What You Can and Can't Say: A Legal Guide for Writers and Publishers

You spent years on this book. Somewhere in it is the thing you were afraid to write, and you wrote it anyway, because the book doesn't work without it. Now an editor, an agent, or a lawyer has told you that part of it is a problem. Nobody has explained which passages carry real danger, which flags are a nervous reader covering themselves, or what your options are besides cutting.

I spent years on the other side of that conversation. I was general counsel of a US book publishing company with global reach. I wrote the standards our editors flagged manuscripts against, and then I read the flags, week after week, and decided what went out into the world. I've seen sharp books survive review with everything that mattered intact, and I've seen books get sanded down to nothing for no legal reason at all. The difference was almost never the manuscript. The difference was whether anyone in the room understood how the decision actually gets made.

So that's what this guide is. Publishing risk, explained from the inside: the legal claims a book can draw, the business costs a publisher weighs, and the specific moves that let you keep the true thing you came to say. It's written for authors first, and for the editors, agents, and publishers who have to make these calls with them.

One note before we start. Defamation, privacy, and publicity law vary by state, and this page is general information, so treat it as a map rather than advice about your manuscript. Where the law splits, I'll say so. None of this is legal advice, and reading it is not a substitute for consulting your own attorney about your specific and unique situation.

How do publishers decide whether a book is safe to publish?

Publishers decide by pricing the risk. The price has two parts: how likely a claim is, and what the claim costs if it comes. Whether the publisher would eventually win in court is a minor input, because most of the cost arrives long before any verdict.

Authors walk into this conversation with the wrong model. The author is thinking about rights, truth, and courage. The acquiring editor is sitting in a margin business where one bad outcome can erase the profit on twenty good books. When a manuscript gets flagged, the questions moving through the publishing house look like this:

• What is the probability anyone complains at all?
• If someone complains, what does the response cost before a suit is even filed? Lawyers answer demand letters at hourly rates, and those bills arrive whether or not the complaint has merit.
• Can distribution be delayed, and what does a slipped publication date cost across retail commitments, review coverage, and the author's own publicity?
• If books are already printed, what do a recall, pulping, and reprint cost?
• What do the retailer and distributor relationships look like afterward?
• What does the exposure do to the imprint's reputation with agents and future authors?
• Does the author have assets behind the indemnity clause, or is that clause decorative?
• Is there insurance, what is the retention, and does a claim affect renewal?

Read that list again, because it explains almost every frustrating thing a publisher will ever say to you. A passage can be legally defensible and still commercially unpublishable, and a passage can carry real legal risk that a publisher happily accepts because the probability of a complaint is near zero and the book is worth it.

Once you understand that the decision is priced, you can argue price. An author who says "the First Amendment protects me" changes nothing, and I say that as someone who now litigates tort cases for plaintiffs. An author who says "this subject has no history of litigation, the passage is documented in these three sources, and I'll take a revision that keeps the substance" changes the calculation. The full framework, including how manuscript fixes and contract terms work together, is in How publishers decide whether your book is safe to publish, coming to this library.

The same logic answers the bravest sentence in publishing: "I don't care if I get sued." Your publisher does, because your publisher gets sued with you, on a thinner margin than yours. I wrote out those economics, from response costs to pulped inventory, in "I don't care if I get sued." So why does your publisher?

What are the five ways a book creates legal exposure?

Five doctrines produce nearly every legal problem a manuscript can have. Everything a lawyer flags in a book traces back to one of these, and each one has its own logic, its own defenses, and its own fix.

The elements above are the common formulation. Each state's defamation law sets its own precise wording and fault standard.

The next five sections walk through each category at the depth a working writer needs. The deep-dive pages linked throughout go further.

What is defamation, and where do book claims actually come from?

Every state defines defamation somewhat differently, but the common formulation is a false statement of fact about an identifiable living person, communicated to a third party, that damages the person's reputation, with the specific elements and fault standard set by the law of the state where the claim is brought. Every word of that sentence is doing work, and each one is a place where claims live or die.

