Who Is John Doe? The 700-Year Story of a Name, From Courtrooms to AI Datasets

You open a website's demo page: it says John Doe under the profile picture. You look at an API documentation page: `{"name": "John Doe", "email": "john.doe@example.com"}`. You're watching an American crime series: the tag on the body in the morgue reads John Doe again.

Is this the same man? No. Actually, he's nobody.

John Doe is the world's most famous person who does not exist. And his story goes back much further than most people think — to 14th-century England.

1. The Short Answer: Who Is John Doe?

John Doe is a placeholder name used across the English-speaking world. He shows up in one of three situations:

• Identity unknown — A person whose name has not yet been determined (for example, an unidentified body).

• Identity withheld — A person whose name is known but not disclosed for legal or ethical reasons (victims, witnesses, plaintiffs).

• There is no identity to begin with — No real person is meant; you simply need an example of "any human" (forms, demos, test data).

The family tree:

• John Doe — Unknown/withheld male

• Jane Doe — Unknown/withheld female

• Baby Doe / Precious Doe — Unidentified infant or child

• Richard Roe — The second party (opposing side, defendant)

• John Stiles, Richard Miles — Third and fourth parties

• John Nokes — An older variant, now obsolete

• James Doe, Judy Doe — When one case has multiple anonymous parties

2. The Origin: The Invention of a Tenant Who Never Existed

These names can be traced back to the 1300s and English common law.

At the time, land disputes were handled through a procedure called an "action of ejectment." The problem was this: the procedure had an extremely rigid formula and could only be brought by tenants. But the person wrongfully dispossessed of land was usually not a tenant — it was the owner.

Lawyers found a clever — and entirely made-up — solution: the legal fiction. They would write a little story into the pleadings:

"Suppose the owner leased his land to a tenant. And suppose that tenant was forcibly thrown off it."

This hypothetical tenant needed a name. By convention, the name chosen was John Doe. The hypothetical party on the other side was named Richard Roe.

Why "Doe" and "Roe"? There's no definitive proof, but the most common explanation is this: doe (a female deer) and roe (a roe deer) — two of the most ordinary, most generic animals of the English countryside. Neutral words that belong to no one and carry the idea of "anyone."

This fiction was formally abolished in England in 1852. But the names stayed. Even after the fiction they served disappeared, John Doe and Jane Doe had settled into legal language with the meaning of "anonymous party."

The written record

• The names appear in documents from at least the 17th century (one instance occurs in a work dated 1659).

• The best-known early example is in Sir William Blackstone's Commentaries on the Laws of England (1765–69) — a foundational text of Anglo-American law, and a major reason the name crossed the Atlantic.

3. "Jane Doe" and the Case That Changed History

Jane Doe is a much more recent addition than John Doe. And she owes her worldwide recognition to a single case:

Roe v. Wade (1973).

The plaintiff was a woman whose real name was Norma McCorvey. To protect her identity, she entered the court record as "Jane Roe." The opposing party was Dallas district attorney Henry Wade. The case established a constitutional right to abortion in the United States (overturned in 2022 by Dobbs) and carried the "Doe/Roe" names far beyond legal circles.

This illustrates the placeholder name's most powerful function: anonymity is an instrument of access to justice. In situations where you could not bring a case at all if you had to disclose your name — sexual assault survivors, whistleblowers, immigrants, children — "Doe" opens the courthouse door for you.

4. Where Does John Doe Show Up Today?

4.1. In Law

• Anonymous plaintiff: A party wishing to protect their identity files suit under the name "John Doe."

• The "John Doe" subpoena: A suit filed when the defendant's identity is not yet known. It's especially common in internet litigation — to find an anonymous user committing copyright infringement or defamation, you first sue "John Doe," then obtain a court order compelling the internet service provider to disclose the real identity.

• The "John Doe" indictment: In some U.S. states, when the statute of limitations is about to expire and you have the perpetrator's DNA profile but not their name, charges are filed against the DNA profile itself. This stops the clock.

