Does changing the name actually protect you?

You called him Ray. In the manuscript he is Ray, he works somewhere else, and you moved the story one state over. Part of you thinks that settles it. The other part is why you are on this page, and the other part is right.

Renaming someone is a courtesy, and a signal to your publisher that you thought about the problem. I would rather you do it than not. As protection it does less than almost every author expects, because the name is not what the law is looking at. In my own review years it was the change authors were proudest of and the one that moved the analysis least.

Identifiability is the question underneath. In defamation it lives in the element courts call “of and concerning” the plaintiff, which asks whether readers understood the passage to be about this particular person. Privacy claims ask a version of the same thing before they ask anything else. Neither one asks what you called him. The full map of how the exposure categories fit together is on the main guide, What You Can and Can’t Say.

What does the law test when you change a name?

The law tests whether readers could tell the passage was about your subject. Every defamation claim has to clear an identification element, and the name is one input into that question rather than the question itself.

The requirement is real, and plaintiffs lose on it. In New York Times Co. v. Sullivan, 376 U.S. 254 (1964), the Supreme Court refused to treat an impersonal attack on a police department as a libel of the commissioner who ran it, holding that an of-and-concerning showing resting on the bare fact of his office was constitutionally insufficient. Criticism of an institution is not automatically criticism of the person running it.

The element runs the other way too, and that half is the one authors miss. A plaintiff does not have to be named at all. The Second Circuit put it in a line worth keeping: “It is not necessary that all the world should understand the libel.” Geisler v. Petrocelli, 616 F.2d 636, 639 (2d Cir. 1980). What matters is whether the people who knew the plaintiff could work out that she was the person meant, and extrinsic evidence is admissible to show it.

One feature of this element surprises working editors. Courts do not agree on who decides it. New York treats identification as a question for the court and threw out a novel case on the pleadings on that basis. Springer v. Viking Press, 60 N.Y.2d 916 (1983). Geisler said it generally goes to the trier of fact, and the Eighth Circuit called it a jury question read from the story as a whole. Ruzicka v. Conde Nast Publications, Inc., 999 F.2d 1319 (8th Cir. 1993). The consequence is money. Where identification reaches a jury, a weak claim lives longer and costs more to answer, which is the subject of “I don’t care if I get sued.” Your publisher does.

Which details actually identify a person?

Role, place, time, relationship, and the events themselves. The people most likely to complain already have all of them.

Ruzicka is the cleanest illustration I know. A magazine promised to conceal a source, then called her Jill Lundquist, described her as a Minneapolis attorney, and reported that she had served on the state task force behind a 1985 law on therapist-patient sex. She alleged she was the only woman lawyer on that task force. The name was invented. The description was a fingerprint. Deciding a broken promise of confidentiality, and borrowing the defamation identification standard to do it, the Eighth Circuit reversed summary judgment and sent identifiability to trial. The test is not the author’s intent and not the subject’s fear. It is the reasonable understanding of the reader.

Appearance and geography do the same work. In Elias v. Rolling Stone LLC, 872 F.3d 97 (2d Cir. 2017), two members of a fraternity chapter were held to have plausibly alleged that an article was of and concerning them individually, based on descriptions of how they looked and where their rooms were. No names appeared. Elias also carries a limit worth knowing: where a plaintiff relies on facts outside the publication to make the connection, those facts have to have been known to the people who read it. A detail only your subject could decode is not identification.

Volume counts as much as any single detail. In Smith v. Stewart, 291 Ga. App. 86 (Ga. Ct. App. 2008), a novelist built a character out of a friend’s life, and the trial court counted at least twenty-six specific similarities: hair color, chronic lateness, chain-smoking and the cough that went with it, her parents’ occupations, her childhood home, and facts about her life that were not public before the book. The defendants conceded the character was based on her and recognizable as her, and pointed to real differences, including the changed name, a sorority the plaintiff never joined, and a club she did not belong to. The Georgia court held those differences made a jury question rather than a defense.