False. A defamation claim needs falsity. Accurate statements defeat it, which is why documentation is the spine of every risky book. Keep your sources, your records, and your interview notes, because the practical question is rarely "is it true" and usually "can you show it."

Statement of fact. Opinions get protection, but the label does you less good than writers hope. The Supreme Court held in Milkovich v. Lorain Journal Co., 497 U.S. 1 (1990), that there is no blanket exemption for anything called opinion: a statement is actionable when it implies a provably false fact. "In my opinion, my business partner cooked the books" implies a provable fact. Writing "in my opinion" in front of an accusation changes almost nothing.

Identifiable. The person has to be recognizable, and recognizable means recognizable to people who know them. A changed name with a real job, a real city, a real timeline, and a real sequence of events does little to protect a subject, because identifiability turns on whether people who know the subject can recognize them from the surrounding details, not on the name itself. Identifiability is the real test, and it's the single most misunderstood concept in publishing risk. I wrote a full page on it: Does changing the name actually protect you?

Fault. How careless you were matters, and the standard depends on who you wrote about. These two rules come from the Supreme Court and apply nationwide:

The practical consequence runs in the author's favor more often than authors expect: sharp, documented criticism of public people on public conduct is among the best-protected speech in American law. The unglamorous claims are the dangerous ones. The brother-in-law in chapter four, private figure, lower fault standard, motivated for reasons that have nothing to do with your book. That asymmetry, and the other factors that decide who is safe to write about, get a full treatment in Why you can write about some people and not others.

One more thing writers ask constantly: calling it fiction does no work by itself. If readers who know your subject can recognize them in your character, a defamation claim survives the novel label. The fiction pages in this library cover it, and the short version is that the identifiability analysis above applies with the names changed.

How can a true story still be legally risky?

A true story can still be legally risky because privacy law punishes disclosure rather than falsehood. Defamation asks whether you lied. Privacy asks whether some truths were yours to publish, and it's the category that surprises memoirists most.

American courts commonly sort privacy into four separate torts, following the framework in the Restatement (Second) of Torts: public disclosure of private facts, intrusion upon seclusion, false light, and appropriation of name or likeness. Adoption varies state by state, and the variance is real rather than academic.

Public disclosure of private facts is the memoirist's tort. The claim runs against publicizing intimate, non-newsworthy facts about a private person: medical history, sexual life, financial details. The defense that matters is legitimate public concern, and courts give real weight to it, but its edges vary by state and the analysis is fact-specific.

Intrusion upon seclusion targets how you gathered material rather than what you published. Recording someone in their home, accessing accounts or files you had no right to open, going through papers that were locked away. An exposé built on well-sourced documents someone gave you sits differently from one built on documents you took, and the intrusion analysis can apply even when everything published is accurate.

False light, where it exists, covers technically-true-but-misleading portrayals that a defamation claim can't reach. Some states refuse to recognize it at all. Texas is one of them: the Texas Supreme Court rejected the tort in Cain v. Hearst Corp., 878 S.W.2d 577 (Tex. 1994), reasoning that it duplicates defamation while evading defamation's free-speech safeguards. Other states apply it regularly. A portrayal that is safe in Dallas can support a claim in another state, and books distribute nationally, which is exactly why a clearance read checks more than one state's law.

Appropriation overlaps with the right of publicity, covered next.

The pattern worth keeping: truth is the complete answer to defamation, and truth is barely a defense at all in privacy. A clearance review reads every passage twice for that reason, once asking "can this be proven" and once asking "should this be published about this person at all."

When does the right of publicity touch a book?

The right of publicity protects the commercial value of a person's name, image, and identity, and it lives in a patchwork of state statutes and common law. Some states protect identity after death, some don't, and the terms vary enough that any concrete question needs a specific state's law in front of it.

