ReTuning the Lute
How opponent processing helps us adapt to a complex world
This week I’ve been stricken by some kind of virus - there’s lots of it going round if you haven’t heard. Rather than afflict you with my fever-brained writing I’m going to repost one of my early posts from back in the day (ie February). I think today is my 6 month anniversary of posting weekly! What better way to celebrate than by recycling some old content. Enjoy.
Siddhartha Gautama was sitting by the bank of a river in India when he heard a fishermen teaching a young boy to tune a lute. The fisherman advised that if you tighten the strings too much they snap, and too little the lute doesn’t make a sound. Siddhartha had abandoned his life as a prince for a life of brutal asceticism in his quest for spiritual enlightenment. After hearing this lute lesson Siddhartha realised that enlightenment lay not on the extremes, but in the correct tuning of the instrument.
This idea become known as the Middle Way, and it had a great impact on Siddhartha - whom we know more often as The Buddha. Whether this story happened or not, the concept of a Middle Way has important resonances and implications for us as we consider cognition and intelligence, and how to build Intelligent Teams.
One question sits at the core of the exploration in this Substack. How do we make sense and act in a complex, changing, and infinitely data-rich world? Today I’m going to introduce Opponent Processing Theory and try to situate it in the context of our meaning making, both as individuals and as teams. Hopefully it will lead to you being better at tuning your lutes and your teams.
Mo problems mo problems
When it comes to problem solving we face a significant obstacle. There are infinite ways to get from a starting point to a goal state. Nonetheless we somehow manage to find a way to search through those infinite options and use the toaster, the oven, the frying pan or whichever way works best for us.
What ‘best’ means varies from situation to situation. The best way to make toast when you’re camping or in a power cut will be different from if you’re at home with the electricity working. Making toast for my gluten intolerant friend might involve a trip to the shops, and if I’m in a hurry the best way will be the quickest that doesn’t burn the house down. We don’t mentally list every consideration in decisions like this - that list would be unlimited. Instead we use a combination of mental shortcuts, bodily emotional intuitions and rational logic to narrow down our options and work our way through the problem. (see Antonio Damasio’s excellent Descartes' Error on this).
Our problem are ill-defined, the list of potential unintended side effects is unlimited and we need to make decisions quickly and effectively with limited cognitive and biological resources. How do we do this? We’ve looked a little at how Predictive Processing is one answer to this question; iteratively, quickly, and using fast and cheap predictions which we can easily update. I’d like to explore another answer in this post.
Staying Relevant
University of Toronto’s John Vervaeke argues that the idea of relevance is central here. Good cognition is the thinker’s ability to consider the relevant factors from an environment. When considering how to make toast, I need to quickly work out what to consider important from an infinitely long list of potential considerations.
Is the weather relevant to making toast? Well if I’m camping it probably is, but if I’m in my house it probably isn’t. Unless there’s a crazy hurricane or a heatwave. The weather might be a relevant factor, it depends. What does it depend on? That also depends. We face a ‘combinatorial explosion’ of possibilities of what to consider and how to choose what to consider with pretty much every decision we face. The body needs a way to make this quick and easy so we can actually do what we need to do to survive.
Vervaeke argues that we use opponent processing to find relevance. We put two heuristics (rules of thumb for fast decision making) in opposition to each other and flexibly move between the two extremes. We realise and use relevance when we find the right tuning in that opponent processing that achieves our goal whilst being an overall efficient use of our cognitive resources.
One example is the autonomic nervous system. We have in us two opposing nervous systems - sympathetic (fight and flight) and parasympathetic (rest and digest). The sympathetic nervous system is trying to increase your heart rate, make you more alert to danger, and get your muscles ready to run or fight. If you’re in physical danger, the more sympathetic response will be the relevant one. I don’t want to be caught napping when the tiger arrives.
On the other hand this is an energetically costly approach. If I’m always in sympathetic mode I might flinch and prepare for battle at the rustle of every leaf just in case it’s a tiger and I’ll never get anything done.
Opposing this is the parasympathetic nervous system, pulling my heart rate and breathing down, trying to make me relaxed and restful. This will let me get stuff done and digest my meal and conserve energy. If I’m safe this will be the relevant mode to use. It is on the other hand the wrong approach if I’m in danger and should be running away. There is no ‘correct’ place between parasympathetic and sympathetic extremes that we should always aim to occupy. Rather as our environment changes, our body intelligently fluctuates in small and large ways between these two modes to find the balance that is relevant to the environment.
