Between Craft and Quality

Something you owned for years finally gives out. You feel like you got your money’s worth. It served you well. So you go online and buy the same thing again, the same model from the same brand. Only when it arrives, you notice immediately that something is different.

Nothing is broken. No one thing is obviously cheaper. You turn it over in your hands looking for whatever changed but cannot find it. After a while, you start questioning yourself rather than the item. You have gotten older. You are fussier now. Maybe the first one was never as good as you remember.

I have seen some version of that story dozens of times through Worse on Purpose’s tip line. A perception of some reduction in quality, impossible to pin down, followed by expressions of self-doubt.

In ‘On Quality’, I argued that quality exists in the connection between an object and the person interacting with it, rather than as some feature of the object itself. That essay focused on the implications of this view for companies. Firms now measure performance more closely than they ever have. The things that make a product good rarely show up in those numbers, so quality quietly drains away while executives optimize only what they can measure.

This essay is about our relationship to quality as individuals. Companies stopped seeing quality because they narrowed their attention only to what they could measure. What happened to us is different.

Over the past two centuries the things we buy were quietly rebuilt so that most of what determines their quality exists where no buyer can see it. At the same time, we lost the skills that enabled us to judge quality on the parts that still show.

Both losses have a single cause.

On the maker’s side, production withdrew into factories and distant trades. When things were made in the household or the town, the manufacturing processes sat in plain sight. We knew what went into the objects we owned because we saw them being produced. Now the things that determine whether a product will last are completely hidden from view.

Meanwhile, the same forces pushed each of us deeper into narrower fields of work. We do one thing for a living and outsource the rest. But the ability to judge how well a thing is made comes from having made things. The fewer kinds of things we make, the fewer kinds of things we can judge.

What follows is an account of how that happened, and what it cost us.

Hand an object to someone who has spent a lifetime around examples of its kind. A jacket to a tailor, a stool to a carpenter or a knife to a chef. They’ll turn it over once or twice and deliver their verdict. “It’s good” or “not something I’d use.” The whole examination takes only a handful of seconds. Those short moments deserve close attention.

The expert-examiner is performing an inspection. They are holding the object against a standard and determining whether the object meets it. The verdict is entirely intuitive. That is expertise.

But where did the standard that the examiner is applying come from?

It came from having lived with various examples of the object in question for years. It came from a loose familiarity with how such objects should go together. It’s a familiarity that can only be acquired through decades of use.

We are all carrying similar capabilities for judgement. In the domains where we have put our hands to work, we can tell when things are not what they were.

That raises something to settle before going further. If the examiner is detecting something real, where is that thing? Inside the object itself? Or somewhere else?

Suppose quality is a property that exists within an object. If that were true, then quality must be findable. We must be able to create some instrument to detect it. Imagine the finest jacket ever made. Take it to a laboratory and try to design a test for its ‘quality’. You can’t. You can measure the tensile strength of the thread or the abrasion resistance of the fabric, but you cannot design an objective test measuring the overall ‘goodness’ of the jacket.

Suppose instead that quality exists only in the perceiver’s mind. Then it is nothing more than a simple preference. Quality just means whatever you happen to like. This can’t be right either. We agree on quality, even if we can’t express why. We have also all had the experience of buying something on the perception of quality, only to find we were wrong in use. If quality is just a feature of our minds, then the gap between our quality perception and the true quality of the object makes little sense.

So if quality is not a property within an object, nor just a perception in our minds, then what is it?

Robert Pirsig provides, in my view, the most satisfying answer in Zen and the Art of Motorcycle Maintenance. Quality, he says, is not a thing, but an event. Quality arises when a subject meets an object in use. It exists only in the connection between the object and its user. It does not exist independently in the jacket or in you. Quality is what happens when you pick the jacket up, put it on and step into the wind and rain.

This raises a problem: if quality happens in the meeting between an object and its user, then any verdict on the quality of the object will depend on who is delivering it. Different users will reach different verdicts. Something must separate a good judge from a bad one, otherwise we are back to simple preference.

Two centuries before Pirsig, David Hume laid out his criteria for the competent judge:

This is essentially a job description. It describes the preconditions of expertise. Pirsig tells us where quality happens; in the meeting between an object and its user. Hume tells us the meeting of those two things only produces a reliable verdict if the user brings a trained faculty to it, built out of lifelong practice and comparison.

So quality itself has a precondition. Somebody must be skilled enough in their domain to detect it.

A quality nobody can detect is a quality nobody pays for. A quality nobody pays for eventually stops being made. Everything, then, turns on what a buyer can detect.

There are some properties of an object that you can check yourself, standing in a store, with no special training whatsoever: how the fabric hangs, whether there is any play in the joint, whether a chair will support your weight. These are observable traits.

Others, you have never been able to check at any time, no matter how skilled you were: the purity of the silver, whether the flour has chalk in it, the true provenance of the grapes in the wine. These are concealed traits.

Societies have taken the problem of concealed traits incredibly seriously in the past. They built protective institutions, and ran them for centuries.

English cloth aulnage began in 1197 and did not stop until 1699. Five hundred years where a salaried official’s entire job was to measure every piece of woolen cloth offered for sale, stamp it with a seal if it passed, seize it if it did not, and collect fourpence for the service. The reason is purely practical. Cloth was sold folded. You cannot measure twenty-four yards of folded broadcloth across a market stall, so the state sent a man with a yardstick to do it on your behalf.

Hallmarking began in London in 1300 and is still the law. If you sell something as silver in Britain today, it must carry a mark from an assay office.

The Trial of the Pyx has been testing the coin of the realm since 1282. It tested 7,960 coins in 2024. It is still operational.

Notice the things no institution was created for. Nobody was ever appointed to certify that a chair was sturdy. There is no medieval office of the Sitter. You could just sit on it yourself.

The institutions cluster on the concealed side and leave the observable side entirely alone. This tells you that past societies were not confused about what a buyer could and could not judge at the point of sale. They knew exactly where the line ran, so they spent resources policing it.

The line marks the ratio of observable traits to concealed traits in the marketplace. For most of history it was stable. Most traits were observable, and the few that were not could be policed.

That ratio has since been overturned, from both ends at once. Products grew complex. More materials, more processes, more of their fate decided in places no buyer will ever see. The concealed share of traits within ordinary products exploded. At the same time, we grew narrow. Each of us commands one trade and buys the output of every other, so even the traits that remain visible are harder for us to judge competently.

Most of what determines quality is now concealed. Most of what still shows, we can no longer judge. A quality that cannot be verified cannot be charged for, and what cannot be charged for will cease to exist.

After the paywall: the cause. The concealment of quality and the loss of our ability to judge are two consequences of a single mechanism. What follows names that mechanism, measures it and shows it at work in the production of a simple shirt across four centuries.