The Principle of Charity

011 min

The problem the principle of charity solves

Words rarely fix their own meaning. A short sentence can be read in several ways, and the reader picks one. If the reader picks the weakest reading, they end up arguing with something the speaker never meant. The speaker then has to defend a point they did not make, or gives up a point they did make.

The cost falls on both sides, but it appears first on the receiving side, because the receiver acts on the reading they chose. A reviewer writes "This flow is too long." The designer can read that as a count of screens, as a comment about the effort a new user needs, or as a sign that the reviewer did not read the brief. Each reading leads to a different reply. Only one of them answers what the reviewer meant.

The principle of charity targets one specific failure: choosing a reading that makes the speaker look irrational, ignorant or dishonest when a more reasonable reading was available. It does not target rudeness or poor listening in general. It is a rule about which reading to choose, and it applies at the moment of choosing.

022 min

How the principle of charity works

The principle of charity is a method for assigning meaning. The interpreter treats the speaker as a rational person who mostly has sensible beliefs, and picks the reading that fits that assumption best. The philosopher Simon Blackburn describes the method this way: "it constrains the interpreter to maximize the truth or rationality in the subject's sayings."

In practice the method has four steps.

1. Find the plain reading. Say to yourself what the words mean if taken at face value.
2. Test it. Ask whether the speaker would be saying something false, self-contradictory, trivial or off-topic. If not, stop. The plain reading is the answer.
3. Look for other readings. Change one thing at a time: the way the speaker uses a word, the context they are in, or the knowledge they assume you share.
4. Choose the reading that makes the speaker most reasonable, then check it. Hold it as a hypothesis and revise it if the speaker's later words do not fit.

The word "charity" is misleading. The method does not ask you to be kind, and it does not ask you to agree. It asks you to be accurate about what was said. A reading that makes the speaker sound clever but that no evidence supports is not charitable in this sense. It is a different mistake.

The principle also works on a whole set of statements at once, not on one sentence alone. One strange statement does not cancel the method. The interpreter weighs it against everything else the speaker has said. The next section shows this with the case that gave the principle its name.

Philosophers have split the principle into several versions, because "most rational" can mean different things. Since the time of Willard Van Orman Quine, other philosophers have formulated at least four versions, and these may conflict with one another, so the right one depends on the goal of the conversation.

When two versions point to different readings, ask which one serves the purpose of the exchange. A code review needs the first and the second. A debate about a strategy needs the third.

031 min

Where the principle of charity comes from

Neil L. Wilson gave the principle its name in 1958–59. He proposed it as a practical rule for working out whom a speaker means when they use a proper name. The principle later became a general method in the philosophy of language through the work of Quine and Donald Davidson. This makes it a methodological proposal from philosophers, not an experimental finding. It has been argued over for decades, and its status is a matter of philosophical judgement.

041 min

A named case: Wilson's Caesar

Wilson's own example is short, and it shows the logic clearly. A speaker, whom Wilson calls Charles, makes five statements that contain the name "Caesar." Charles says that Caesar conquered Gaul, that Caesar crossed the Rubicon, that Caesar was murdered on the Ides of March, that Caesar was addicted to the ablative absolute, and that Caesar was married to Boadicea.

The question is who "Caesar" refers to when we cannot ask Charles. Wilson's rule was: "We select as designatum that individual which will make the largest possible number of Charles' statements true." A designatum is the person or thing a name points to.

Apply the rule. The first three statements fit Julius Caesar, the Roman general. The last one does not fit him, because Boadicea lived long after his death. Most of the statements still fit the general, so the interpreter chooses him and concludes that Charles holds one false belief about him.

Two points follow from the case. First, the method tolerates error. It does not require every statement to come out true, only as many as possible. Second, the method gives a verdict. The interpreter does not stay undecided between several candidates, and does not decide that Charles is simply confused.

051 min

Before and after

The scenario is illustrative. A product manager writes in a design review thread. The designer replies twice.

The comment from the product manager was: "This flow is too long." The first reply chooses a reading in which the reviewer made a claim that is easy to refute, and then challenges it. The second reply names two readings that the reviewer could reasonably have meant, and asks which one is correct.

What changed is structural. The first reply tests the weakest reading and puts the burden of defence on the reviewer. The second reply states the readings it considered, so the reviewer can correct the interpretation at once. The "After" version also shows that the designer took the comment seriously before disagreeing with it.

061 min

A script to use tomorrow

Use this when a message or an argument could be read more than one way, and you are about to answer the weaker one.

If the answer is yes, reply to that point. If the answer is no, you have found the gap before it cost you a wasted reply.

For a written thread, add one line before your disagreement: "The strongest form of your argument, as I see it, is [their position]. My concern is [your objection]." Keep your objection in a separate sentence, so the reader can tell where the restatement ends and the disagreement starts.

072 min

When the principle of charity goes wrong

The principle can fail in two opposite directions. Too little charity produces the straw man: a weak version of the other position that is easy to defeat. This is the failure the principle was written against, so the rest of this section covers the failure in the other direction.

The risk of too much charity is that the interpreter replaces what the speaker said with what the interpreter would have said. A related principle, named the principle of humanity by Daniel Dennett, makes this risk clear. It asks the interpreter to attribute "the propositional attitudes one supposes one would have oneself in those circumstances." A propositional attitude is a belief or a desire about something. The danger is that your own beliefs become the standard for what is reasonable, so the reading you choose is really your opinion.