Another example is explore versus exploit. Explore is about finding new resources, new ideas, new options. It’s a resilient strategy that hedges our bets and makes it more likely that we’ll find what we need, even if it’s not very efficient. If I don’t like my job any more and I’m not getting anything out of it, maybe it’s time to go exploring and update my CV.
Exploit is about staying put and maximising use of the existing resources in front of us. It’s an efficient strategy. Why go find a new job when I’m learning loads in my existing one? I’d be better off spending my time and attention exploiting what I already have.
When I’m making toast I can either explore and spend hours looking for new methods of making toast and get into the weeds of all different types of toaster and bread, or I can exploit my tried and tested methods and use the toaster in my kitchen and bread in my bread basket.
If I’m trying to win a toast making competition, maybe I should explore more, and if I’m just trying to have breakfast before work I should focus on exploit. Like with the autonomic nervous system there’s no ‘correct’ answer between explore and exploit, I trust my body’s unconscious, self organising machinery to find the relevant balance in this opponent process, given the relevant constraints. When everything is working well, my body, emotions, intuitions and experiences help me self-organise into finding the right things relevant so I don’t waste time researching toasters when I’m already late for work.
Opponent Processing in your team
If we take the principles behind opponent processing and apply them at a team level we come up with some interesting and counterintuitive suggestions. Much of the conversation that I witness is about finding the right framework, or the ‘correct’ way to implement scrum. Based on our understanding now of relevance realisation and opponent processing we know that a one size fits all approach isn’t going to be best.
Different teams are in different situation, requiring different balances between opposing poles. Let me describe two teams for you.
Team Elixir is a 4 person startup, building an app to help people sharpen their knives with the help of AI (can you tell I’m writing this from my kitchen?). Their technology is untested and they don’t have any immediate competitors in the home-knife sharpening app space. They don’t know of anyone who has paid for a similar service like this before and are still looking for who their potential users are.
Team Fantasma works in a giant telecomms team. They are trying to migrate one of the ancient databases to a slightly less ancient server cluster. These databases contain reams of sensitive customer data. They are tightly integrated with five other teams working on different parts the system.
Do you really think both these teams will use Scrum exactly the same way? Or use it at all? Or that they both would *definitely* benefit from Kanban?
How could these teams choose how to use Scrum, or which bits they need to change?
I’m proposing here that opponent processing could give us a generic model for helping teams think about how to change their practices. We ask teams to think about the two opposing extremes governing their decisions and work out if they are well situated on the spectrum.
In this way we can do a kind of conscious opponent processing. Just as the brain self-organises neuronally to find the right balances in opponent processes for each situation, we can help our teams self organise to experiment with different positioning on the extremes.
We can offer some of the same constraints from the cognitive science literature as starting points. John Vervaeke describes a few opponent processes from which general relevance realisation (and cognition) arises.
General Purpose vs Special Purpose Are we optimised to solve general problems (a hand) or are we optimised for particular purposes (a screwdriver)
Exploiting vs Exploring Should we be looking to make the most out of known and existing resources (stay in the old job) or should we look for new resources (get a new job)
Compression vs Particularisation Should we be compressing lots of data down to a simple representation of that data drawing a line of best fit (giant surveys) or should be looking at the particular stories behind each bit of data (customer interviewing)
I believe that if we help bring this logic to our teams we can help them find the right practices based on the relevant factors to them and their circumstances. We should be enabling teams to experience and influence their opponent processing to make sure that they are finding the best available ways to solve their problems.
Maybe Team Elixir will realise that they are over compressing their data and missing the opportunity to learn about particular customer experiences. Instead relying on surveys.
Perhaps Team Fantasma will remove some Lean User Experience practices designed for a more general purpose team and optimise more to follow some good particularist industry patterns for moving servers.
When they realise they are in the wrong tuning in the opponent process they have a rational sense-making tool to help them pick their next team working experiment, which is to try something they think will move them closer to a desirable extreme.
Like the Buddha understood, the answer is a Middle Way. We don’t want to tune too loose or too tight. We need to work out the right tunings with our teams.