The Social Infrastructure We're Optimizing Away

More than just good looks.

A few months ago Provo had a fight about how wide streets should be. The city council considered a proposal that would increase the minimum allowable street width from 24 to 30 feet, at the fire marshal’s request, primarily in the interest of giving emergency vehicles easier access. Technically, the council acknowledged, the existing 24-foot minimum could be sufficient, so long as people followed the laws that restricted parking to one side of the road. But enforcement had been spotty, and who could argue against more drivable streets?

In truth, the case for wider streets was easy to make. Fire trucks need turning radius. Snowplows need clearance. Parking has to go somewhere, and so does traffic. Many of the claims came with quantitative data.

The case for narrower streets was much harder, and though many people from historic neighborhoods came to try, the argument never quite got a foothold. They (okay, we) reached for the functional arguments that were available: narrow streets slow cars, cost less to build and maintain, and leave more room for housing. Then we reached for the aesthetic ones: narrow streets are charming, cozy, in keeping with the older neighborhoods. All of these were true, but none of them were the point.

What nobody, myself included, quite managed to say was that a street’s width decides what kind of street it is. It decides whether crossing to a neighbor’s porch is a thing you do without thinking or a thing you look both ways for. It decides whether kids treat the pavement as a play area or as a boundary. It decides whether the people on opposite sides of it live on the same block or on two blocks that face each other. That was what was at stake, and it never quite made it onto the record, because the record has no column for it.

We’re fluent in two of the three things that make a place worth living in, and the third one, the one people actually mean when they say they love their neighborhood, gets its lunch money taken every time.

I’ve been thinking about that fight ever since, because it’s the clearest local example I’ve seen of a problem I keep running into: we’re fluent in two of the three things that make a place worth living in, and the third one, the one people actually mean when they say they love their neighborhood, gets its lunch money taken every time.

The two languages.

When we evaluate the built environment, we speak two languages. There’s the aesthetic: style, materials, form, the things renderings are made of. And there’s the functional: traffic counts, parking ratios, utility capacity, the things staff reports are made of. Nearly all professional work on cities lives in one of these two, and the best of it bridges them. A beautiful street that moves people well is a real achievement.

Zoning codes are written in these two languages, as are street standards. That’s an understandable constraint, but it means that whatever a place has beyond function and form has to be smuggled in as one or the other, or left outside.

The third dimension.

Every place has a relational dimension: the layer where it connects you to the people in it. It’s the porch that’s close enough to the sidewalk for conversation. It’s the bench where the same three men have held court for twenty years, or the corner store where the owner knows which kids are whose. In Provo, it’s places like the Shakespeare Streets—a beloved little neighborhood near downtown with small houses and streets even narrower than the 24-foot minimum, so human-scaled that I know people who drive there just to park on the edge and walk around.

It’s the layer you’re describing when you say you love your neighborhood, and the layer you’re grieving when you say it doesn’t feel like home anymore. It’s what residents mean when they reach for words like “character” and “charm.” At the hearing, we all pointed to the Shakespeare Streets and were, perhaps understandably, told, “nice anecdote.”

Words like “charm” gesture at a hundred concrete observations without naming any of them in a way that registers as immediately important. They describe an impression, not a mechanism, and that turns out to matter enormously.

Why “cozy” has no heft.

Ordinary citizens carry causal models of traffic in their heads, and they can argue from them. We all have aesthetic preferences we’ll actually get heated about. But almost nobody carries a causal model of why one street feels like it belongs to the kids and another doesn’t. Most people haven’t read the great documentarians of the relational dimension, like Jane Jacobs, Christopher Alexander, or Jan Gehl, and there’s probably no reason they should have (though I wish they would). So the relational case arrives at the microphone as merely someone's feeling or impression. Narrow streets feel cozy. True, and virtually useless, because the next question is why should that matter, and there’s nothing to say.

On a stage built for legible metrics, a claim without a mechanism doesn’t register as a weak argument, it registers as a luxury. I’d love for your kids to feel safe playing in the street, but I have to deal with traffic patterns. And the person saying that isn’t wrong, exactly. They’re responding rationally to the arguments as presented. That’s the whole problem. Every individual decision to defer the cozy thing in favor of the countable thing is defensible, and the sum of those decisions is a place nobody wants to be.

Every individual decision to defer the cozy thing in favor of the countable thing is defensible. The sum of those decisions is a place nobody wants to be.

I wrote a few months ago about the 4th of July parade, and about how the things that bind a community tend to look like waste until you’ve lost them. This is the same math, applied to the built environment instead of rituals. We are very good at optimizing. We are not very good at noticing what we optimized away.

Where it comes from, and why we can’t see it.

