The Parthenon Principle

Yesterday, The Guardian and The Independent reported that a mathematician has challenged one of architecture’s favourite stories.¹ The Parthenon is the ancient Greek temple on the Acropolis in Athens. Built in the fifth century BC, it has become one of the most studied buildings in the world, and an enduring symbol of ancient Greece, Western civilisation, and democracy.²

»A hypothesis can be true and inert. It can be true and productive. It can be false and forgotten. Or it can be false and enormously productive«

But look closely and it is full of deviations from perfect geometry. Its base rises towards the centre. Its columns swell slightly in the middle. Its walls and columns lean inwards. These deviations are real. Why are they there?

The favoured explanation has been that they were deliberate optical corrections. The Greeks, it was said, knew that a perfectly straight base would appear to sag, while columns with perfectly straight sides would appear to taper too sharply. Alain Goriely, a mathematician at Oxford, has now modelled the supposed effects and found them either unsupported by evidence or too small to perceive.³

Maybe the Greeks got the optics wrong. But more interestingly: maybe we got the Greeks wrong.

The optical theory was not simply handed down from antiquity. The Parthenon’s refinements were measured closely in the nineteenth century, and in 1851 Francis Penrose proposed that they were intended to correct optical effects. Meanwhile, historians of Greek architecture have shown that curvature, inclination, and other refinements were already established when the temple was designed.⁴ We have assumed that an explanation came first, and the architecture followed. But maybe it was the other way around – architecture came first, then came observation, then came explanation.

There’s a difference between saying that something is the case and explaining why it is the case. »The roof is white« is a description. »The roof was painted white to reflect heat« is a hypothesis. It introduces something invisible: a cause. A hypothesis doesn’t even have to be believed. It can be held in suspension long enough to ask what would follow if it were true. That’s when an explanation becomes a model: something we can experiment with, design from, predict from, or use to make something else.

Let’s call what happens next the Parthenon Principle.

A building produces an interpretation. The interpretation becomes a model. The model produces another building. That’s a strange reversal between the synchronic and the diachronic: between looking at something as it exists at a particular moment, and looking at how things unfold through time. We see the building as it is, then explain it. But once the explanation becomes something we use, it enters the history of what comes next. Our account of the first building becomes part of the cause of the second.

The Parthenon story is interesting even if it’s wrong. If its curves really were optical corrections, an architect could turn that explanation into a design rule: straight lines need correcting, so introduce curvature; straight-sided columns look wrong, so give them entasis (make them slightly convex). But if the explanation is wrong, the mechanism still works. An architect can believe it, teach it, design according to it, and produce buildings that wouldn’t otherwise have existed. The false explanation becomes a real cause.

And this happens outside architecture. Phlogiston was an incorrect explanation of combustion, but it organised generations of experiments. Researchers built investigations around it until the evidence helped destroy it. The explanation was false, but it had become an experimental machine. A hypothesis can be true and inert. It can be true and productive. It can be false and forgotten. Or it can be false and enormously productive. Take a bank run: if depositors believe a bank is about to fail, they withdraw their money. If enough people follow suit, the bank eventually fails. The belief has changed the thing it was supposed to describe.

Buildings are explanations made physical. Today’s generative design makes the loop explicit: we make something, assess it, infer something from the result, and make again. The assessment isn’t merely judgement after the fact. It becomes part of the mechanism producing the next thing.

Perhaps that’s what the Parthenon Principle really says. An explanation doesn’t merely tell us what something is. Once we start using it, it becomes part of what the thing becomes. Maybe Goriely is right, and the Parthenon wasn’t built according to the explanation.

Still, other things were.

References