You are standing in a permit office, third floor, fluorescent hum overhead, watching a structural engineer circle a clause about anchor bolt spacing in red pen. The contractor shrugs. The architect checks the footnotes and finds nothing useful. The clause is followed, the bolts go in, and the moment passes without ceremony. Nobody in that room knows they are complying with the aftermath of a fire that killed dozens of people before any of them were born.
That is how building codes actually work. They are not logical systems derived from first principles. They are sedimentary records, layer pressed onto layer, each stratum deposited by a catastrophe that was, at the time, considered unthinkable.
The rule that outlived the memory of why it exists
Take the requirement for outward-swinging exit doors in places of public assembly. It appears in virtually every modern building code descended from American or British practice, stated flatly, without explanation, as though it were obvious. It is not obvious. Outward-swinging doors are harder to weatherproof, they interfere with pedestrian flow on busy streets, and they cost more to frame correctly. There are genuine engineering reasons to prefer inward-swinging doors in many contexts.
The rule exists because of the Iroquois Theatre fire in Chicago, in which more than six hundred people died, many of them crushed against doors that opened inward and jammed under the pressure of a panicking crowd. The survivors who testified described people stacked four and five deep against doors that would not yield. Legislators and fire marshals rewrote the rules within months.
The doors in your local cinema swing outward because of what happened in that theatre.
Ask a junior fire-safety consultant why exit doors must swing in the direction of egress travel. If they are honest, they will say: because the code says so. The mechanism is remembered only by historians. I find that genuinely unsettling, not as an abstract observation about institutional memory, but as a practical fact about the people making decisions in buildings right now.
Concrete, rebar, and the ghost of San Francisco
Seismic provisions offer a sharper illustration, because earthquakes tend to produce building failures that are specific and photographable. After the 1906 San Francisco earthquake, engineers observed that unreinforced masonry buildings collapsed while wood-frame structures nearby survived. The lesson was absorbed locally, gradually, and imperfectly. Reinforced concrete became standard for larger structures, but the detailing of that reinforcement varied enormously.
Then a major California earthquake collapsed freeway overpasses and hospital wings built to codes considered modern at the time. Engineers examining the rubble found that columns had shattered because the spiral ties holding the longitudinal rebar together were spaced too far apart. The concrete core had no confinement. Under lateral load it simply exploded outward, the way a stack of coins explodes when you strike it from the side, fast and total and irreversible. Within a few years, California codes mandated tighter hoop spacing in column plastic-hinge zones, a requirement now known as confinement reinforcement.
The specific number in the code is not derived from a tidy formula. It is derived from photographs of columns that failed and columns that did not, from engineers measuring the difference and writing it down. A structural engineer specifying column ties at a spacing of four inches or the diameter of the longitudinal bar, whichever is smaller, is following a rule whose precise origin is a collapsed hospital wing and the people who walked through it afterward.
That is still the optimistic case, because at least someone in a seismically active region probably knows the general outline of the story. Most code provisions are not that well remembered.
What people get wrong about why codes are so complicated
The common assumption is that building codes are complicated because bureaucracies love complexity, or because lawyers have made everything defensive, or because the construction lobby has inserted protections for its preferred materials. All of those things are true to some extent. They miss the dominant mechanism.
Codes are complicated because disasters are varied. A provision about draft stops in concealed roof spaces exists because fires have spread invisibly through void spaces and emerged simultaneously at multiple points, baffling firefighters. A rule about tempered glass in shower enclosures exists because ordinary glass, when it fails, produces long shards at heights and angles that cause serious lacerations. A requirement for a secondary drain on a flat roof exists because primary drains have blocked, ponding water has exceeded the design load, and the roof has collapsed onto occupied space below.
Each of those sentences is a compressed incident report.
Consider two engineers, call them Marcus and Diane, who trained at the same university and graduated the same year. Marcus spent his career in the midwest, primarily on low-rise commercial work. Diane worked for a firm in Japan for a decade before returning. Ask them both about base isolation systems and their answers will diverge sharply, not because of intelligence, but because Diane has absorbed, through practice and proximity, the institutional memory of a seismic culture. Marcus follows the same seismic provisions Diane does when a project requires them. He follows them as rules. She follows them as conclusions.
That gap matters when an unusual situation arises that the code does not directly address.
The slow erosion of why
Codes are revised on cycles, typically every three to six years for the major model codes. With each revision, language gets cleaned up, cross-references get renumbered, and the explanatory commentary that once accompanied a provision gets trimmed for concision. The provision survives. The rationale shrinks. After two or three revision cycles, the rationale may exist only in the original committee minutes, technically public and practically inaccessible.
This is not unique to building codes. Food safety regulations carry the memory of contamination outbreaks. Aviation checklists encode the memory of crashes. The difference is that pilots are explicitly trained to know which checklist item came from which accident. That institutional practice is not standard in construction education, and the profession is worse for its absence.
The result is a field that is, in a precise sense, protected by knowledge it does not know it has. The bolts go in at the specified spacing. The door swings the right way. The column ties are close enough together that the concrete core will hold when the ground moves.
So here is the question worth sitting with: if the code is functioning as a kind of collective memory, what happens to a profession that has stopped reading its own past?
In the ordinary run of things, whether anyone in the room can explain why is irrelevant. The day something genuinely new happens, something the existing sediment does not cover, is the day it becomes very relevant indeed which practitioners understand that the code is a record of past failures, and which ones believe it is simply the rules.