For working writers the shape of it is simpler than the patchwork suggests. Books are expressive works, and courts have generally given expressive works breathing room that they refuse to give commercial products. Writing about a celebrity in a biography sits near the protected core. Putting that celebrity's face on the cover of a book that has nothing to do with them, using their name in advertising, or spinning off merchandise moves toward the danger zone. The publicity questions I saw as a publisher's GC came from covers, marketing, and title pages far more often than from the text itself.

What do you need permission to quote or reproduce?

Permission questions come down to three things: what you're copying, how much of it, and what your use does to the market for the original. Copyright is federal law, uniform across the country, which makes this the one category where the rules don't shift under you at the state line.

Fair use is the doctrine everyone has heard of and almost everyone overstates. The statute, 17 U.S.C. § 107, sets out four factors courts weigh: the purpose and character of your use, the nature of the work you're quoting, the amount and substantiality of what you took, and the effect on the market for the original. Fair use is a case-by-case defense, decided on specific facts, and no honest lawyer will certify a use as fair in advance. Anyone who gives you a safe word count is making it up. The Copyright Office says the same thing in its Fair Use Index: there is no formula.

A few practical patterns from years of watching permissions fights:

• Song lyrics and poetry are the expensive category. The works are short, so a line or two is a meaningful fraction of the whole, and an active licensing market means your unpaid use has an obvious price tag. Publishers routinely require licenses for quoted lyrics, and the license can cost more than the epigraph is worth to you.
• Prose quotation has more room, at least in the sense that courts weighing the purpose-and-character factor give real weight to whether a quotation engages with, criticizes, or builds on the original, as opposed to using it as decoration. That factor is one of four, and no single factor decides the outcome.
• Letters, emails, and texts belong to their writer. As a general rule, whoever wrote the words owns the copyright in them. Receiving a letter gives you the paper, and publishing its text is a separate question. Memoirists quoting a parent's letters are quoting someone else's copyrighted work, and the same logic reaches emails and text threads.
• Photographs belong to the photographer. Owning a print, appearing in the picture, or paying for the shoot doesn't by itself give you the right to reproduce it, absent an agreement that transferred rights.
• Naming real brands is usually a different problem than writers think. Mentioning a trademarked product in prose is not, by itself, trademark infringement, since trademark law targets confusion about source or sponsorship rather than the mere use of a name in a narrative. The trouble starts when the story's treatment of the brand reads as a false factual claim about it, which circles back to defamation and its cousins rather than trademark.

Deep dives on lyrics, quotation, letters, photographs, and ownership questions like ghostwriting and AI are all part of this library, each built around the specific permission fight it covers.

What did your contract already promise your publisher?

Your publishing contract almost certainly contains a warranty section and an indemnity clause, and together they move publishing risk from the publisher to you. Almost nobody explains those clauses to authors, and I've come to believe the silence is structural: the person who drafted them sits on the other side of the table.

The standard architecture works like this. You warrant that the book is original, that it doesn't infringe anyone's rights, that it isn't defamatory, and that its statements of fact are accurate. Then the indemnity clause says that if anyone brings a claim tied to a breach of those warranties, you cover the publisher's losses, and depending on the drafting, its defense costs. On a book that earns you a fifteen thousand dollar advance, you can be standing behind a company many times your size.

Three things follow, and they're worth more than any other contract advice you'll get this year:

• Read the warranty clause before you deliver the manuscript, and behave as if every warranty is a checklist. The warranties are promises about the words on the page, so the clearance work has to happen before delivery, when revision is still cheap.
• Scope matters and is negotiable. Whether the indemnity covers claims or only breaches, whether it reaches settlements made without your consent, whether there's a cap: these are drafting choices, and agents concede them more often than they should.
• Insurance changes the picture. Publishers carry media liability coverage, and authors can sometimes be added to it. Whether that protection actually reaches you depends on the policy and the endorsement, and it's a question to raise before signing rather than after a demand letter arrives.

