Empathy Puzzles

How does art involve us? And when it does, is that good?


There’s something to be said for thinking of art as an engineering problem1. When we look at art as engineers — what can we say about how this works? — we avoid the lazy school of thinking that says, art is really just too personal, it’s all subjective, we can never really know anything about it. To that I say: really? Because some works seem to resonate with us, our friends, and thousands or millions of others, and some works don’t.

On the other hand, all engineering and no art is also a recipe for disaster. Some…interesting takes have been produced by pure engineers who have little appreciation for art themselves. They proceed to analyze “art” as this type of information that, for some mysterious reason, people insist on going around making and consuming. If you attempt to model art as a transfer of data, you’ll get the idea that you make art by (1) taking True and Beautiful Stuff, and (2) crunching it into some output format (perhaps serializing it into Protobufs, for you SWEs in the audience).

This metaphor is tempting, but it’s not a good model of what artists do, or even a good model of how art is received. As I wrote about last time, we can’t just beam stuff into each others’ brains; we have to work indirectly, choosing specific artistic techniques to evoke specific kinds of understanding in the reader.

In A Swim in a Pond in the Rain, noted engineer-artist George Saunders writes that “the fundamental unit of storytelling is a two-part move.” First, the writer sets something up that causes the reader to have certain expectations — by creating a mystery (“a man in a hooded cloak watched them from across the room”), by evoking genre tropes (“she was laser-focused on her high-powered career, and love was the last thing on her mind”), or whatever. Then, the writer makes use of those expectations in an interesting way — not by straightforwardly satisfying them, but also not by subverting them in an arbitrary or disconnected way.

I think this is a fine example of how engaging with art — in this case, reading fiction — is a process that is more than the sum of its parts. We don’t just absorb all the text and the process it after, like you’d do if you were writing a computer program to analyze texts. Think of a piece of fiction, instead, as a system with three parts: the words on the page, the things they literally refer to, and the effects that reading has in the reader2.


One of my blogging inspirations, Hotel Concierge, once wrote:

Young adult fiction can have complex characters, worldbuilding, and rules of magic/ethics/rationality as long as the complexity is spelled out for you. “Snape was mean to Harry…but [flashback] that’s because, deep down, he was still in love with Lily…” What such stories will never ask you to do is intuit that Snape had tribulations, they assume that you have no instinct for cognitive empathy […]

Adult fiction, or literary fiction, by contrast, does make such demands on us. Beyond transmitting information, literary fiction often provokes us to do certain kinds of work, to ask ourselves why the characters do what they do.

Faulkner’s short story “A Rose for Emily,” which you (like me) may have been assigned in high school or undergrad, is a very vivid example of this — if you’re not familiar, no worries, I’ll explain as we go3. It operates as a puzzle on a couple of levels. Most obviously, it’s told in five chronologically-disconnected segments, so it’s a basic logic puzzle: in what order do these five things happen? This particular puzzle is not hard, but it forces you to consider the text and create your own, separate mental model. The story doesn’t come with an accompanying timeline or study guide, at least not in my copy.

The story concerns Emily Grierson, the last member of a dying Southern aristocratic clan. She’s treated by the town as a relic of a bygone era, with a mixture of reverence, pity, and irritation (because, e.g., she doesn’t pay taxes). Throughout her youth, her father drives away any potential suitors, leaving her unmarried. After his death, she meets a Northern businessman, Homer Barron. He’s socially “beneath her,” but she’s too old to have many other options; they get engaged. Then he disappears.

At the end of the story, spoilers, it turns out she poisoned him, and has been sleeping in the same bed as his corpse for the past 20+ years.

This leaves us with a few questions. Like: why did she do that?

Without assuming anything about “authorial intent,” I would submit that there are more and less correct answers to this question. When I learned this story at age ~15, I said something like “Emily had a psychological need to control her life, and to put it into stasis.” This may or may not be The One True Answer, I’m not a Faulkner scholar, but I’m somehow completely confident that it’s better than “Emily really wanted to bang a dead guy.”

We intuitively sense that a good answer should be made up of the material in the story, and obey certain laws of form: like a well-crafted puzzle, the story asks us to accept certain premises, and draw them toward a conclusion. And certain ways of doing that “feel better” than others.

Still, stating this answer is a fundamentally different thing from reading the story, just as both are fundamentally different from my one-paragraph summary of the major plot points. Even if we assume that this answer is “correct,” it skips the process of you generating it. Or, perhaps, the true “answer” is so intricately indicated by each word of the story that nothing less than the full story actually expresses it. For a strictly economical form like the short story, that might actually be true.


I don’t necessarily mean that a story can be “solved,” or that being an empathy puzzle is the only function of a good story. Actually, I can easily think of great stories that aren’t very much like this at all. But I think this experience is something we encounter uniquely in narrative art, this experience that feels like “solving” for human behavior.

Remember the much-maligned Game of Thrones “subverting expectations,” above? It’s an interesting example, because the meme above unleashed “subverting expectations” as a critique into the popular consciousness. Like any popularly-understood critical framework, it gave people a powerful weapon. It gave people the option to say, “aha, I don’t like this thing because it’s objectively wrong.” That’s turned out to be a mixed blessing.