There are four tells that a charitable reading has gone too far:

• The speaker says "that is not what I meant."
• Your reading is more precise or more sophisticated than anything the speaker has written or said.
• You have to ignore a clear statement in order to keep your reading.
• You give generous readings to people you agree with and plain readings to everyone else.

Charity also has a limit that comes from the evidence. A reading must be consistent with what the speaker has actually done and said. If a person repeatedly states a claim in plain words and the evidence shows that they mean it, the correct reading is the plain one. Charity does not require you to excuse a wrong statement, and it does not ask you to assume good intentions in the face of evidence of bad ones. It governs the choice of reading. It does not govern the verdict on the claim.

081 min

The principle of charity vs. nearby concepts

Cooperative principle. The cooperative principle, from the philosopher Paul Grice, describes what speakers normally do: they say what a conversation needs. The principle of charity is a policy for the listener. The two fit together. A listener can work out what a speaker meant because the listener assumes the speaker is cooperating. The deciding axis is who the rule describes: the speaker's habit in one case, the interpreter's method in the other.

Common ground. Common ground is the knowledge two people share and know they share. Charity needs it, because step three of the method varies the knowledge the speaker assumes. The deciding axis is that common ground is a body of knowledge, and charity is a way of choosing a reading.

Steelmanning. Steelmanning means building the strongest version of an opponent's position before responding to it. It is close to the argument version of charity and is often used as a synonym. Steelmanning can go further than charity: it may add support the opponent never gave. Charity in its original sense only selects among readings of what was said.

Hanlon's razor. Hanlon's razor is a proverb that advises you to explain a bad outcome by carelessness before you explain it by bad intent. It concerns why something happened. The principle of charity concerns what something means.

091 min

A second case: Rapoport's rules

A second case differs from the first in kind. Wilson's case concerns the meaning of a word. This one concerns an argument, and a person's habit of caricaturing it.

Daniel Dennett, in his book Intuition Pumps and Other Tools for Thinking, presents a set of rules from the social psychologist and game theorist Anatol Rapoport. Dennett calls them the best antidote for the tendency to caricature one's opponent. They set out how to write a critical commentary. Dennett gives the first rule as: "You should attempt to re-express your target’s position so clearly, vividly, and fairly that your target says, “Thanks, I wish I’d thought of putting it that way.”"

The remaining rules are to list points of agreement, especially those that are not widely shared, and to mention anything you learned from your target. Only after these steps may you offer rebuttal or criticism.

The variable that changes between the two cases is the test of success. In Wilson's case the test is how many statements come out true. In Rapoport's rules the test is the other person's own judgement: they must accept your restatement. The contrast teaches something that neither case teaches alone. A reading can be chosen by counting, as Wilson did, but when the other person is available, asking them is a stronger check, and the rules are built around that check.

102 min

Where the principle of charity is contested

Philosophers agree that interpreters need some assumption of rationality. They disagree on how strong it should be, and what it should be called.

Davidson held that charity is not optional. The Stanford Encyclopedia of Philosophy says that on the Davidsonian account the principle remains "a presupposition of any interpretation whatsoever." Each particular attribution of belief that charity produces can be revised, but the principle itself stays. The same entry says the principle "admits of various formulations and cannot be rendered in any completely precise form." That is the first source of disagreement. If the principle has no precise form, two interpreters can apply it and reach different readings.

A second objection comes from argumentation theory. Piotr Sikora and Jakub Pruś argue that the principle, in Davidson's contemporary version, meets with some serious problems. They still defend a modified principle, because critical discussion needs one. Their proposal is based on a rule written in the sixteenth century by Ignatius Loyola, and it gives a short procedure that participants follow so that they understand each other and are not misled by cognitive bias. Their paper is a recent academic proposal, and it is not yet widely used.

A third problem is that the versions conflict. The version that assumes true statements and the version that assumes valid arguments can give different readings of the same sentence. A practitioner who ignores this will apply whichever version gives the reading they already preferred. That links to confirmation bias: the tendency to favour evidence that supports an existing belief makes it easy to pick the version of charity that agrees with what you wanted to conclude.

What does this mean for someone at work? The disagreement is about how to state the principle, not about whether to read colleagues fairly. While the theory is unsettled, name the reading you chose, say which version of charity you used, and let the speaker correct you.

112 min

How the principle of charity changed since

The principle has changed its meaning several times, and the original rule is much narrower than the present-day use.

In Wilson's version, from 1958–59, charity picked the referent of a name. It was a counting rule: choose the individual who makes the most statements true. Quine then used it to make sense of any utterance whose meaning is unclear, which is a much wider domain than names. Davidson widened it again. In his essay "Radical Interpretation," as the Stanford Encyclopedia of Philosophy records, the principle counsels us to interpret speakers as holding true beliefs, true by our lights at least, wherever it is plausible to do so. Davidson gave it a larger role than the earlier philosophers had. For him, charity was both a constraint and an enabling principle in all interpretation, not a convenient habit.

A different line of development moved the principle out of the philosophy of language and into everyday argument. In that use, charity means answering the best version of the other person's argument. Rapoport's rules, as Dennett reports them, are one example. Another is the dialectical principle of charity published in 2023 by Sikora and Pruś in the journal Argumentation.

Current practice does not match the original meaning. Wilson wanted a way to identify who a name refers to. Most people who use the term today mean something closer to "argue against the best version of the other view." These are related ideas, and they share an attitude of respect for the other speaker's rationality. They are not the same method, and the second one has no counting rule to settle disputes. When someone says "be charitable," it helps to ask which one they mean.

122 min

Frequently asked questions

?7 questions

Questions people ask

§4 sources

Sources