Stories and Theories

As readers of this blog have surely noticed, I prefer to discuss theories rather than tell stories. Systems-thinking comes naturally to me; narratives, not as much.

I'm ambivalent about this. (Ambi-valent — having both positive and negative feelings.) On the one hand, systems-thinking is incredibly powerful, and I'm glad I have a facility with it. On the other hand, stories are also hugely important — different from theories but similarly powerful — and my deficiency with stories is a real weakness.

I'd like to understand stories better, and I know of only two ways to improve at something: theory and practice. We'll get to practice in a minute, but first I'm going to do theory — theory about stories. And also (bear with me here) theory about theories.

Ultimately, what I care about is making sense of the world, and both stories and theories are instruments to that end. But they are different instruments, different lenses, each with their own strengths, weaknesses, and idiosyncrasies.

Let's start with a brief discussion of what stories and theories have in common.

(N.B.: Philosophers have been discussing these ideas for centuries — from Plato to Dreyfus and everyone in between — and with far more depth and precision than I care to attempt. My goal here is just to set the stage for our two main players.)

Both stories and theories are ways of representing reality (re-present — to show again, in a different way). With theories this is fairly straightforward. But stories too — even the wildest fantasies — are representations of reality. Whether a story aims to educate or (merely) to entertain, it must ring true on some level or we wouldn't recognize it. Even the most genre-bending experimental fiction will have notions of time, objects, and events, and characters with human-like agency.

Both stories and theories ultimately derive from, and connect back to, direct first-hand experience. Humans are embodied creatures, situated in the world, and our m.o. is participation. Participation gives us our first and most fundamental understanding of reality, an understanding that is primary and visceral, detailed and highly contextual, intimate and deep. It's the rich raw material out of which we construct our stories and our theories. [1]

Now on with the show.

Stories

The question I want to ask about stories is: what makes them uniquely powerful tools for understanding the world?

First, stories are engaging, even captivating. A great story holds its audience in rapt attention, helping them attend to the details and absorb as much information as possible.

Second (and more important), stories are great at inciting empathy. Empathy — putting yourself in someone else's shoes in order to feel what they feel — is a quintessentially human skill. It liberates us from our first-person prisons by showing us the world through others' eyes. And stories give us the opportunity to empathize with many different characters. Through stories, we get to see the world from every angle.

But reality is also highly engaging and offers plenty of chances to empathize with our fellow humans. Why turn to stories when we can just watch the world unfold?

One reason is scope. Stories give us access to a wider range of the experience than we could ever hope to reach by ourselves, directly. Stories illuminate every facet of the human condition, from the highest highs to the lowest lows, and every crevasse and corner in between. Stories show us what it's like to be male and what it's like to be female — what it's like to be rich and what it's like to be poor — what it's like to be black and white, yellow and red, and all the shades of brown — on the side of the good guys and the bad guys — to be young and to be old — to live in Beirut and Beijing, Hong Kong and Honolulu — in the distant past and the unreachable future — to live through wars and famine, the roaring 20s and the Great Depression — births, deaths, dreams, hallucinations, and a hundred thousand other things.

Most stories focus on what's exciting and unusual — a bias we can wholeheartedly forgive. But on rare ambitious occasions, a story will even cover what it's like to be bored.

Stories give us access to hypotheticals and counterfactuals. What if someone had a ring that made him invisible? What if Charles Lindberg had been elected president in 1940? What if a prince became a pauper? What if an asteroid was headed directly for Earth?

Stories let us inspect reality at different scales and speeds. The storyteller's lens (whether literal or metaphorical) can record high-frequency/narrow-focus/slow-motion video or low-frequency/wide-angle/time-lapse video. A story can zoom in on the smallest decisions, the most minute facial expressions and body language, pausing to tell the entire backstory to a single thought. Or it can zoom out and trace the rise and fall of a civilization, spanning centuries and touching millions of lives.

And the scope of stories is not only broad, but deep. It's uncanny how small, simple phrases can channel so much intimacy. A great story will take you inside someone else's head in a way no other experience can, inciting feelings you didn't realize other people felt (maybe not even yourself).

Another thing stories provide (that we can't get from the real world) is enhanced legibility. The world is messy and hard to decipher. Stories simplify reality, presenting it in a way that's easier to understand. Personalities are more pronounced, action more decisive, cause and effect more clear-cut. These things make stories easier to reason about than direct experience — and easier to share and discuss with others.

Legibility (by itself) is good, but it's in tension with something equally important: realism. Legibility and realism aren't diametrically opposed, but they aren't complementary either. The more a story tries to paint in terms of black and white (good and evil), the less realistic it will be. But the more it sticks to an aesthetic of realism, the messier it gets.

All stories omit details, but the devil lies in which details are omitted. Done well, legibility casts the world in high relief, enhancing the important details and omitting those that would distract. But it's tempting to oversimplify, to leave out something essential. Legibility is a delicate art.

Theories

If stories are one way of approaching the world (for the purpose of making sense of it), then theories are another. Theories look at systems as a whole, rather than (or in addition to) tracing the arc of individuals or components within a system. Stories can and do inform theories — and vice versa — but they are fundamentally different ways of trying to understand the world.

We've discussed what makes stories so powerful. Now I'd like to ask, why are theories so powerful?