4.2. In Forensics and Medicine

Unidentified bodies enter hospital records and morgue tags in numbered form, like "John Doe #14, 2024." Unconscious emergency room patients carrying no identification are logged the same way — because treatment requires a file name, and you cannot wait for a name.

In the U.S., databases such as NamUs (National Missing and Unidentified Persons System) work to match thousands of "Doe" files against missing person reports. Some of these files are solved decades later through genetic genealogy, and a "Doe" finally recovers their real name.

4.3. In Popular Culture

• Frank Capra's 1941 classic "Meet John Doe" permanently bound the "ordinary American" figure to the name.

• In Se7en (1995), the killer's identity is given as "John Doe"; namelessness is itself part of the character's message.

• The series John Doe (2002) follows a man who has lost his memory.

The cultural weight matters here: John Doe means not just "unknown person" but also "everyone." The average person, the ordinary citizen, the face in the crowd.

5. So What's the Connection to AI and Software?

This is where the name's second life begins. And this use — less well known than the legal origin — is far more frequently encountered today.

5.1. Demo data, test data, sample data

A software developer constantly needs an "example user":

{ "id": 1, "name": "John Doe", "email": "john.doe@example.com", "role": "user" }

Why not a real name? Three reasons:

• Privacy and law. Writing a real person's name, email, or phone number into a test environment violates GDPR (and Turkey's KVKK). When a screenshot gets shared, a log file leaks, or a demo environment goes public, a real human's data spills out. The combination of `John Doe` + `example.com` (a domain reserved for testing purposes by RFC 2606) reduces that risk to zero.

• Cognitive load. When a reader sees a familiar pattern, they don't pause to ask "is this real?" The name becomes invisible and the code becomes visible — which is the whole point of documentation.

• Not accidentally hitting a real person. A randomly invented name may belong to someone who actually exists. "John Doe," by contrast, culturally signals "this is nobody."

5.2. In the AI context

John Doe appears at several different layers when working with artificial intelligence systems:

a) In prompt and documentation examples.

In nearly every example showing how to write a prompt for a language model, "John Doe" stands in for a real customer/patient/employee. The reason is the same: demonstrating structure without sharing real data.

b) In synthetic data.

One of modern machine learning's central dilemmas is this: training a model requires enormous amounts of data, but using real personal data is both risky and often illegal. The solution is synthetic data — artificial records that mimic the statistical properties of real data while belonging to no actual individual. The names in these datasets are the classic "John Doe" logic applied at industrial scale — except now it's not one name, but millions of realistic-but-fake names generated by libraries like `Faker`, `DataFaker`, or `Mimesis`.

There's an important distinction here: "John Doe" is a single generic placeholder; synthetic data imitates a distribution. If you train a machine learning model 100,000 times on "John Doe," the model never learns name diversity — in fact, it will weight that name abnormally. So John Doe is fine for demos and documentation, but model training requires realistic variety.

c) In anonymization and data masking.

When sharing a dataset for analysis, personal fields have to be replaced. In this process, real names are swapped for placeholders — or for pseudonyms (pseudonymization, separately defined in GDPR Article 4(5)).

d) In AI output.

Ask a language model to "write a sample employment contract" and John Doe will very likely appear. This is a pattern the model learned from the millions of templates, forms, and sample documents in its training data. In other words, AI's fondness for John Doe is no accident — he is the common resident of every example document humanity has ever produced, so that's how he settled into the model.

5.3. One thing worth watching

"John Doe" looks neutral, but he isn't entirely neutral. It is a white, male, Anglo-Saxon name. If all of an AI system's sample data is built on John Doe / Jane Doe, then the way that system gets tested is also confined to a narrow demographic. Different alphabets (ç, ğ, ş, ü / Chinese characters / Arabic), single-word names, very long surnames, names with apostrophes (O'Brien) — these exist in the real world, and they usually surface as bugs for the first time in production. A good test dataset goes beyond John Doe.

