Taste and the Making of Things

My mum could tell, from across the room, whether a hem had been finished by hand or by machine, and, if by hand, how carefully it had been done. She never went to design school; she wasn’t taught what to look for. She sewed for thirty years, and somewhere in that time her eye learned things her mouth still can't fully explain. She always said the word "well made" just as people use "expensive," as though the two had nothing to do with each other.

I think about her whenever someone says taste is just opinion, just like preferring tea to coffee. It's a comfortable thing to believe. It ends arguments fast and offends nobody. It also falls apart in the presence of anyone who has actually spent years making something with their hands.

Ask a structural engineer whether his load calculations are a matter of personal preference, or a tailor if the perfect cut is just an opinion. Nobody believes that in those rooms. Somewhere along the way, we decided that the moment a craft involves paint, prose or melody, the standards dissolve and every eye becomes equally valid. That belief is convenient. It is also wrong.

Taste is a skill, in the same category as a golf swing or a surgeon's hand. It can either be untrained or sharp. A person who has spent years close to a craft can often sense that something is slightly off before they can say why. That is the residue of attention. A radiologist reads an X-ray differently than someone seeing one for the first time because that eye has been built by ten thousand earlier films, not because of better eyesight. The same thing happens in creative communities, an artist surrounded by excellent work for long enough begins to see through those standards, even before they can name them.

So here's the question worth asking. What does this trained eye actually do? What changes for one person working alone, what changes when a group of people have it together, and what does either have to do with the rare cases where something genuinely new gets made?

The eye that won't let a thing alone

In 1926, much of elite fashion still ran on visible abundance. More beading often meant more money, and many dresses were built to say so. Coco Chanel looked at that whole logic and found it ugly, not in some vague sense but the way a mathematician finds a clumsy proof ugly: too much was happening, and none of it was doing real work. She published a sketch in American Vogue of a plain, straight, calf-length black dress, almost bare of ornament, cut from a simple fabric that stood apart from the heavily embellished couture of the period. Black at the time belonged to mourning and to maids' uniforms. Vogue called it "Chanel's Ford," after the Model T, because like Ford's car, it looked the same no matter who could afford it, and that sameness was the point. Her rival Paul Poiret, who had built an entire career on the abundance she was cutting away, was unsettled enough by it to call the look "deluxe poverty."

What's worth noticing here is the order of operations, not the dress. Chanel didn't start by trying to invent something new. She started by being unable to stand something old. Nothing about the ornament was provably wrong. It simply felt, to a trained eye, like decoration covering for a garment that had nothing else to say. The discomfort came first, and the cutting came after.

That's the first job taste does for someone working alone. It acts like a low hum that won't stop, going off before the maker has any idea what a better answer looks like. Most people, faced with something that already works, leave it alone. A trained eye keeps circling back to the thing it can't yet fix, and that circling is usually the only reason anyone goes looking for a better answer at all.

But taste does a second thing, and it doesn't reduce to the first one as neatly as I would have liked it to.

By the late 1980s, every company making handheld game consoles was racing toward the same targets: faster chips and raw power, and color screens. Gunpei Yokoi went the other way. He built the Game Boy around an old, cheap, gray-scale screen and a processor that was already out of date, on the theory that once a game was fun enough, a kid stopped caring about colors. Nothing here was fixing an error. The color screens his rivals were chasing worked fine. What Yokoi noticed was a kind of open door nobody had tried, because nobody needed to. Many of the color-screen handhelds racing against it lasted only a fraction as long on a set of batteries. The Game Boy went on to sell past a hundred million units and outlasted every machine built on paper to beat it.

That's a different motion than Chanel's. Her eye was repelled by something already there. Yokoi's eye found an empty space nobody else had bothered to walk into, because the whole industry's ambition had marched straight past it without looking down.

These two things taste does aren't really the same mechanism in different clothes, much as it would be tidy if they were. One is judgment: a trained eye measuring something against a standard and finding the gap. The other is closer to noticing room that was sitting there unused the whole time. A creator who only ever does the first produces work that is clean and forgettable. A creator who only reaches for the second, without the discipline of the first, produces work that collapses under its own cleverness. The creators worth remembering are those who can do both.

When the eye stops belonging to one person

A good look at Bauhaus. It ran for fourteen years, moving across three small German cities, and shaped how design, furniture, typography, and buildings look a century later. Plenty of art schools existed at the same time, with comparable students and comparable money, and left almost no trace. That difference is worth sitting with.

It isn't a question of who let more talented students through the door. What the Bauhaus did differently was structural. Walter Gropius set up every workshop under two teachers at once: a "form master," usually a painter or sculptor concerned with composition and idea, and a "craft master," a working tradesman concerned with material and technique. The two didn't always agree about what the work in front of them should become, and a later account of the school describes this dual system plainly as one that was, in its own words, "not free of tensions." A weaving student or a metalworking student answered to two different sets of eyes with two different sets of standards, and had to satisfy both before a piece counted as finished.

