How Buildings Learn: What Happens After They're Built

by Stewart Brand

SATISFICING doesn’t try to solve problems. It reduces them just enough.

Perfection is frustratingly temporary always,

Style cramps life, and life erodes style. All too soon the unified look is polluted by use, and it’s time to hire someone to supply another alien unity.

Paradoxically, habit is both the product of learning and the escape from learning. We learn in order not to learn. Habit is efficient; learning is messy and wasteful. Learning that doesn’t produce habit is a waste of time. Habit that does not resist learning is failing in its function of continuity and efficiency. Buildings keep being changed until they get to a point where they don’t have to be changed so much.

To change is to lose identity, yet to change is to be alive. Buildings partially resolve the paradox by offering the hierarchy of pace—you can fiddle with the Stuff and Space plan all you want while the Structure and Site remain solid and reliable.

What do the linked debacles of deep office buildings, sealed office buildings, and “smart” office buildings have in common? Each was a clever and comprehensive design solution, but each tried to solve just one primary problem and acted as if the problem would hold still over time. These were classic cases of overspecificity, overcentralized control, and “tight fit.” Each took a conspicuous trend of the moment—open offices in the 1960s, energy efficiency in the 1970s, information technology in the 1980s—and, at astronomical cost, shaped whole buildings tightly around it. When the trends moved on, the buildings were left standing, good at something that no one wanted any more. Their failure is the failure of optimization as a design strategy.

Several dynamics seem to be at work. Change is often followed by reversal of the change, because the prior pattern lingers as the most conspicuous alternative, because people are understandably conservative about their physical space, and because most change is really undertaken as a trial, no matter what people say at the time. And most trials are errors.

The severely ecological architect William McDonough imitates them with his insistence that any new office building he designs be potentially convertible into housing, since he regards that as the most fundamental use of buildings, for which there will always be a need and which always guides you toward humane design.29 His ingredients? Modest depth, high ceilings, operable windows, massive construction, raised floor rather than dropped ceiling for services, and individually controllable amenities such as window awnings. “It won’t do, any more, to think in terms of cradle-to-grave,” he insists. “From now on we should think of our buildings in terms of cradle-to-reincarnation.”

Perfection of detail is the enemy of change.

The common attribute of vernacular remodeling (and construction) is that it is done without plans. You proceed by improving on what already exists, following wherever usage demands. “Wanderer,” wrote a Spanish poet, “there is no path. You lay down a path in walking.”

SCENARIO PLANNING leads to a more versatile building. It takes advantage of the information developed by programming (detailed querying of building users) and offsets the major limitation of programming (overspecificity to immediate desires). The building is treated as a strategy rather than just a plan.

Architect William Rawn has found a way to get around clients who become entranced with defining every detail of the building. He gets the client to discuss the vision of the building at the very beginning in some depth, and then he can harken back to that vision when obsessive details threaten to overwhelm the building with fussiness. Vision is generic, and generic is adaptive.

The iron rule of planning is: whatever a client or an architect says will happen with a building, won’t. Architects always want to control the future. So do clients. A big, physical building seems a perfect way to bind the course of future events. (“Once we move the company into the new building, then we can use it to limit our growth.”) It never works. The future is no more controllable than it is predictable. The only reliable attitude to take toward the future is that it is profoundly, structurally, unavoidably perverse. The rest of the iron rule is: whatever you are ready for, doesn’t happen; whatever you are unready for, does.

SCENARIO PLANNING reduces the likelihood of being pushed around by a building obdurately clinging to a future that never happened. It reduces surprise in a good way. When something untoward happens, the building is ready for it.

The thing to avoid is a “bet-the-company” strategy that wins in only one scenario and loses in all the others. One approach is to develop a “robust” strategy that is viable in the variety of futures. Sometimes it is discovered by “regret analysis.” You ask, “What if we get it wrong? What would we regret not having done? What would we regret locking in?” You’re seeking to balance your risks intelligently. Another approach is devising an “adaptive” strategy that is exceptionally alert to changing events and can adjust quickly.

When and how would scenario planning fit into a standard building project? Two points in the usual design sequence come to mind. Sometimes both might be used, since scenarios frequently are developed by stages, with research in between to pursue the new questions that scenarios inevitably raise. The first opportunity would be the initial forming of the vision for the building. The second would be after there has been some preliminary programming, which might be used as the “official future” of the building, and which a fresh look at scenarios should challenge.