Identification also fails, on facts that look uncomfortably close. Springer involved a minor character who shared the plaintiff’s given name, height, weight, build, grooming habits, and recreational activities, and New York’s high court held that combination insufficient as a matter of law. In Bates v. P.C. Cast, 316 P.3d 246 (Okla. Civ. App. 2013), a young-adult novel about a school for vampires gave a character a real woman’s full name, and the author had publicly identified her as the real Erin Bates. The court found no identification anyway, because the character was a teenager, the plaintiff was in her mid-twenties, and the setting was invented. Readers can see the name and still not believe the character was meant to be her.

Read the four together and a working rule falls out. Identification tracks the density and specificity of what surrounds the name, not the name.

What each change actually removes

Each row is a tendency, not a legal grade. Identification is decided case by case, on the work as a whole, under the law of the state where a claim is brought.

Does calling it fiction change the analysis?

No. Courts apply the same identification test to novels that they apply to journalism, and a disclaimer does not supply a defense the facts have not already earned.

Bindrim v. Mitchell, 92 Cal. App. 3d 61 (Cal. Ct. App. 1979), is the case to know before you start renaming. A novelist attended a psychologist’s therapy sessions and wrote a novel about a version of them. In the book the therapist was heavy, bearded, and white-haired, under a different name. The real psychologist was clean-shaven with short hair. Identification was found anyway, on testimony from people who recognized his way of talking and his methods. The court asked whether a reasonable person reading the book would understand the character “was, in actual fact, the plaintiff acting as described.” Changing how a character looks is the change that matters least to readers who know your subject.

One caveat on Bindrim belongs in the open. It is a 1979 decision from a California intermediate appellate court, and citator research flags it for caution, so treat it as illustration rather than as the rule in your state. The identification test it states appears in the same form in later decisions elsewhere, including Smith v. Stewart and Bates.

The Restatement (Second) of Torts states the rule at section 564, comment d, and it cuts both ways. A libel can be published of a real person through a work meant to deal only with invented characters, where the resemblance is close enough that readers reasonably understand a character is intended to portray that person. Similarity of name alone is not enough, and neither is recognizing a resemblance, unless readers also reasonably believe the character was meant to portray the person.

The copyright-page disclaimer does no work on its own. Geisler framed the question as whether a reasonable reader would rationally suspect the protagonist is the plaintiff, whatever the author and publisher say about the book being fiction. Smith v. Stewart made the point from the other side: where an author sets the action in real restaurants and hotels, researches details for accuracy, and builds a character out of a real person’s biography, the label will not stop readers from taking the book as describing actual events. Keep the disclaimer. Do not mistake it for protection.

Can someone sue if only a few readers recognize them?

Yes, and the small readership is usually the dangerous one. The identification element asks whether people who knew the subject would recognize them, so a portrait that means nothing to a stranger and everything to a sister is exactly the shape that produces claims. Wide recognition is not the test, and it never has been.

The California court in Bindrim treated it as a publication question. Publication is satisfied when a statement reaches one person besides the subject, so whether most readers made the connection was beside the point. Formulations vary by state, and the lesson does not: your book has to be understood by the wrong reader, not a wide one.

Authors often pair a rename with a second move, folding a person into a group. That has its own doctrine, and it is narrower than it sounds.

Restatement (Second) of Torts § 564A states the rule, and Florida’s court collected the numbers in Thomas v. Jacksonville Television, Inc., 699 So. 2d 800 (Fla. Dist. Ct. App. 1997): recovery has usually involved groups of twenty-five or fewer, and a class of 436 commercial fishermen was far too large. New York has gone further than the Restatement comment suggests. In Elias the Second Circuit allowed a small group claim by members of a fraternity chapter of roughly fifty-three, weighing size alongside whether the statement impugned all of them and how prominent the group and its members were locally.

The comment to § 564A carries the line memoirists should tape to the wall. Even where the group is large, circumstances known to readers can give the words such a personal application to one member that he is defamed as effectively as if he alone were named. In a town of four thousand people, a role can work as a name. So can a job title inside a company, a seat on a board, and the phrase “my first editor.”

Does renaming protect you from privacy claims?

Less than it protects you from a defamation claim, and for a different reason. Defamation asks whether you got the facts wrong. The privacy torts ask whether the disclosure was yours to make, so a passage about a renamed but recognizable subject can still draw a claim in the states that recognize public disclosure of private facts.

