The Problem with Perspective

If a lion could talk, we could not understand him.

— Ludwig Wittgenstein, Philosophical Investigations

This past summer I was interning in Austin, sharing a rented house in East Riverside with roommates my age. We got along most of the time. One evening the dishwasher had been closed since the night before. I didn’t even open it to check. I said I thought it had been run the day before, since I remembered asking someone to. One of my roommates wasn’t convinced. Nobody actually knew, and he was frustrated, and at some point I told him, “it’s not that deep.”

It made him even more agitated, which I shoud’ve expected. I knew it would piss him off, and that’s partly why I said it. Why’d I do it? I think I wanted to downplay the whole thing, to turn it into a joke so it would stop from becoming an argument. But joking about the situation meant joking about how he felt. And I never asked him why it was such a big deal to him, because why would I? I made a claim that night about something I couldn’t begin to understand: how much it mattered to him.

To decide how much something matters to someone, you’d have to know what it’s like for them. How would I know that? How could anyone possibly know that?

I fear that we may never know how much our experiences differ, even between two (seemingly) similar people.

How do we know if a set of experiences between two people is identical, or even remotely similar? Is the colour red that I see the same red you see? Picture a piece of chocolate. I can do my best to describe what it tastes like: rich and sweet. You might taste both too, but at different levels. What if it’s more bitter than sweet for you, and that’s why you don’t enjoy chocolate?

And I never said which chocolate. You might’ve pictured dark, or white, which has barely any bitterness at all, and a word as simple as “chocolate” has some ambiguity in what it describes even before taste comes into it. Whichever one you imagined, though, is your bias at play, the one that defines chocolate for you.

The physicalist answer is that we could analyze the taste receptor cells in your taste buds, record the neurotransmitters they release, and see which neurons fire in your brain as a result. Does that guarantee you and I experience chocolate the same way? The uncomfortable answer is no. We don’t truly know.

And honestly, this bothers me because I’m a physicalist. If I believe everything can be reduced to its fundamental particles, then in principle your experience should be observable. I’d be able to map every neuron, every signal, and I’d have it. But even if we could, how would I know the map is the feeling? I could know exactly which neurons fire when you taste chocolate and still not know what it’s like for you. There’s still so much we don’t understand about the brain, about consciousness, about what it even means to experience something, and I’m not sure physicalism has an answer for that yet.

Wittgenstein’s “Beetle-in-a-box”#

Wittgenstein has a thought experiment for this. Imagine a group of people who each have a small box, and inside each box is something everyone calls a “beetle.” There’s only one rule: each person can only look inside their own box. Everyone talks about what’s in their box using the word “beetle,” but it’s entirely possible that everyone has something completely different in there. It could be constantly changing, or the box could be entirely empty.

His point was really about how words for private feelings get their meaning, but it feels like we can generalize this to perspective. We all share the names of emotions like joy, fear and love, but what each of us actually feels when we use those words is a private object inside our own heads, and they’re surprisingly hard to measure and even harder to put into words.

I feel like the reason everyone’s box is different comes down to how we live. Every small thing that has happened to you—your arguments with parents, your first love, your greatest achievements to date, your lowest of lows—goes inside the box. It shapes not just your conscious mind but your subconscious too. And it compounds as we live. One experience changes how you see the next, which changes how you see the one after that.

It reminds me of training a neural network. A model learns from the data it’s shown, and it gets good at whatever that data covers. Picture two models trained on data that overlaps, like a Venn diagram. On inputs inside the overlap, they’ll mostly agree. Outside it, where one model has seen examples the other never has, they can give completely different answers.

Although it’s an oversimplification, that’s roughly how I think we understand each other. When someone tells me how they feel, I run it through my model, the one trained on my life, and I get back a guess at what they’re feeling. Where our lives overlap, the approximation is pretty good. If someone’s family member dies, ask a hundred people and nearly all of them would strongly agree that it’s sad and the grief is justified.

But something strange also happens here. If someone’s dog dies, you’d probably still feel for them, even if you’ve never had a pet, even if you’ve never lost anyone close to you. So where does that empathy come from, if not from overlap? Maybe we’re generalizing. Somewhere along the way, just from living in this world, we learn that death is associated with sadness and grief, and we apply it to losses we’ve never had. Is that understanding? Or is it just a really good guess, the kind a model makes when it’s seen enough nearby examples? I don’t know. Maybe that’s all empathy ever is. But the further you are from a shared experience, the worse the approximation gets.

And it gets stranger: two networks trained on the exact same data, starting from different random initializations, end up with different weights, yet they can still give the same answers. That might explain how two people could live the same lives and one comes out pessimistic while the other comes out hopeful: they both have the same data, but have very different starting points. It also means that identical answers tell you nothing about what’s going on inside. We can both call chocolate “rich and sweet” and be describing different beetles.

Neural networks are also surprisingly bad at knowing when they’re out of their depth. Give one an input unlike anything it was trained on and it doesn’t say “I don’t know.” It gives a confident answer anyway.

I think we do the same. That night in Austin, I had no data on why a closed dishwasher bothered him so much.

I’ve been on the other side of it too, and it felt like a personal attack. Not because of the words, but because of who said them. I was emotionally vulnerable with someone I thought truly cared for me and would be there for me no matter what, and they decided it was best to stop talking about it, dismissing my feelings even though it mattered to me. It felt like someone who genuinely cared for me wasn’t interested in what I had to say.

But it’s a weird balance. Sometimes you genuinely need pushback. If I dropped my ice cream and it made me sad for the next few weeks, would it be best for my friend to try to understand me, or to tell me to move on? Probably the latter. But I think the difference is whether they tried first. Pushback without attempting to understand feels like dismissal. Pushback after someone has tried to understand feels like empathy.

The perspective I’m missing#

A while ago I was hiking Yellow Aster Butte, near Mount Baker, with two friends, and someone asked, “Would you rather have high IQ or high EQ?” My first thought, cliché as it is, was this shouldn’t be a binary. Then admitted I’d still lean IQ, which, in hindsight, was a very IQ answer. Logic just feels more powerful to me than emotion.

But the reason I think the way I do is a result of my own experiences and how I’ve perceived them. My trust in logic comes from my past experiences too: for a couple years, I weighed my food and tracked every single macro- and micronutrient that went into my body, because that seemed like the most logical thing to do, and the fastest path to achieving a healthy body. So at times, I do wonder if there’s a perspective I’m missing. I feel like spending a few years of mostly math has trained me to look for the right answer or at least “optimal” choices. But situations like these don’t reduce to a finite number of choices, let alone a binary.

Does it even matter?#

But does it even matter that we can’t open each other’s boxes? Do we actually need to know if two perspectives are identical?

As humans, we’re tribal. We want to relate to the people who get us, and we push away the ones who don’t, and maybe that’s why this gap feels like a problem. But imagine someone told you chocolate tastes sour to them, more like an orange than anything sweet. You wouldn’t feel alienated. You’d want to know more. It’s the same as looking at the clouds with someone under the same sky. They see shapes you’d probably never see.

If our boxes were all identical, nobody could show us anything we couldn’t find ourselves. The difference is what makes other people worth talking to.