This is the part of taste that stops belonging to any single skull. When people with genuinely different training look at the same object, the gap one person misses is often exactly what the other person was trained to catch. What a painter's eye walks past without noticing, a metalworker's eye won't let go of. Put enough differently trained eyes on the same piece of work and you get something closer to a fine net than any one person could weave alone.

A different version of the same thing, with no friction in it at all, is when Jørn Utzon sketched those famous shell-like roofs for the Sydney Opera House. People loved how the design looked on paper, but it was almost impossible to build. Each shell needed its own curve, and each curve needed its own mold. That would’ve been insanely expensive. Utzon and the engineers at Arup spent years wrestling with the problem together before the answer came: cut every shell as a section of one shared sphere, and all the curves become the same curve, reusable with the same mold. The building only became possible once the team let one underlying order generate all the pieces at once rather than treating each piece as its own problem.

No single mind solved that by force of will. It took years of an architect and a room of engineers turning down each other's near-misses, all pointed in roughly the same direction, until the entire problem reorganized itself around something simpler than anyone had been picturing. That's not one eye catching one error. It's a group of trained people slowly converging on the same answer until what was impossible for any of them alone became obvious to all of them together.

What this means for where new things come from

Set those two pictures next to each other and something becomes visible that’s easy to miss looking at only one.

A single trained eye sets a ceiling. It decides what one person, working alone with whatever sense they’ve built over years, is capable of noticing: the open door nobody else saw, the false note nobody else in the room could hear. Without it, a person can be skilled and busy and still never make anything that surprises even themselves.

A group of trained eyes sets a floor, and it sets a pace. A room full of people with genuinely demanding, genuinely different standards catches a weak idea faster than any one of them could alone, and beyond catching errors, it stops people from re-solving a problem someone nearby already solved well. Once Utzon's team found the shared-sphere solution, nobody building a shelled structure after them had to find it again from scratch. Once Yokoi's instinct proved itself, later decisions at Nintendo inherited permission to skip a race their competitors were still running. A field moves quickly to the extent that the people in it are fast to recognize something good and fast to refuse letting something that is merely adequate pass for it.

But the two don't always pull the same way. A group's instinct for catching errors tends to be suspicious of additions that fix nothing, additions that exist purely to delight rather than to repair. A committee is good at finding what's wrong and bad at protecting what's unnecessary but alive. Anyone who has watched an odd, promising idea get sanded flat in a review meeting, by people whose judgment was entirely sound, has felt this directly. Nothing they removed was technically wrong to remove. Something still went missing.

That tension might be why the rarest kind of new thing, the kind that produces delight and not just improvement, tends to need either a creator with some shelter from the instinct of the collective to smooth things over, or a culture unusual enough to let an addition survive even when it can't point to a flaw it fixed. That shelter is expensive to give. From the outside it looks like waste, until the thing it protected turns out to be the only part anyone remembers.

What the eye needs before it can be trusted

To train the eye, trust matters first, in deciding what it even gets shown before there's anything to correct or invent. A person raised on nothing but mediocre work develops a mediocre sense of where the gap is, because they've never stood close enough to the alternative to feel it, which is why the form masters and craft masters sharing a Bauhaus workshop floor were each betting the other's standards were worth absorbing, even mid-disagreement. An eye shaped that way doesn't stay inside one discipline either. It starts borrowing standards from wherever it's been allowed to look, refining a problem in one field with an instinct picked up from another, the way a furniture maker who has spent real time with prose might suddenly notice a rhythm problem in a chair leg that nobody trained only in furniture would catch. None of this knows when to stop on its own. Chanel's eye could have curdled into a different excess, stripping everything on principle regardless of what a garment actually needed, and it didn't, because somewhere underneath the looking and the borrowing sits the part of taste that knows when to show restraint. A creator can have a sharp eye for all of it and still ruin the work by missing that last part.

So if there’s a practical version of all this, it’s not “trust your gut” or “get more feedback.” It’s narrower than that, and harder to act on cleanly. The error-correcting kind of taste is teachable in something close to a straight line: show someone enough bad work and good work side by side, sharpen the comparison enough times, and their eye gets faster at finding the gap. Most design education, most apprenticeship, most editing, is built around exactly this kind of training, because it works. The second kind resists that same training. You can protect the conditions for it. You can refuse to kill an odd idea the moment someone asks what problem it solves. You can build a review process with a rule in it that says some fraction of what gets made doesn’t have to justify itself by what it fixes. But you can’t drill someone into having it, the way you can drill them into spotting a clumsy line. That asymmetry is probably the most useful thing to take from any of this: if you’re creating something, the first kind of taste will improve on its own through exposure, practice, and a refusal to let a flaw sit unfixed once you've seen it. The second kind needs you to deliberately make room for it in your work, or it won’t survive long enough to show up at all.