Contracts also restrict speech from the other direction. An NDA, a separation agreement, or a fiduciary obligation can make a passage legally dangerous even though every word is true and no tort applies. Business memoirs live and die on this, and the business and confidentiality pages in this library treat it at length, including what your old employment paperwork actually covers and what it can't reach. Newer contracts add one more layer: AI clauses, with warranties about how the manuscript was made. Those clauses are evolving quickly enough that they get their own pages here, kept on a quarterly review cycle.

What does Texas law add to the picture?

For a book with Texas connections, four Texas rules change the practical math: a one-year clock, a correction-request statute with teeth, an early-dismissal law that shifts fees, and a publicity right that runs fifty years past death. I practice in Texas and this site is grounded here, and the state also makes a useful demonstration: when lawyers say defamation law varies by state, this is the kind of variance they mean.

The clock is short. Texas gives a defamation plaintiff one year to sue for libel or slander, measured from when the claim accrues. Tex. Civ. Prac. & Rem. Code § 16.002(a). When accrual happens has its own case law, so treat the precise start date as a lawyer's question, and treat the practical point as settled: the Texas window is short.

The retraction letter has statutory teeth. Under the Texas Defamation Mitigation Act, a plaintiff may maintain a defamation action only after making a timely and sufficient request for a correction, clarification, or retraction, or where the publisher has already made one. Tex. Civ. Prac. & Rem. Code § 73.055(a). A plaintiff who lets ninety days pass after learning of the publication without requesting one gives up exemplary damages. § 73.055(c). And a correction made in compliance with the statute takes exemplary damages off the table unless the publication was made with actual malice. § 73.059. In Texas, the letter demanding a retraction is a statutory event with deadlines and consequences running in both directions, and the response deserves counsel rather than instinct.

Weak claims can die early, at the claimant's expense. The Texas Citizens Participation Act, the state's anti-SLAPP statute, lets a defendant move for early dismissal of a legal action based on or responding to the exercise of the right of free speech, the right to petition, or the right of association. Tex. Civ. Prac. & Rem. Code § 27.003. If the court dismisses, it must award the moving party court costs and reasonable attorney's fees, and it may add sanctions calibrated to deter similar suits. § 27.009(a). The statute was amended effective September 1, 2019, so older commentary about its reach needs checking against the current text. For authors and publishers, the TCPA moves the probability side of the risk math on Texas-filed claims.

The dead have publicity rights here, for a while. Texas protects a deceased individual's name, voice, signature, photograph, and likeness, and the protection has an end date: the statute permits any use after the fiftieth anniversary of the individual's death. Tex. Prop. Code § 26.012(d).

And the Texas tort map from earlier on this page, gathered in one place: private figures suing media defendants must prove negligence, a standard the Texas Supreme Court has not extended to non-media defendants, Neely v. Wilson, 418 S.W.3d 52 (Tex. 2013); the private-facts tort runs on publicity given to private matters, offensiveness to a reasonable person, and the absence of legitimate public concern, Star-Telegram, Inc. v. Doe, 915 S.W.2d 471 (Tex. 1996); and false light claims fail outright under Cain v. Hearst Corp., 878 S.W.2d 577 (Tex. 1994).

Texas is one state. Every state has a stack like this, its own clock, its own procedural gates, its own tort map, and a book distributes into all of them at once. A clearance read checks the states that matter to your subject, and your home state is only the first of them.

Why is memoir the category publishers worry about most?

Memoir concentrates every category on one page: real people, identified or identifiable, in private moments, described by someone they know, sold as true. A novelist can argue invention and a journalist can point to institutional standards and editors. A memoirist is a single witness selling recollection, and publishers price that accordingly.

The publishing houses aren't wrong about the category, but the right response is engineering rather than retreat. A memoir becomes acquirable when the author does four things: gets a clearance review before delivery, obtains an attorney review letter the publisher can rely on, prices media perils insurance, and, where the deal allows, gets added as an additional insured on the publisher's policy. Each of those moves is explained in Why memoir is the riskiest category in publishing, and how do you de-risk yours?