Identification is still an element. Where nothing in a publication points to the plaintiff, a privacy claim fails on that ground alone, as it did in Bayer v. Ralston Purina Co., 484 S.W.2d 473 (Mo. 1972), where the Missouri Supreme Court held that a plaintiff whose privacy remained inviolate had nothing to recover for. That case involved a photograph rather than a written portrait, and the principle carries: de-identification that actually works is a defense, and de-identification that only feels thorough is not.

The most useful decision I have read on renaming in memoir is Bonome v. Kaysen, 17 Mass. L. Rptr. 695 (Mass. Super. Ct. 2004), a trial court decision, persuasive rather than binding. Susanna Kaysen published a memoir about chronic pain that described her sexual relationship with a boyfriend she did not name. He sued for invasion of privacy. The court dismissed the claim: where an author’s own story addresses a matter of legitimate public concern, disclosures about a third party are protected where a sufficient nexus links the private detail to that concern, because intimate experience often cannot be separated from the person it was shared with.

Then the court did something authors should notice. It treated the absence of his name as reducing the degree of interference, pointing out that the people who could identify him were the friends, family, and clients who already knew about the relationship. The rename did not defeat identification. It moved a weight on the scale the court was balancing.

The honest version is narrow. Renaming rarely defeats identification, and it can still matter to how a court weighs the intrusion and to how your publisher prices the passage. Legitimate public concern gets weighed differently from state to state, and a memoir distributes into all of them at once, which is why a clearance read looks at more than one state’s law.

What does real de-identification take?

Changing every coordinate the subject’s own circle uses, not just the label. Role, place, time, relationship, and often the events themselves.

Here is the exercise I ran on renamed portraits. It takes about twenty minutes a person.

1. Name the five people most likely to read the book who also know your subject. Be specific. Their sister. Their old boss. The two people who were in the room.
2. For each one, write what they would recognize in the passage as written. Not what identifies your subject in the abstract. What identifies them to that reader.
3. Change or generalize items on that list until it is empty, and keep every change consistent across the whole manuscript. One unchanged detail in chapter fourteen undoes eleven changes in chapter three.
4. If the list will not empty out, stop renaming and start planning. That passage needs documentation, narrowing, a release, or a priced decision, and renaming is not going to be the answer.

The other move is the composite.

A hypothetical, and a composite rather than anyone’s actual manuscript. An author writes about the stepfather she calls Ray. In the draft he manages a Ford dealership in the mid-size city where she grew up, the confrontation happens the week of her college graduation in 2011, and her mother and two brothers are in the room. The rename does nothing for her brothers, and very little for anyone who knew the family. Move the business to another industry, move the city, move the year, and cut the family to a smaller cast, and the portrait stops being a fingerprint for readers outside the household. That is the kind of change that moves a flag toward clearance. It clears nothing by itself, and the passage still has to be accurate and supportable.

"Authors change the name and feel safe. The subject's sister reads the job, the divorce, and the summer of 2019, and the name never mattered. I flagged renamed portraits for years, and the rename was almost never the question."

Guy Muller, former general counsel of a US publishing company with global reach

When can a portrait not be de-identified?

When the identifying fact is the story. Some portrayals cannot be separated from the details that make them worth publishing, and no amount of renaming reaches them.

In my own review practice, three shapes came back as effectively impossible. The person defined by a unique position: the only woman on a nine-person board, the holder of a named office in a given year, the one survivor of a documented event. The immediate family member, because “my father” is not a detail you can change without changing the book. And the portrait whose force is its specificity, where the passage matters precisely because of the identifying facts, and generalizing them leaves a paragraph that says nothing.

Naming those cases early changes what you spend your time on. When a portrait cannot be de-identified, the manuscript needs a plan built on evidence rather than concealment: documentation supporting every factual claim, narrowing to what the evidence shows, attribution of claims to their sources, a release, or a disclosed and priced decision with your publisher. Which plan fits depends on who the person is, and the factors that sort people into those categories are the subject of Why you can write about some people and not others. When you ask for a release and the answer comes back no, the options and what each one costs the manuscript are in What to do when they refuse to sign a release.

---