I think there are two main reasons. The first, once again, is scope. The scope of stories is vastly greater than that of direct experience, but the scope of theories is vastly greater still.

Theories provide truly mind-altering shifts in perspective. They allow us to think about smaller and larger objects, at smaller and larger time scales, than anything we can begin to fathom through direct experience. We can theorize about the tiniest particles in the universe, billions of times smaller than anything visible to the naked eye. We use theories to reason about the evolution of galaxies, the earliest microseconds of the Big Bang, and the eventual heat death of the universe. Theories help us understand how the earth formed — where babies come from — what goes on inside our cells — how we're related to every plant and animal, every fungus and bacterium — how our social systems function (or fail) at civilizational scales — how our brains work — how to build globe-spanning computer networks — how to deliver earth-born machines to other planets — ... and a hundred thousand other things.

Stories let us see the world through many different eyes, but they are all human eyes, human perspectives. Theories show us reality on its own terms:

The other reason theories are so powerful is that they make predictions.

Predictions are great because they tell us what's going to happen before it happens (pre-dict — to speak in advance). A story can also help us reason about the future, but only by means of a theory (implicit or explicit) that we've managed to extract from it. Stories, by themselves, are just case studies.

Predictions are also great because they help us go back and check whether a theory is true or not. This is the criterion of falsifiability — the linchpin of science — and its importance can't be overstated. All knowledge, all progress, depends on falsifiability. (OK, maybe that's an overstatement — but not by much.)

We spoke about the tension between legibility and realism in stories. Theories suffer no such tension — they are legible (for the most part) by construction [2]. Theories do admit to greater or lesser realism — they can be false (inaccurate representations of the world), or oversimplified (to the point of being irrelevant or even dangerous). But weak theories are exposed, sooner or later, when they make bad predictions.

An anecdote(!)

A few years ago I ran across a fun little book called Year Million, a great case study in stories vs. theories.

The book was a collection of 15 essays commissioned from 15 different futurists, each to answer the question, "What will our world look like in the year 1,000,000 C.E.?" Not ten years from now, not a thousand — a million.

A million years is an unimaginable stretch of time, but the authors chose to try their imaginations anyway. They all chose to tell stories.

All of them, that is, except Robin Hanson (our fearless protagonist!). He decided to take the question seriously — seriously enough to develop a theory and extract predictions from it.

Robin began his essay with a word of caution about using stories to peer into the future:

The future is not the realization of our hopes and dreams, a warning to mend our ways, an adventure to inspire us, nor a romance to touch our hearts. The future is just another place in spacetime.

He then went on to develop a theory powerful enough to wrangle a million years' worth of sprawling possibility (an un-imaginable amount) into something approaching probability:

In this chapter we will use evolutionary game theory to outline the cycle of life of our descendants in one million years. What makes such hubristic conjecture viable is that we will (1) make some strong assumptions, (2) describe only a certain subset of our descendants, and (3) describe only certain physical aspects of their lives. I estimate at least a 5 percent chance that this package of assumptions will apply well to at least 5 percent of our descendants.

The rest of Robin's essay was dry and technical, devoid of moral judgments — an ascetic celebration of theory. But he gave it a colorful title, "The Rapacious Hardscrapple Frontier" — a wink, perhaps, to literary sensibility.

Of the authors who chose to tell stories, Rudy Rucker's was the worst. The weakness of his method bled into nearly every paragraph. For instance:

On a practical level, once we have telepathy, what do we do about the sleazeball spammers who'll try to flood our minds with ads, scams, and political propaganda? We'll use adaptive, evolving filters.... Another issue with telepathy has to do, once again, with privacy. Here's an analogue: a blogger today is a bit like someone who's broadcasting telepathically...

I'm sorry — I can't go on.

Those concerns are, in the words of Feynman, "too local, too provincial" — out of proportion with the vastness of a million years. It's distasteful, almost irresponsible, to use a story about Year Million to moralize at those of us back in Year Two Thousand.

As Robin said, the future is not a warning to mend our ways. It's just another place in spacetime.

Oops

I didn't mean to end on a note so flattering to theory and so disparaging of stories. I'll follow up soon with some thoughts on why (and more importantly, where) stories are powerful. Stay tuned.

___

Endnotes

[1] experiences are raw material for theories. Actually this is a simplification. Kant argues (persuasively I think) that we actually can't have any pre-theoretical experience of the world — that even our most basic perceptions are shaped by innate 'theories' about how to turn sensory information into a model of the world. For example, we naturally/instinctively chunk our sensory input into 'objects'. If our brains didn't come pre-wired (out of the womb) to learn how to perceive objects, we'd never get past William James' stage of "blooming, buzzing confusion" — our so-called 'direct' experience would never get off the ground.

[2] theories are legible by construction. Again a simplification. In this post, I'm mostly referring to explicit theories — the kind we can discuss and consciously reason about. But as Dreyfus (and many others) have noted, theories need not be explicit to be powerful and useful. We have many implicit/illegible theories about how the world works that we carry around individually, in our heads, or collectively, in our culture. But here we're discussing the kinds of theories that are comparable to stories — i.e., the explicit ones.

Special thanks to:

• Tristan Brand for prompting me to think about stories.
• Robin Newton for reintroducing me to Dreyfus's critique of AI.