# The Name Was the Credential
> Smith, Baker, Cooper, Fletcher. A name was once a checkable claim about what you could do, who trained you, and who was liable when the work failed. Nothing since has done all three.
Author: Kariaa Research
Published: 2026-07-25
Category: Insights
Reading time: 14 min read
Canonical URL: https://www.kariaa.com/reports/the-name-was-the-credential
---
import {
Timeline,
CompareColumns,
RankedList,
FlowSteps,
Takeaways,
} from "@/lib/report-charts";
Smith, Baker, Cooper, Fletcher, Chandler, Wright, Thatcher, Mason, Turner,
Fuller. Roughly a third of the most common English surnames are job titles that
stopped being job titles. For several centuries a person's name was a public,
checkable claim about what they could do, who had trained them, and who could be
held answerable if the work failed. That system was hereditary, exclusionary,
and largely closed to anyone born outside it. Dismantling it was correct. It is
also the last time verification and identity were the same object, and nothing
built since has done the job as well.
## A name that made a claim
**Occupational surnames were not decoration. They were the output of a
registration system that ran on everyday speech.** When surnames became
hereditary in England, over roughly the twelfth to fourteenth centuries, a
substantial share of them fixed a person's trade permanently into the way
everyone referred to them. A Fletcher made arrows. A Cooper made barrels. A
Chandler made candles. A Wright built things, which is why the name survives
mostly in compounds like wheelwright and shipwright. A Fuller cleaned and
thickened cloth. A Barker tanned hides using bark.
The naming was not unique to England, and the parallel construction across
unrelated languages is the point. German surnames encode the same trades:
Schmidt the smith, Müller the miller, Schneider the tailor, Zimmermann the
carpenter. In the Arabic-speaking world the *nisba* attached origin, lineage, or
occupation to a name in the same way, producing al-Attar the perfumer, al-Khayyat
the tailor, al-Warraq the paper seller and copyist. In much of South Asia,
occupation was bound into social identity far more rigidly than a name, in a
structure whose immobility is precisely the cost this essay is about.
Independent societies with no contact converged on the same design because they
faced the same problem. Before written records were general and portable, the
cheapest place to store a claim about someone's competence was inside the thing
everyone already had to say out loud.
## What the guild actually certified
**The name pointed at a training system that took years and produced a person
someone else was liable for.** Guild apprenticeship in England was formalized by
the Statute of Artificers in 1563, which set a seven-year term for entry into a
wide range of trades. Seven years is not a credentialing formality. It is long
enough that the master and the apprentice live in the same house, and long
enough that the master's assessment of the apprentice is based on thousands of
observations rather than an interview.
Three features of that arrangement matter, and none of them survive in modern
credentials.
The training was conducted by someone with skin in the outcome. A master who
certified an incompetent journeyman damaged their own standing in a guild they
could not leave, in a town where everyone knew whose workshop the person came
out of. The certification carried the certifier's reputation, not just their
signature.
The assessment was of demonstrated work. Advancement to master typically required
producing an actual object judged by people who made the same objects for a
living. The evaluation was of the thing, by the only people qualified to judge
it, under conditions where a bad judgment cost them.
And the credential was revocable by a body with real power. Guilds could exclude,
and exclusion from the guild meant exclusion from the trade in that town. The
certificate and the sanction lived in the same institution.
## The mark on the metal
**The purest verification system anyone has built was a punch struck into
silver.** English hallmarking, which statute law required from 1300, worked by
attaching the verification to the object rather than to a document about the
object. A piece of silver carried a mark for the assay office that tested it, a
mark identifying the maker, and eventually a letter recording the year. The
metal's purity was not asserted. It was tested, and the test was recorded on the
thing tested, permanently, by a body that could prosecute.
Consider what that gives a buyer who knows nothing about metallurgy. They do not
need to trust the seller. They do not need to understand assaying. They need only
recognize a mark, and the entire chain behind it, the assay, the office, the
named maker, the guild with power to punish, resolves into a claim they can act
on in a second. The cost of verification for the buyer is approximately zero
because the cost was paid once, upstream, by parties who could be punished for
paying it dishonestly.
That is what a credential is supposed to do, and it is worth naming the property
precisely: the mark was traceable to a person. Not to a school the person
attended, not to a category the person belonged to, but to the individual whose
hands did the work.
## A credential is a cache
**Every credential is stored trust: someone did the checking once, and everyone
downstream reads the record instead of repeating the work.** This is an
efficiency, and a large one. Verifying competence directly is slow and expensive,
often impossible for the person who most needs it done, and a cached result that
is right most of the time beats a direct check that never happens.
