Express Confidence and Uncertainty in Calibrated Language
In the last chapter you saw why honest qualification is worth the effort. Now comes the practical part: the actual words. Confidence is not a simple on-off switch between “sure” and “no idea.” It comes in degrees, and the trick is to reach for the phrase that matches how much the evidence really supports your view. This chapter gives you a ladder of those phrases and shows you how to pick the right rung.
Picture Confidence as a Ladder, Not a Switch
Most of us talk about being sure as if there were only two settings: certain, or clueless. Real evidence almost never works that way. Sometimes you have a hint, sometimes a strong case, sometimes a pile of proof, and sometimes genuinely nothing to go on. Calibrated language is simply the set of phrases that let you say which of those you have.
The key idea is that these phrases describe the strength of your grounds, meaning how much the evidence actually backs the claim. They are not a hidden percentage you have to calculate. Think of the ladder below as a beginner-friendly scaffold, not a formal scale you must score exactly.
Climb the Ladder One Rung at a Time
Here is a simple ladder of calibrated phrases, from the least evidence to the most. Read from the first rung to the last, and notice how each one claims a little more than the rung before it.
“We do not yet know whether…” Use this when you genuinely have nothing solid either way. It is not a failure to say it; it is the honest floor of the ladder.
“The evidence weakly suggests…” Use this when you have a little evidence pointing one way, but it could easily be outweighed by something you have not seen yet.
“The evidence strongly suggests…” Use this when you have solid, consistent evidence, while still leaving room for a reasonable doubt.
“On balance…” Use this when there is evidence on both sides, you have weighed it, and one side comes out ahead. Think of this one less as a higher rung than as a phrase for contested cases: it signals that you saw the competing points and still landed somewhere, which may be firm or only slight.
“It is very likely that…” Use this only when the evidence is strong, consistent, and points clearly in one direction, with little room left for doubt.
Notice that even the top rung stops short of “it is certain.” Outside of things you can check directly, “very likely” is usually as high as honest calibration should climb.
Match Your Words to the Evidence, Not to Your Mood
The whole skill lives in one move: pick the rung the evidence warrants, not one higher and not one lower. That means resisting two pulls at once.
The first pull is to climb too high because a bold claim feels more persuasive. It rarely is, and it is the single most common calibration mistake.
The second pull is to slide too low out of caution, dropping a well-supported claim down to “who can say” so no one can ever pin a mistake on you. That is the false modesty you met in the last chapter, and it wastes good evidence. Matching your words to the evidence means going neither higher for effect nor lower for cover.
Rewrite a Claim to the Right Rung
The fastest way to feel this is to watch a claim move up and down the ladder as the evidence changes. Here are a few everyday examples, each rewritten to the rung the grounds actually support.
A sports prediction. Overclaimed: “Our team will definitely win on Saturday.” What the grounds support, if the team has won its last five games against weaker opponents but a key player is injured: “The evidence strongly suggests our team will win on Saturday, though the injury is a real question mark.”
A product review. Overclaimed: “This is the best budget headphone on the market, guaranteed.” What the grounds support, after reading a dozen consistent reviews but testing none yourself: “On balance the reviews point to this being one of the better budget options, though I have not tried it myself.”
An election poll. Underclaimed: “Polls are always wrong, so who really knows.” What the grounds support, with several recent polls agreeing within a small margin: “It is very likely the race is close, though a late shift is still possible.”
A health question. Overclaimed: “One doctor said it is nothing, so it is definitely fine.” What the grounds support, from a single opinion with no follow-up: “The evidence weakly suggests it is minor, but one opinion is thin, so this is not settled.”
In each case, nothing changed about the topic. Only the phrase changed, so that the words now carry the same amount of confidence the evidence does.
See Why Both False Certainty and Vague Hedging Mislead
It helps to see that the two errors are mirror images, and that both break the same thing: they make your words stop matching your evidence.
False certainty borrows confidence you have not earned. It reads well until the claim is tested, and then it costs you, because a listener who trusted your “definitely” now doubts your next one. Vague hedging does the opposite. By refusing to commit, it hands the listener no information at all; a shrug cannot be acted on, however carefully you actually thought. One overstates, the other understates, and both leave the listener with a false picture of how much you really know.
Calibrated language is the cure for both. When your phrase honestly names the rung, “very likely” and “we do not yet know” become useful signals rather than noise, and people can finally take your confidence at face value.
Let’s Recap!
Confidence comes in degrees, so calibrated language uses a ladder of phrases rather than a single “sure or not sure” switch.
These phrases describe how strong your grounds are, not a hidden percentage you have to calculate.
The ladder runs from “we do not yet know whether” up through “weakly suggests,” “strongly suggests,” and “on balance” to “it is very likely that.”
The core skill is picking the rung the evidence warrants, resisting both the pull to climb higher for effect and the pull to drop lower for cover.
False certainty and vague hedging are mirror errors, and both mislead because they stop your words from matching your evidence.
You can now place a claim on the right rung of the ladder. In the next chapter you will sharpen that skill by looking hard at the two ways it slips, so you can catch false certainty and false modesty in your own writing.