Scenarios can provide a way around some of the major design errors that commonly occur with buildings. The worst mistakes come not from wrong decisions but from not doing the right thing that never occurred to anybody to do. Scenarios attack a design from so many directions that gaps and oversights are likely to show up. Also, scenarios severely test fond notions that might otherwise get by unchallenged. (“We will have the stairs run clear around the periphery of the atrium like a magnificent spiral staircase.”) Grandiose ideas that don’t work are more hated by building occupants than almost anything else. Scenarios are conservative—they keep you from betting the whole building on some potential folly—but they are also innovative. Ingenious design ideas can emerge from the creative process of scenario-crafting and strategizing. (“Both grandkids said they love playing in low-ceilinged rooms, so we’ll save money and bulk with an old-fashioned half-height cottage attic, and they can have a kid heaven up there.”)

“Favor moves that increase options; shy from moves that end well but require cutting off choices; work from strong positions that have many adjoining strong positions.”9 More specific to buildings: overbuild Structure so that heavier floor loads or extra stories can be handled later; provide excess Services capacity; go for oversize (“loose fit”) rather than undersize. Separate high- and low-volatility areas and design them differently. Work with shapes and materials that can grow easily, both interior and exterior. “Use materials from near at hand,” advises Massachusetts builder John Abrams. “They’ll be easier to match or replace.”

“We overestimate technology in the short run and underestimate it in the long run.” So design loose and generic around high tech. You will be wrong about what is coming, and whatever does come will soon change anyway.

Postponing some of the design yields more building for less money, since less is spent on detail design and detail finish. And adaptivity is built in.

The Darwinian mechanism of vary-and-select, vary-and-select has one enormous difference from the process of design. It operates by hindsight rather than foresight. Evolution is always away from known problems rather than toward imagined goals. It doesn’t seek to maximize theoretical fitness; it minimizes experienced unfitness. Hindsight is better than foresight. That’s why evolutionary forms such as vernacular building types always work better than visionary designs such as geodesic domes. They grow from experience rather than from somebody’s forehead.

The cosmetic practice of hiding wires and pipes in walls makes maintenance and improvement a major hassle. The conservative tactic—at higher initial cost—of installing over-capacity electrical feeders and breakers, oversize chases, and an apparent excess of outlets is nearly always rewarded. The general rule is: oversize your components. Another adaptive trick—both for incremental and radical change—is to keep wiring accessible in what’s called wire mold on the walls or in cable troughs hanging from ceilings.

PHOTOGRAPHING WALLS BEFORE THEY CLOSE has become a standard practice for design/builder John Abrams, because it makes later adjustment of the building so much easier. The photos reveal exactly where the Services go and what are the hidden Structural elements.

Abrams recommends doing the same thing for earlier stages, photographing the foundations and their drainage before fill-in, the septic system before fill-in, and all trenches. Everything that is buried will be dug up one day. It’s better to leave a treasure map.

LIVE/BUILD is the intimate, traditional way. Dwelling at the construction site can save money, can add to site-sensitivity and use-sensitivity in the design, and can either spread the construction costs and hassle out over time or goad the dwellers to hurry up and get the damned thing finished.

“Finishing is never finished,” but at some point you have to just stop, let the builders go away, and start living in the place. Inhabitation is a complex, thrilling, prolonged event. The way to treat a new building is as if you might move away tomorrow and as if you expect to be there the rest of your life. Inhabit early, build forever. Part of relishing a place is tinkering with it.

People are happiest also in buildings where change occurs at every scale from weeks to centuries. Such buildings are fractal in time.

Besides serving its immediate purpose, Alexander suggests, each project should serve a larger goal of “healing the whole.”13 And it should prepare the way for a larger and more significant whole.

The point is to make adjustments to a building in a way that is always future-responsible—open to the emerging whole, hastening a richly mature intricacy. The process embraces error; it is eager to find things that don’t work and to try things that might not work. By failing small, early, and often, it can succeed long and large. And it turns occupants into active learners and shapers rather than passive victims.

What makes a building learn is its physical connection to the people within.