Form Follows Constraint

Yesterday, Nature Physics published a paper describing a strange thing that happens to waves in a peculiarly shaped cavity. Researchers built an irregular enclosed space from a hyperbolic metamaterial, whose response to waves depends on their direction.

»The designer wants the system to do something; the scientist wants the system to reveal something«

That’s a mouthful worth picking apart: a »hyperbola« is an open curve that runs off toward infinity in two directions rather than closing in on itself, and a »metamaterial« is one that gets its behaviour from engineered structure rather than from chemistry.

The scientists then sent waves through it – not light, and not quite the sound you’d hear with your ears: more like the vibrations of an earthquake, or the ultrasound a doctor scans you with, moving through a solid rather than through air.

Instead of scattering chaotically, the waves converged onto a particular, repeating path. The path has a »handedness« – it always loops one particular way, clockwise or counter-clockwise, rather than mirroring itself – it persists across a broad range of wavelengths, and survives imperfections in the cavity.¹

Ordinarily, this kind of settling-onto-one-path behaviour only turns up in systems that either lose energy as they go, the way a swinging pendulum slows down and eventually stops, or that respond to themselves in complicated, self-amplifying ways, the kind of feedback that produces chaos. This system does neither. It loses no energy, and its response is always simple and proportional. Nothing amplifies, nothing feeds back on itself.

They’ve created something that behaves like an attractor without the usual ingredients. The cavity is irregular, the material is directionally biased, and together those constraints organise the waves into a path that nobody designed. Nothing tells an individual wave where to go. What the researchers have done is construct a world in which certain things happen.

Which is where the experiment becomes more interesting than the result.

We tend to imagine science as observation: a world already there, examined closely enough to reveal how it works. But experiments rarely work like that. Instead, the scientist builds an apparatus, selects materials, and controls initial conditions. The observable behaviour emerges from a deliberately constructed set of constraints.

The physicists didn’t find their wave attractor lying around in nature. They made the conditions under which it could appear, and then watched what followed. This is remarkably close to what a designer does.

Architect Frei Otto understood this well. His soap films and hanging chains weren’t representations of forms already designed. They were systems that generated forms: form-finding instead of form-giving.² Establish the conditions, introduce the material, let gravity and surface tension do their work, and a shape appears. The designer doesn’t design the form so much as construct the conditions it emerges from.

Form follows constraint. But constraint follows form, too. Once a form exists, it becomes a constraint on what can happen next. A roof directs water. A structural frame limits where loads can go. Every designed form is also a small behavioural machine.

And suddenly, design and science seem to touch. Design is the deliberate construction of constraints from which desired behaviour emerges. Science is the deliberate construction of constraints from which observable behaviour emerges. The difference is tiny, but it matters: the designer wants the system to do something; the scientist wants the system to reveal something.

Molecular biologist and historian of science Hans-Jörg Rheinberger’s idea of the »experimental system« captures the scientific side of this particularly well. An experiment isn’t just confirmation of something already known. It’s an arrangement that generates phenomena, making something observable that wasn’t before.³

The wave attractor is an unusually clean example. The researchers constructed an environment in which a particular behaviour could emerge, and then discovered what that environment made possible. We tend to think that control means issuing instructions, telling something what to do. But there is another kind of control: instead of specifying the behaviour, construct the world in which the behaviour emerges.

The wave doesn’t follow an instruction. It’s just never given anywhere else to go.

References