Empathy
Some years ago I realised that I have never understood hunger.
This was a slightly confusing revelation, since I had always assumed that I didn't understand hunger. How could I? I was raised in a comfortable, middle class household in the UK. Struggling to afford food was never a concern that neither I, nor anybody I had any knowledge of, had ever experienced. I always assumed my own experience of hunger to be a shadow of the true meaning, one which I would - hopefully - never live to really understand.
It was something of a revelation to understand that, while the above is undoubtedly true, I had also failed to understand hunger in a far more pedestrian manner.
For me, hunger has always been a very transient concern. It's not that I don't feel hungry, but it has about the same effect as feeling mildly cold or horny. I'm aware of it and it's a bit distracting, but not really unpleasant, and if I focus on something else then it rapidly goes away. On a scale of irritants it ranks a little bit below noticing a stain on my trousers and well below having an itchy foot.
When others described being hungry I always took it as interesting information, but not something particularly time sensitive or calling to action. Often it might seem superfluous - if we already had a plan to eat later, for example - and so I might wonder why somebody mentioned it. When somebody made a fuss about being hungry my instinctive reaction was to think of them as being pampered and a little pathetic, perhaps. They - like I - had no real comprehension of hunger, and making a big issue of not eating for a few hours seemed melodramatic at best.
It took some years of - undoubtedly rather frustrating - persuasion from my partner, and others, to convince me that other people's perception of hunger was somewhat different to mine. That, for many, even the "trivial" hunger experienced by those of us in the affluent west was still a very salient concern - triggering profound changes in mood, concentration, perception. To them, hunger represented a serious problem that needed to be solved relatively soon, or else suffer a significant drop in life quality.
Aside from the simple fact of being wrong about this, I find this revelation to be interesting for a couple of reasons:
I had had considerable experience of dealing with people experiencing feelings of hunger, their reactions and subsequent behaviours. Few of these people were otherwise prone to melodrama or to exaggerating problems. A simple application of Occam's razor would have suggested that the obvious explanation for their behaviour - of seeming upset and irritable - would be that they are currently experiencing upset and irritation.
Even having accepted this state of things, I still find it difficult to actually internalise this acceptance. I still cannot really conceive of feeling hunger to the extent that it fundamentally alters my mood; rather, I have managed to code an override that says "if you think somebody is being melodramatic about being hungry, ignore that and treat it as if they're experiencing a genuine problem".
Lately, I have been thinking about empathy, what it means and how to practice it.
An often used expression in English is that to practice empathy is to put yourself into somebody else's shoes. To see somebody in a situation and to imagine yourself going through the same experience and feelings. But I have lately realised that this expression is, for me at least, actively unhelpful. I have always found it relatively easy to put myself in somebody else's shoes - to call to mind the relevant context, the situation they are in, and to imagine how I would feel. And, all too often, the conclusion has been that the way they are acting seems to be incongruent with the situation.
Rather, I feel like the more accurate expression would be "to imagine somebody else in their own shoes". Which is somewhat less pithy, alas. Modeling myself in another situation seems a relatively straightforward exercise. It turns out that modeling somebody else is - perhaps unsurprisingly - rather more difficult. What's more, it feels like there is a step beyond modelling being discussed here.
There's an old joke where a mathematician, a physicist and an engineer are asked to measure the volume of a red rubber ball. The mathematician uses a pair of calipers to measure the diameter of the ball and computes $\frac{4}{3}\pi r^3$. The physicist fills a beaker with water up to its spot, drops the ball in, and then measures the volume of the water that's displaced.
The engineer looks up the serial number in his red rubber ball table.
At this point, I feel like I have a relatively good model for how my partner will react to being hungry. To execute this model, I try to imagine being in the same situation, consider all the context I'm aware of - and then look up "hungry" in my strange behaviours table and apply the override above. I am the engineer with the red rubber ball.
The glibness of the above formulation I feel rather undersells the extent to which this has been an extremely useful understanding. But it is hardly comprehensive. I feel like I now have, in my head, a very useful table with entries such as "hungry", "anxiety", "flowers". But faced with an entry not present in the table, I feel as helpless as the engineer encountering a blue tennis ball, or a red iron cube.
I recall some years ago watching 'The Good Doctor'. The main character - Dr Shaun Murphy - is an autistic savant - a brilliant surgeon, but incapable of relating to people and easily disturbed by unfamiliar situations. A frequent plot point would be his continued surprise at people's irrational responses to situations. This always seemed a little unrealistic to me. Dr Murphy was repeatedly shown as a genius, capable of modeling very complex biological and physical systems in his head. Given his repeated observations of characters behaving in "odd" ways, how was his model of them not updated to reflect this - how could this behaviour remain a surprise? Surely, a savant like Dr Murphy must be a committed Bayesian and updating his priors based on observation!
Reflecting on my own experience, I feel like I now have a couple of explanations for this:
This feels very much like a problem of a trapped prior. Somehow, I have in my head such a strong prior on "how people behave" that evidence of different behaviour is either dismissed as observational error or attributed to deliberate perversity. That is, while my conscious mind can - post facto - make the conclusion that the obvious explanation for behaviour A is a corresponding real experience of A, my subconscious concludes that the most logical explanation for A is "deliberate faking of A in order to elicit some reaction in me". Each time the situation occurs, my perception of it turns it into an experience of "somebody exhibiting perverse behaviour to wind me up", thus strengthening the posterior believe that this is a thing that otherwise reasonable people whom I have a lot of love for do on a semi-regular basis.
What's more, this further feels like a reference class problem. That is, even when I can deliberately act to escape the trapped prior, my reference class is too small to allow the appropriate update to take place. Rather, we end up with a hodge-podge composite model where specific exceptions get coded for specific inputs. Stuck in the wrong reference class - roughly, a reflection of my own behaviour - I lack the appropriate updates to be able to model others.
So what then is empathy? To me, it feels like a necessary (but not sufficient - some measure of will to empathise must surely also be present) condition is to have a sufficiently general model of human behaviour to be able to easily specialise to a given person given relatively few observations.
What's less clear to me is how necessarily to make use of the above reflections. Many of the suggested activites used to develop empathy - identifying similarities, reading fiction, exposing yourself to different perspectives, for example - feel like they fail to address the trapped prior issue. Perspectives outside the reference class feel like they generate - at best - an entry in the strange behaviours table.
Scott Alexander in the above post suggests that psychedelics might also be useful in adjusting trapped priors, though I'm not sure how much weight to put on one such opinion. Likewise, some meditation techniques seem to potentially be useful. For me, I hope, the act of identifying this bevahiour might itself help, moving the issue from "strange behaviour in other people" to "strange (but potentially understandable) issue in my head".