Memoirists also collide with the two most maddening features of publishing risk. The first is that publishers seem inconsistent about who needs a release and whose name gets changed, and the apparent inconsistency is actually a set of factors nobody wrote down for you: public figure status, newsworthiness, provability, whether the person is dead, whether they're realistically litigious, and whether the portrayal is one no release would ever cover. I unpack that logic in Why you can write about some people and not others. The second is the release that doesn't come. When the person says no, you still have options, and the practical playbook, with what changes in the manuscript when a release is off the table, is in What to do when they refuse to sign a release.

What does legal review actually do to your book?

A competent legal read inventories the manuscript's factual claims about real people and entities, flags the ones carrying exposure, and recommends a fix for each flag. A weak legal read does something simpler: it deletes whatever made the reader nervous. Authors need to know the difference, because the two look identical in a margin comment.

From the inside, I watched the same two failures repeat for years, and they're the reason this guide exists. Authors give up material they were entitled to keep, because nobody told them the flag was precautionary. And authors burn their credibility defending material that actually endangers them, because nobody separated the two categories of flag. Both failures come from the same gap: the publisher has counsel, a risk framework, and an indemnity clause pointing at the author, and the author has a manuscript and no vocabulary for the conversation.

The vocabulary is buildable. When a flag comes back, ask three questions. Which claim, exactly, and about whom? What's the theory: defamation, privacy, contract, permission? And what revision would clear it? A flag that can't survive those three questions was reflex. A flag that can survive them deserves your full attention, and usually has a fix short of amputation: attribute the claim to its source, narrow it to what your documentation supports, generalize the identifying detail that does no narrative work, or move the assertion from fact to clearly-signaled perception. The sharp thing in your book is almost always keepable. What has to change is the unsupported thing wearing the sharp thing's clothes.

"A good legal read protects the book without removing the reason it matters. I've watched both versions happen from the general counsel's chair, and the difference isn't caution. The difference is whether the reviewer knows the law well enough to fix a sentence instead of cutting it."

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

Who wrote this?

I'm Guy Muller. Before I started my own firm, I was Associate General Counsel of a technology company and then General Counsel of a US book publishing company with global reach and roughly thirty years of operating history. I handled the publisher's contracts, intellectual property, licensing, work-for-hire agreements, board governance, and M&A, and I wrote the manuscript review standards our editorial team worked from. Editors flagged passages against those standards, and the flags came to me. I read them every week, at volume, and made the recommendations that decided what we published.

Today I'm a plaintiff trial lawyer in Texas, and the cases I litigate are personal injury cases. A personal injury is a tort: a civil wrong, one person's conduct injuring another, that the law remedies with money damages. Defamation, invasion of privacy, false light, and misappropriation are torts from that same family. Those are the injuries I work to prevent on the front end rather than litigate after the fact. The combination shapes everything on this page: I priced manuscript risk from inside a publishing house, I try tort cases, and I spend my working life thinking about how claims actually get brought.

Texas is strict about specialization claims, so I'll put it plainly instead: this is work I did, weekly, for years, from the chair where the decision gets made. More on the background is at About.

Where should you go deeper?

6 deep-dive pages are live now, built in the same spirit as this one: what's actually dangerous, what isn't, and what to do about it.

• Part 2Does changing the name actually protect you? Identifiability as the real test, and what genuine de-identification requires.
• Part 3"I don't care if I get sued." So why does your publisher? The economics that drive every publishing decision about risk.
• Part 4Why is memoir the riskiest category in publishing, and how do you de-risk yours? The four moves that make a memoir acquirable.
• Part 5Why can you write about some people and not others? The factors behind the apparent inconsistency.
• Part 6What to do when they refuse to sign a release Options and tradeoffs when the answer is no.
• Part 7Can you write about your ex, your parents, or your siblings? You can write about family without their permission. Relatives are private figures, they recognize themselves through renaming, and proof matters most.

Common questions