Caches are also governed by a small number of properties that determine whether
they are worth reading, and these apply to credentials with no modification.
**What is stored.** A cache holds an answer to a specific question. The guild
mark answered whether this person can do this work to this standard. A degree
answers whether this person completed a course of study at this institution.
Those are different questions, and only one of them is the question the reader is
actually asking.
**Who wrote the entry, and what it cost them.** An entry written by a party who
suffers when it is wrong is worth more than one written by a party who does not.
The master's certification cost him standing if it was false. An institution that
graduates a weak student bears no consequence traceable to that student.
**How it is invalidated.** A cache with no eviction policy accumulates entries
that were true once. The guild could revoke. A degree, once granted, is
permanent regardless of whether the holder has practised the skill since, and
most professional credentials behave the same way.
**How fast it goes stale.** The rate at which the underlying reality changes,
against the rate at which the entry updates. A cooper's competence in 1600
changed slowly across a career. Most technical competence now changes faster
than any credential recording it.
## Liberalization was correct, and it broke the cache
**The old system's accuracy came from the same property as its cruelty.** It
worked because it was closed. Membership was largely hereditary, entry was
controlled by incumbents who benefited from restricting it, and a person born
outside a trade generally stayed outside it. The system knew exactly who was
competent because it decided in advance who was permitted to become competent.
Opening that up was one of the clearer moral improvements in economic history and
nothing here argues against it. But it is worth being precise about what
happened mechanically. Access widened enormously. The verification need did not
change at all. Employers still had to answer the same question, whether this
person can do this work, about a vastly larger and more mobile population, and
the institution that used to answer it had been correctly dismantled.
What filled the gap was not a better verification system. It was a set of
proxies that were available at scale: the name of an institution, the name of a
previous employer, a job title, a duration. Each of these correlates with
competence. None of them measures it. And critically, none of them identifies
anybody who can be held to the claim, which is the property that made the
hallmark worth a second of anyone's attention.
## What we actually replaced it with
**The modern credential answers a question adjacent to the one being asked, and
the gap between them is where hiring fails.** An employer wants to know whether
someone can do the work. The available evidence says where they studied, what
they were called, and for how long. Everyone involved knows the substitution is
happening. The résumé exists to make the substitution look reasonable, and the
interview exists to partially undo it.
The consequences follow directly from the cache properties.
Because the entry records category rather than capability, people who possess the
capability without the category are invisible, and the loss is not distributed
randomly. It falls on anyone who acquired their competence outside a recognized
institution, which correlates almost perfectly with not having had access to one.
Because there is no eviction policy, a credential earned decades ago reads
identically to one earned last year. The cache returns a hit either way.
Because the issuer bears no cost for a wrong entry, there is no force acting to
improve accuracy over time. Institutions compete on selectivity at intake, which
is a claim about the input rather than the output.
And because the entries are cheap to produce and getting cheaper, the ratio of
signal to assertion in the whole system is falling. A credential that costs
little to obtain and nothing to fake conveys correspondingly little, and every
party involved compensates by reading it more skeptically, which degrades it
further.
## What direct verification has to beat
A cache is only worth replacing if the direct check is cheap enough to run and
better than the stored answer. That sets a specific bar, and most proposed
replacements do not clear it.
It has to assess the work rather than the biography. The hallmark's real
innovation was testing the object. Any system that evaluates a description of
what someone has done, rather than something they have actually done, is
producing another proxy with extra steps.
It has to name someone who is answerable. A verification that cannot be traced
back to a party who suffers when it is wrong is an assertion. The maker's mark
was compulsory from 1363 for exactly this reason: anonymous claims about quality
were understood to be worthless, and the response was to make anonymity illegal.
It has to expire. Competence is a rate, not a possession. Any credential that
never goes stale is guaranteed to be wrong about a growing fraction of its
holders, and the ones it is wrong about are indistinguishable from the ones it
is right about.
And it has to be cheap enough at the point of reading that people actually use
it. The hallmark's genius was not the assay. Assaying was expensive. The genius
was compressing an expensive upstream process into a mark a buyer could read
instantly, which is the only reason the expensive part was worth doing.
None of this requires rebuilding a closed system, and the closure was never the
source of the accuracy. The accuracy came from testing the work, naming the
tester, and keeping the power to revoke in the same hands that granted. Those
three properties are separable from hereditary exclusion. We have simply not
bothered to separate them.
---
*This report was produced by **Kariaa Research**. All data, resources, and
analysis are proprietary. For questions, contact
[research@kariaa.com](mailto:research@kariaa.com).*