Maintenance
Part of Misquotes and fact checks: the questions that matter
Best Misquotes and Fact Checks: A Citation-Loop Guide
Best misquotes and fact checks start by tracing a quote's sources, spotting circular reporting, and testing whether evidence reaches an independent source.
For the best misquotes and fact checks, start by asking whether the sources are independent. Suppose four sources agree. That is the moment most checks stop, and it is the moment they should slow down. Agreement between sources is only evidence when the sources are independent, and in quotation work they usually are not. What looks like four confirmations is often one claim, copied three times, wearing four different mastheads.
This page is about that failure specifically: how a single unsourced sentence acquires the appearance of corroboration, how to see the loop from inside it, and what to do when the loop is all there is.
Selected misquotes and fact-checking resources
These familiar cases are useful tests of exact wording. Compare the popular version with the source, rather than counting pages that repeat it.
- “Play it again, Sam.”In Casablanca, the spoken wording is “Play it, Sam. Play ‘As Time Goes By’”; “Play it again, Sam” is a familiar paraphrase, not the line.
- “Luke, I am your father.”In The Empire Strikes Back, Darth Vader’s line is “No, I am your father.”
- “Elementary, my dear Watson.”This familiar Sherlock Holmes catchphrase does not appear in Arthur Conan Doyle’s Holmes stories as that exact line.
For wider fact-checking, search Snopes, PolitiFact, FactCheck.org, and The Washington Post Fact Checker. These are established fact-checking outlets, but any quotation verdict still needs to be traced to the work or record it cites; an outlet name alone does not establish provenance.
What to take away
- Counting sources measures circulation, not evidence.
- Every source you accept has to be reduced to what it actually cites.
- A loop has a shape you can draw in about ten minutes on paper.
- When the loop closes with nothing inside it, that is a finding, and it is publishable.
Why four sources can be one source
Independence is a strong claim. Two sources are independent when each could have arrived at the statement without the other. Most pages carrying an attributed line fail that test badly. They were written after the line was already common, by writers who took it from the common stock, and they add nothing except another date and another byline.
The failure has a name outside this subject. Circular reporting describes a claim that returns to its origin through intermediaries and is then treated as confirmed by them.
Quotations are unusually vulnerable to it because the unit copied is short, memorable, and carries its own supposed provenance inside itself. The name is part of the sentence, so copying the sentence copies the attribution automatically.
How a loop forms
A citation loop can form through a sequence of ordinary-looking repetitions:
- A line begins circulating without a source that can be checked.
- A writer attaches an attribution or repeats one without verifying it against the original.
- Later writers copy the wording and attribution from that account.
- A later page cites one of those repetitions, making the line appear to have a traceable source.
How a Citation Loop Forms
- Line appears with no source
- Second writer adds a name
- Third writer cites the second
- Reference work picks it up
- Everyone cites the reference work
By stage five the line has a citation history that is entirely internal. Nothing in it ever touched a text the credited person produced. The general problem of provenance applies here exactly as it does to an object in a museum: what matters is the unbroken chain back to origin, not the number of people who have handled it since.
Drawing the graph
Do this on paper. It takes less time than arguing about it.
Write the line at the top. Under it, list every source you have found, each with a date. Then, for each one, answer a single question: what does this source itself point to? Draw an arrow to that. Three outcomes are possible.
| Arrow lands on | What it means | What to do next |
|---|---|---|
| Another source in your list | This node adds nothing | Cross it out and follow the arrow |
| A source not yet in your list | New evidence | Add it, then ask the same question of it |
| Nothing at all | Terminal node | Mark it, and note its date |
Cross out every node whose arrow lands inside the list. What remains is your actual evidence. Very often it is one node, and that node has an arrow pointing at nothing.
Signs you are inside a loop
- Every source postdates the same year, and none of them explains why that year.
- The wording is identical across sources, including an odd word that no independent writer would have chosen twice.
- The same small error travels with the linea wrong middle initial, a wrong decade, a garbled place name. Shared errors are the strongest evidence of copying there is, because a correct fact can be arrived at twice and a mistake usually cannot.
- Sources cite each other in a ring, so following the trail returns you to where you started.
- Nobody quotes more of the passage. If a real page existed, someone would have quoted the sentence before it.
That last one deserves weight. Genuine sourced material tends to accumulate context over time, because writers who open the book see what surrounds the line. A line with no context accumulating around it in decades of use was probably never in a book anyone opened.
Signs You Are Inside a Loop
- All sources postdate the same unexplained year
- Identical wording, including one odd word
- The same small error travels with the line
- Sources cite each other in a ring
- Nobody quotes more of the passage
Breaking out
There is only one exit, and it runs backward in time. Loops are made of secondary material, so the way out is a corpus predating the loop's earliest node and searchable by date.
For newspapers, search the Library of Congress's Chronicling America: Historic American Newspapers collection by date and wording. For a quotation from congressional proceedings, check the Congressional Record. Quote Investigator can point you toward prior quote research, but follow its citations to the earliest source rather than treating a later page as proof.
Search for the terminal node's date minus twenty years, not for the line's fame. You are looking for the sentence before anyone thought it was worth attributing. If a loop points to a book, use WorldCat to locate an edition, then check the cited edition and page. An earlier appearance, if it exists, is usually anonymous, longer, and worse written than the version you started with. That is what a real ancestor looks like.
As a quotation is shortened or copied, its wording and attribution can change. Compare each version with the earliest source you can locate, and note any wording or attribution that appears only in later repetitions.
If the search returns nothing earlier, say so precisely. A loop with an empty center, described honestly, is a stronger piece of work than a fifth repetition would have been. State the terminal node, its date, the corpora you searched, and the fact that the chain closes on itself.
When the loop is the whole record
Some lines will never resolve; the honest position is that attribution rests on a chain of secondary sources. The earliest member gives no evidence, and the corpora you searched contain nothing earlier.
That is a real finding and a reason not to use the line, especially if the credited person is alive and the sentence makes them say something.
When a saying has no single author, record its earliest traceable appearance and avoid assigning an author without evidence.
Common questions
How many sources is enough?
The question has no number in it, which is the point. One source that reproduces the passage with an edition and a page beats forty that repeat a sentence. Reduce every source to what it cites, then count only what is left.
Does a shared error really prove copying?
It is the strongest single indicator, though not proof on its own. Two writers can independently make the same error if the error is an obvious one, such as confusing two people with the same surname. An arbitrary error, such as a misspelled place name, is much harder to arrive at twice.
If a well-regarded publication printed it, is that not evidence?
It is evidence about the publication's process, not about the line. Reputable outlets carry unsourced quotations constantly, usually in commentary and feature writing where nobody expected a citation. Ask what that specific piece cited, not what the masthead is worth in general.
Can a loop be broken by finding the true author?
Sometimes, and it is worth trying, because a line lifted from a lesser known writer often survives in that writer's own book. More often the search ends with an earlier anonymous appearance and no author at all. Report the earliest appearance and leave the authorship question open rather than filling it.