The relational dimension didn’t used to need defending, because it didn’t need producing. For most of history the people who lived in a place were the ones who shaped it, and relational quality was the byproduct of that process. A shopkeeper widened her stoop because customers lingered there. A block negotiated its own fences and shortcuts. The built environment absorbed the relationships of the people making it. Nobody designed the result, and the result was designed by everybody. Provo’s older neighborhoods were built at the tail end of that era, which is a large part of why they have what newer districts struggle to acquire.

Jane Jacobs caught this layer in the act. Her “sidewalk ballet” on Hudson Street wasn’t a description of architecture or traffic. It was choreography: the locksmith watching the kids, the sequence of regulars, an intricate order produced by people who knew each other in a place built at the scale of their knowing. What made her radical wasn’t her policy program. It was that she looked at the relational dimension directly and reported it as evidence that something important was missing from the top-down calculations of powerful institutions.

The political scientist James C. Scott, who spent a career studying why confident plans to improve places so often ruin them, drew a line between two kinds of knowing. There’s techne: knowledge that can be stated as rules and applied anywhere. Thirty feet of pavement. Two stalls per unit. It travels because the physics travels. And there’s what the Greeks called metis: practical, local knowledge that exists only in particulars—the fisherman’s feel for one stretch of coast, the neighbor’s sense of which block will absorb a change and which will break. You can learn it, teach it, even write about it. What you can’t do is apply it without judgment on the ground.

TECHNE / METIS — James C. Scott’s two kinds of knowing. Techne is knowledge that can be stated as a rule and applied anywhere. Metis is practical, local knowledge that exists only in particulars and can’t be applied without judgment on the ground.

Institutions, Scott argued, can only manage what they can render legible, what fits on a map, in a code, in a count. So they remake the world in the image of their instruments, and whatever the instruments can’t capture gets treated as if it doesn’t exist. His grim addendum: illegible things don’t merely get overlooked. They get destroyed, usually by well-intentioned schemes to improve the very places that depend on them. The relational dimension is the least legible thing about a place, and the most load-bearing.

LEGIBILITY — Scott’s term for what an institution can see: whatever fits on a map, in a code, or in a count. What can’t be made legible isn’t merely overlooked. It tends to be destroyed by well-meaning plans to improve the places that depend on it.

None of this means the relational dimension can’t be understood. Christopher Alexander spent a career doing exactly that: a pattern, in his sense, is a recurring relationship between a configuration and a human outcome, stated so you can use it a million times without ever doing it the same way twice. Narrow streets slow cars enough that crossing to a neighbor becomes casual. Porches within conversational distance of the sidewalk produce encounter. Those are causal claims, and they’re true nearly everywhere.

PATTERN — In Christopher Alexander’s sense, a recurring relationship between a physical configuration and a human outcome, stated so it can be used endlessly without ever being built the same way twice.

What can’t be done is turning the pattern into a spec. The functional effect of a narrow street can be abstracted away; the relational effect depends on the relationships that are already present.

That’s why what happened at the street hearing happens at nearly every major hearing. Residents perceive a relational threat and, finding no admissible vocabulary for it, translate: into parking, into process complaints, into architectural objections. The other side hears aesthetic and functional concerns and offers aesthetic and functional concessions, and nothing much improves, because the issue was never the thing we were all talking about. One side concludes the stated concerns were pretextual, which in a sense they were. The other concludes nobody listened, which in a sense they didn’t. Nobody is acting in bad faith. The vocabulary is failing everyone at once.

What to do if you care about something illegible.

Convincing institutions that their ROI calculations are not just missing something but destroying it is about as upstream as a battle gets, and it’s a large part of why I started Peregrine. But you don’t need a research firm to do something about it in your own pocket of the world.

Participate locally. Efficiency comes from scale, and scale means detaching decisions from context. The reverse is also true: the smaller the scale, the more room there is for knowledge that makes sense intuitively rather than objectively. Act locally not because it’s the best place to influence the relational dimension, but because it’s the only place.

Get organized. The best way to change local policy is often to change local norms. Rather than trying to convince the council yourself, gather likeminded neighbors and help people notice what has mattered to them only unconsciously. A claim without a number is a luxury. But that same claim made by several dozen residents is a constituency.

Sharpen the lens. Articulating the “squishy” things starts with seeing them clearly. Learn why a porch produces encounter and why a narrow street belongs to the kids—not technically, but relationally. Yes, I’m basically asking you to study anthropology, and I won’t apologize for it. “Charm” has no heft, but “Community attachment and a strong sense of belonging” often does. “Making downtown more walkable” feels like a luxury, but “encouraging people to shop casually at local businesses by having them out of their cars and in the city” feels like an economic development plan.

Local context comes from local voices. There is no column for the things that matter most, but there are many ways to make them legible anyway. ◆