How do I get to know you?

Until some time in the last century, if you were wandering in the highlands of New Guinea, and you bumped into someone else, it would be really important to have a conversation.

The purpose of this conversation would be to avoid having to kill one another because you were from different tribes: the topic of the conversation would be familial relations that you have in common. So long as you have family in common, you don’t have to kill each other.*

Humans, like other animals, use sounds to establish group identity, select mates, protect territory (for example when an academic writes down their sounds and publishes it in a journal).

If you live in a modern society, though, it is unlikely that you will be able to establish familial connections through conversation, so we have developed a system of groupings based on shared concerns. When we meet, we try to establish shared group membership through identifying common concerns.

Often this means starting with common, culturally normal concerns. So, for example, we remark on the weather. Or a man might say ‘did you see the match last night?’. Once a general level of affective coherence is established (i.e. there is something in common that we care about) we have opened up a conversational domain, and we can use this to deepen our sense of group membership – for example by talking about players or strategies.

Another example: ‘did you see Love Island last night?’ might be followed by ‘Oh my god! Can you believe what happened with [X and Y]!?’ Notice how it is important to express strong emotions if the process is to work well – just saying ‘yeah, I did.’ would risk decoherence. If decoherence happens, the implication is that we are members of competing groups and – just like our New Guinea tribesmen – we try to avoid this. So the response ‘no – I don’t watch Love Island’ might be followed up with ‘are you on Netflix, then?’ in an effort to find a new ingroup/affective domain.

If we cannot find common ground, the conversation will peter out uncomfortably and we will likely avoid one another and more prone to say negative things about one another ‘I just find them a bit stuck up…’. Whilst we are unlikely to actually kill each other, we will compete in other, subtle, ways.

This all happens because, as complicated social creatures, we group according to the things we care about.

But what about education?

Humans encode experience according to their concerns, i.e. with respect to affective significance (Affective Context Model). In other words, what we care about governs our reactions, which determine what we remember. So it’s vital that if we are to learn effectively, we understand what matters to a person. So – how do I get to know you?

Often the old ways are the best ways – and at the very least we are silly to overlook methodologies that humans have developed during the course of millennia. A conversation is the tool that humans have developed to establish shared concerns. As part of a conversation, storytelling is used to reveal what a person cares about. When talking about a night out, the things that people highlight – to which they attach emotional significance – will tell you what they care about. So a good place to start an educational process would be a conversation and stories that a person tells. This is why coaching and mentoring continue to be popular.

But there’s a problem. This process is coloured by the desire to establish coherence; as such it is reciprocal in nature – the educator must tell their stories, the whole thing proceeds much like a dance. And of course it cannot be that every educator cares about every thing that a student cares about, and to the same degree. Of course without establishing some connection to what the student cares about, education cannot proceed – but if we try to imagine a digital education, one in which concerns are mapped and experiences provided in line with those concerns, then this starts to look like a mismatched process.

‘What about data?’ some people will say. ‘Can’t we use big data to establish patterns of concerns?’. The problem with this approach is that big data is invariably a proxy for concerns: to see this, consider what we know about someone if we know that they have purchased a number of Yves Saint Laurent bags. If we look at aggregate purchasing data we can see that lots of people in a given culture are buying this bag. Does this tell us that Yves Saint Laurent is a shared concern? Not really – because we don’t know what the underlying concern is that is driving this behaviour. Now you and I might reasonably posit that people are buying the bag because of the social status that it accords – but see how we have introduced a theoretical layer beneath the data layer. We can develop a sound predictive model using big data, but this will just be a statistical model. It will have its limitations, because in the absence of a theoretical basis, we won't necessarily know what data to collect.

In an educational context, we might notice that a child loves playing with Lego – but we can’t necessarily conclude that this is because they love all things mechanical. Perhaps their friends also play with Lego. Perhaps they only like ‘Harry Potter’ Lego. Perhaps it is because they like animals, so they build lego zoos. How would we establish what is driving the observable behaviour? Possibly we would need to have a conversation.

This may sound familiar if you know Clayton Christensen’s ‘Jobs to be done’ approach – in order to really understand the observed behaviour, we have to understand the concern that is underpinning it. 

To conclude, though, it’s important for us to tackle this problem – the problem of how to quickly and accurately map concerns, to a high degree of specificity, since this is the starting point for the design of any learning system. We should pause to reflect on the ways that we do that today – through conversation – and on the difficulty in developing a model that will provide a framework for the totality of concerns expressed within a culture (i.e. the extent to which this depends on a ‘Theory of Concerns’). Where would such a framework come from? Comparative psychology or Social Evolutionary Psychology might provide a starting point, but are likely to lead us astray. Looking for patterns in big data will give us some clues; but today we still need to talk to people to make sense of what we see. Of course we can overlook the theoretical model entirely and adopt a data-driven ‘people like you liked this’ model, but this will implicitly reinforce norms and drive a kind of toxic coherence in my view.

*From Jared Diamond’s ‘The World Until Yesterday’

Picture: Mahrael Boutros