to body size is another. Our lifelong curiosity and playfulness is a behavioural characteristic of neoteny.
Neoteny is a short-cut taken by evolution– a route that brings about a whole bundle of changes in one go, rather than selecting for them one by one. Evolution, by making us a more juvenile species, has made us weaker than our primate cousins, but it has also given us our child’s curiosity, our capacity to learn and our deep sense of attachment to each other.
And of course the lifelong capacity to learn is the reason why neoteny has worked so well for our species. Our extended childhood means we can absorb so much more from our environment, including our shared culture. Even in adulthood we can pick up new ways of doing things and new ways of thinking, allowing us to adapt to new circumstances.
In the world of artificial intelligence, computer scientists have explored how behaviour evolves when guided by different learning algorithms. An important result is that even the best learning algorithms fall down if they are not encouraged to explore a little. Without a little something to distract them from what they should be doing, these algorithms get stuck in a rut, relying on the same responses time and time again.
Computer scientists have learnt to adjust how these algorithms rate different possible actions with an ‘exploration bonus’ – that is, a reward just for trying something new. Weighted like this, the algorithms then occasionally leave the beaten track to explore. These exploratory actions cost them some opportunities, but leave them better off in the long run because they’ve gain knowledge about what they might do, even if it didn’t benefit them immediately.
The implication for the evolution of our own brain is clear. Curiosity is nature’s built-in exploration bonus. We’re evolved to leave the beaten track, to try things out, to get distracted and generally look like we’re wasting time. Maybe we are wasting time today, but the learning algorithms in our brain know that something we learnt by chance today will come in useful tomorrow.
Obviously it would be best if we knew what we needed to know, and just concentrated on that. Fortunately, in a complex world it is impossible to know what might be useful in the future. And thank goodness – otherwise we would have evolved to be a deadly-boring species which never wanted to get lost, never tried things to just see what happened or did things for the hell of it.
Evolution made us the ultimate learning machines, and the ultimate learning machines need a healthy dash of curiosity to help us take full advantage of this learning capacity.
Or, as Kurt Vonnegut said, “We are here on Earth to fart around. Don’t let anybody tell you any different.”