Our brains hate surprises
I'm doing a deep delve into Robin Carhart-Harris's research at the moment (an awesome psychopharmacologist at Imperial). More on him in later posts no doubt, but through reading his work I came across Karl Friston's free energy hypothesis, which provides a nice little springboard for a wildly extrapolated train of thought.
Friston's paper (cited below) is full of clever equations and fancy stats concepts which I don't understand but I think I get the gist.
Friston (and many others) think that the point of the brain is to use sensory inputs to generate a model of our environment so that we can predict what will happen next and act accordingly. So far, so obvious.
The cool leap he’s made is to frame the goal of the brain as being to ‘minimise free energy’. Free energy is basically a stats term for ‘surprise’ which in this context means prediction error: the difference between what we perceive to be happening and what we predicted would happen. The brain wants to be an accurate prediction machine in order that we make the right choices, so being surprised by events (e.g. the unexpected presence of a sabretoothed tiger) is a failure.
So why is this cool?
Well, it’s cool because it then follows that brains have three strategies open to them to minimise free energy - and which strategy the brain is using can explain a lot of behaviours which are very familiar to us.
If I’m a brain looking to avoid being surprised, I can:
Make a good model - gather lots of data about the world and build an accurate predictive model of it so my predictions match the way the world is most of the time.
Change the world - act on the world in order to bring it in line with my predictions.
See what I want to see - attend to only those perceptions which confirm my model of the world, rejecting the others, thereby minimising surprise.
I think we employ different balances of these strategies as we develop.
A baby starts with little or no model of the external world, so can’t make useful predictions. Anyone who has watched a baby will know that they are voracious gatherers of information - tasting, prodding, and staring at anything and everything. This makes sense - the baby’s brain is desperate to suck in as much information about the world as possible so it can start building a model of how the world works and begin making predictions in order to stop being so bloody surprised the whole time. Babies are essentially forced to use only strategy 1.
As we get older, we have a better model of the world and can start making some useful predictions. This then opens up the options of using strategies 2 and 3.
It’s a big leap from Friston’s paper to what I’m going to say next - his paper is strictly about information processing in the brain - but this got me thinking: what might the different strategies ‘look like’ to an outside observer?
Strategy 1: the Scientist
This is a person whose model of the world is ‘unlocked’ - i.e. open for revision. They will be as hungry for new information as a baby, and as prepared to revise their assumptions about how the world works. They are happy to have their views challenged and to entertain the possibility of new ideas. Action is primarily aimed at attaining new information about the world.
Scientists, philosophers, writers, and artists may do their best work in this mode.
Strategy 2: the Activist
This person’s model of the world is ‘locked’. They recognise that the world is not always the way they think it should be, and are prepared to change that. They will welcome new information but only insofar as it will help them to target their actions in order to remake their world as they see fit. Action is primarily aimed at remaking the world to better fit their model.
Activists, entrepreneurs, and early-stage tyrants might be most effective in this mode.
Strategy 3: the Zealot
This person’s model of the world is also ‘locked’. Unlike the activist, they do not admit the possibility that they might be wrong about the world. They will judge information which fits their model of the world to be ‘correct’, and discard any conflicting information as ‘wrong’. Action is primarily aimed at avoiding any information which doesn’t fit their existing model.
Zealots of any stripe use this mode: from religious fundamentalists, to left-wing diehards, to tyrants who ignore or suppress conflicting data (and are therefore less successful that tyrants employing Strategy 2, e.g. Pol Pot, who suppressed data suggesting people weren’t meeting their agricultural targets, and as a result millions of his subjects died of starvation).
In practice, of course, very few people will be entirely operating in one mode or another - mostly they will use a blend - sometimes gathering more information, sometimes acting to change their world, sometimes burying their head in the sand / running away. They might also operate in different strategies in different domains - e.g. I think I’m a Scientist on literature (open to pretty much anything), an Activist on the environment, and (though I’m trying to change) a Zealot on Brexit.
This is a simplistic model, obviously - but it sort of makes sense. As people get older and their models of the world become more developed, they often tend to shift towards strategies 2 and 3. We all could probably do with a little more strategy 1 in our lives…
Friston’s paper: Friston, Karl. "The free-energy principle: a unified brain theory?." Nature reviews neuroscience 11.2 (2010): 127.