6. What Is John Doe Called in Other Countries?

Every culture has its own "man with no name":

• USA / Canada — John Doe — Jane Doe

• United Kingdom — Joe Bloggs, John Smith — Jane Bloggs

• Germany — Max Mustermann — Erika Mustermann

• France — Jean Dupont, Monsieur Tout-le-monde — Marie Dupont

• Italy — Mario Rossi — Maria Rossi

• Spain / Latin America — Juan Pérez, Fulano de Tal — Fulana

• Sweden — Medelsvensson — —

• Japan — Yamada Tarō (山田太郎) — Yamada Hanako (山田花子)

• Russia — Ivan Ivanovich Ivanov — —

And in Turkey?

Turkish has no single standard; it varies by context:

• Everyday speech: Falanca, filanca, feşmekânca — "Falanca said that..." These words come from the Arabic root fulān; the same source as Spanish Fulano.

• Examples and paired narration: Ali and Veli, Ayşe and Fatma, Ahmet, Mehmet — "Ali has 5 apples..."

• Official forms and demo screens: Ahmet Yılmaz, Ayşe Yılmaz, Mehmet Demir — the statistically most common first-name/surname combinations are preferred.

• Forensic/medical records: "Kimliği tespit edilemeyen şahıs" ("person whose identity could not be determined"), often abbreviated K.T.E. Turkish legal and forensic practice uses a phrase, not a name. This is a meaningful difference: the Anglo-Saxon system gives the unidentified person a name, while the Turkish system documents the absence of one.

• Legal anonymity: In Turkish court decisions, people are generally referred to by initials (A.Y., M.D.) — an abbreviation is preferred over a Doe-style pseudonym.

7. Frequently Asked Questions

Is John Doe a real person?

No. That said, there are people in the world who genuinely bear the name "John Doe," and it's a constant source of trouble for them — systems may assume their name is fake data and reject it.

What does "Doe" mean?

In English, a female deer. "Roe" is a roe deer. The general view is that these words were chosen precisely for their ordinariness and neutrality.

Why is it still used? It isn't needed anymore.

Because its function changed. In the 1300s it was needed to get around a procedural rule; today it's needed to protect privacy, enable access to justice, and keep technical documentation readable.

Why does AI keep writing John Doe?

Because the name appears in a very large share of the human-produced templates, examples, and documentation in its training data. The model has formed a strong association between this name and the context of "I need an example person."

Should I use John Doe in my own project?

For documentation, demos, and interface mockups: yes, it's a safe and familiar choice. But if you're building a Turkish-language product, a local placeholder (Ahmet Yılmaz) will feel more natural. For model training or comprehensive testing, generate synthetic data with real variety instead of a single name.

8. Conclusion: A Name for Namelessness

John Doe has been doing the same job for 700 years: filling a gap in a way that makes the gap visible.

In the 14th century he was an imaginary tenant invented to get around a rigid procedural rule. In the 20th century he became a shield that let people who could not disclose their names go to court. In the 21st century he works as a quiet layer of protection in API documentation, synthetic datasets, and AI output, preventing personal data from leaking.

The common thread never changes: John Doe stands in for a real person in order to protect that real person.

The next time you see that name on a demo screen, you'll actually be looking at a tenant from 1300s England who never existed at all.

Sources

• John Doe — Wikipedia

• John Doe: Origins of America's Placeholder Name — World Wide Words

• Why Are Anonymous People Called "John Doe" in Legal Matters? — A Way with Words

• Why Are Unidentified People Called John or Jane Doe? — Mental Floss

• John Doe: The History of Placeholder Names — Fakenamely

• How Do Other Countries Refer to 'John Doe'? — Blend

• What Is Synthetic Data? — IBM

• How Generative AI Is Revolutionizing Training Data with Synthetic Datasets — Dataversity