In this piece4, Sam Kriss writes about the second-to-last episode of GoT. After Daenerys wins the war, she refuses to accept surrender, and brings seemingly-unnecessary carnage and destruction to the people of King’s Landing. I remember that this choice in particular was reviled on release: “Daenerys wouldn’t do that!”

Look, the last couple seasons of GoT sucked, I don’t blame anyone for checking out. But more broadly, when this kind of thing happens in a work of fiction, there are two ways to react. You can say: this doesn’t make any sense, the producers/writers are “subverting expectations” again, they did it wrong, I reject this. Or, you can say: okay, this seems weird to me; can I guess why this character might act like this?

Again, I don’t feel strongly about this specific example. But if we get too used to reaching for the “subverting expectations” hammer, we lose the possibility of understanding anything new about human behavior.


I’m dancing around a pretty old justification for literature: building empathy as a noble end in itself. This is what you’d get from DFW’s “This is Water,” or what Saunders espouses in A Swim in a Pond in the Rain, or what your English teacher told you. But I’m not quite satisfied with this.

I have a friend who wants to — if I may editorialize, really should — break up with her partner. This guy sucks. She goes to couples’ therapy sessions with him. She complains that he always brings up some new bullshit reason to justify his bullshit complaints, always something that wasn’t mentioned in the initial phase of couples’ therapy, where they were supposed to put all the complaints on the table. I find all this very easy to imagine, because this guy sucks.

But it’s hard for her to up and leave, and despite knowing that this guy sucks, I get it. You want someone to say, okay, I’ve seen everything, and *referee’s whistle* YOU’RE RIGHT!

I am, like, 99.9% sure that she should break up with him. But her inquiry (“can this relationship work?”), like every earthly inquiry, is a process which begins at some point, proceeds under some conditions, and arrives at an end, also defined by some conditions. And the inquiry she’s subjected herself to, in couples’ therapy, will never arrive at the end that she needs to pull the trigger.

When we read a lot of literary fiction, we learn that it is noble to start at point A of empathy (“why did this person behave this way?”), and proceed toward point B of empathy (“ah, they had a hidden aspect that I didn’t understand”). Perhaps we arrive at point B, or perhaps we approach point B, or perhaps point B is unreachable but even so the effort is worthwhile. In some sense all of these teach the same lesson, because even if the exact destination is different, we learn that this approach is generally, directionally, Good. So if we can’t solve a problem in this framework, it feels Bad.

Unfortunately, in real life, people will fuck you over, and they often won’t have the decency to give you a satisfying reveal of their deeper motivations. They are the heroes of their own psychodramas, which, for all your efforts, may remain inscrutable to you. If you say, “well, they must have had a bad childhood” — sure, maybe they did. Now what are you going to do?

As Hotel Concierge wrote, provocatively, on another occasion:

When someone slaps your hypothetical girlfriend’s ass in the proverbial club, what does humanism say you should do? At least toxic masculinity has an answer.


Yeah, these days, I’m not so satisfied with the noble defense of empathy in itself. Let me suggest, instead, that empathy has a particular practical purpose.

The average medieval peasant encountered only 100 narratives in their entire life. If you’re active on social media — especially, God forbid, Twitter — you encounter at least 100 every day, and you definitely don’t have time to fact-check them all. If you’re not active on social media, well, all your friends are (and please let me know how you found this essay). So how do you sort through all this garbage, ideally without becoming an unwitting puppet of some deranged cause?

My hope for humanity is that we each have, deeply built-in, an astoundingly good narrative bullshit detector. It’s the sense that warns you you’re being played, when you’re told how outraged you need to be about person/group X. It’s the sense that warns you you’re being coddled, when you’re told what a perfect little person you are, especially if you behave just like so. It’s how you know when you’re being conned, and when someone is trying to get something from you.

So here’s my pitch: empathy puzzles hone that sense. You feel a twinge of aesthetic disgust when you’re confronted by a poorly-constructed puzzle, or an unconvincing answer. Doing empathy puzzles makes your bullshit detector louder, so when it warns you of something, you have a better chance to hear it.

Just as people do crossword puzzles to stave off dementia, doing empathy puzzles staves off Internet-induced schizophrenia. Empathy puzzles exercise a certain mental process. This process gives us the power to, first, accept people who behave differently than we expect, and then, to consider why they might act the way they do. Moreover, as we get an instinctive sense for the patterns traced by this process, we get a chance to think twice when we catch someone cutting corners.

Can empathy be bad in excess? Sure, like any paradigm, including the paradigm that everything is bad in excess. But it’s a useful tool to have.

I doubt I will ever be comfortable making sweeping statements about “what art does,” but for now, let’s say this is one of the things.


Some of my thinking here is inspired by @SuspendedReason’s collection “Art (That) Works,” which brings together a lot of thinking in this vein from the early 2010s. Luminaries of this movement include Suspended Reason himself, Hales a.k.a. The Sublemon, Gabe Duquette, and many others. Worth checking these folks out!


My resemblance to Peirce here, which I only noticed in editing, was totally unintentional. I’m not sure that what I’m saying lines up with his model, but I appreciate that he distinguishes between the object (the thing in the world that is being signified by the sign) and the interpretant (the thing that happens in the receiver when they receive the sign).


If you’re very curious, you can read it online here. Fair warning, it’s set in a Mississippi town in the late 19th / early 20th century, so the characters are big racist, and the N-word is used liberally.


h/t @sympathetic_opp