
In the last chapter you gathered evidence on both sides of a decision. Now comes the final move: reasoning from that evidence to a conclusion. This is the moment where even a careful thinker can still slip, not by holding weak evidence, but by claiming more than the evidence can carry. This chapter shows you how to reach a conclusion you can actually stand behind.
Reasoning is the step from the last part where your grounds are supposed to support your recommendation. Most of the time the evidence itself is fine; what fails is the size of the conclusion drawn from it.
Overreach is drawing a conclusion that is broader, bolder, or more certain than your evidence can support. The evidence points in a direction, and you run further down that road than it actually takes you. Suppose you move a struggling fern closer to the window and, over two weeks, it perks up. The evidence supports something modest: “this fern seemed to do better with more light.” Overreach turns that into “ferns need direct sunlight” or “light was definitely what fixed it,” claims about all ferns, and about cause, that two weeks with one plant cannot carry.
Notice that overreach is not the same as being confident, and this is the distinction that trips most people up. You can state a small, well-scoped claim with total confidence and be completely in the clear. Overreach is not about how sure you sound; it is about how big your claim is compared with the evidence under it. A quiet voice can overreach, and a firm one can stay perfectly honest.
Behind most overreach sits a simple confusion: mixing up what the evidence shows with what you wish it showed. You wanted the fern trick to be a reliable rule, so your mind quietly upgrades a hopeful hint into a settled fact. Catching that upgrade, and refusing it, is the whole skill of this chapter.
The rule that prevents overreach is short: match the strength of your conclusion to the strength of your grounds. Weak or narrow grounds earn only a modest, narrow conclusion. Strong, broad grounds earn a firmer one. The evidence sets the ceiling, and your conclusion must stay under it.
It helps to see that a conclusion has two dials you can turn too far. One is scope: how widely the claim reaches, from “this one fern” all the way out to “all ferns everywhere.” The other is certainty: how sure the claim is, from “seemed to help” up to “definitely caused.” Overreach is pushing either dial past what the evidence supports, and you can overreach on one while staying honest on the other. “This diet definitely worked for me” may be fine on scope but overreach on certainty; “this diet probably works for everyone” overreaches on scope while hedging the certainty.
Fixing overreach is usually not a matter of throwing the claim away. It is a matter of tightening it until it sits exactly on top of its evidence. Here are three claims that reach too far, each paired with a tightened version that says only what the evidence earns.
Reaches too far: “My two flatmates tried the new diet and lost weight, so this diet works.” Tightened: “Two people I know lost weight on this diet, which makes it worth a closer look, not proof that it works in general.”
Reaches too far: “This blender has a five-star review, so it is the best on the market.” Tightened: “One reviewer rated this blender highly, which is a point in its favor, but a single review cannot rank it against everything else.”
Reaches too far: “I revised with flashcards and passed, so flashcards are the best way to study.” Tightened: “Flashcards seemed to help me this time; whether they beat other methods, or work for other people, I cannot tell from one exam.”
Look at what the tightening does in each case. It shrinks the scope, from “everyone” or “the market” down to the handful of cases you actually saw. It softens the certainty, from “works” and “best” down to “worth a look” and “seemed to help.” And it keeps the honest core intact, so you still have a real claim, just one you can defend. A tightened claim is not a weaker claim; it is a claim that will not collapse the moment someone pushes on it.
The everyday tool for tightening is qualifying language: the small words that size a claim to its evidence. “Some,” “often,” “so far,” “in this case,” “tends to,” “seems,” “based on what I have seen,” these are not filler. Each one marks the edge of what your evidence covers, so the claim stops exactly where the support stops. “Flashcards work” and “flashcards seemed to help me so far” rest on the same evidence, but only the second one tells the truth about how much.
This is only a first taste of qualifying a claim well, which is a craft in its own right. For now, the point is simply to notice that these words are honesty tools, not signs of weakness.
If I hedge every claim like this, won’t I just sound wishy-washy and unconvincing?
It feels that way, but it usually runs the other way around. A claim scoped to its evidence is harder to argue with, not easier, because there is no exposed overreach for anyone to attack. “This diet cures everyone” invites a single counterexample to demolish it; “two people I know lost weight, so it is worth a look” cannot be knocked down that way, because it never claimed more than it can show. The overreaching claim sounds stronger and breaks on contact; the scoped one sounds humbler and holds.
Let’s put the whole chapter to work on one small case, spotting each overreach in turn and then tightening the claim to fit.
Imagine you spend four nights with your phone left in another room at bedtime, and you notice you fall asleep faster on those nights. Pleased, you announce: “Phones cause insomnia, so everyone should keep them out of the bedroom.”
Now weigh what the evidence actually supports, checking both how much it covers and how far the claim stretches on scope and certainty. The overreaches stack up quickly:
Coverage: the evidence is four nights, for one person, you. That is a thin slice, nowhere near enough to speak for “everyone”;
Other causes: plenty else could explain the better sleep, less caffeine that week, a quieter mind, simple chance, so the phone may not be doing the work you credit it with;
Certainty: “cause” is a strong word. Falling asleep faster on four nights is a hint that the phone might matter, not proof that it causes anything;
Scope: “everyone should” leaps from your own bedroom to a rule for all people, which four nights cannot possibly reach.
Strip the overreach out and a defensible claim is still standing: “Over four nights, keeping my phone out of the room seemed to help me fall asleep faster, so it is worth trying for longer to see if the pattern holds.” That version says everything your evidence earns and not a word more. It is quieter than the announcement you started with, and it is the only one of the two you could actually defend.
Overreach is drawing a conclusion broader, bolder, or more certain than your evidence can support, and it is separate from how confident you sound;
The rule that prevents it is to match the strength of your conclusion to the strength of your grounds, keeping both the scope and the certainty under what the evidence supports;
Tightening a claim shrinks its scope and softens its certainty while keeping the honest core, producing a claim you can actually defend;
Qualifying words like “some,” “so far,” and “seems” size a claim to its evidence, and a scoped claim is harder to knock down than an overreaching one;
Qualifiers only work when they match the evidence: a hedge in front of a wild claim is still a wild claim.
With your reasoning now scoped to what the evidence bears, the next chapter shows you how to organize all your evidence under a single claim, so each piece has a clear place and job.