Think Again

Think Again is Adam Grant's guide to revising assumptions, updating beliefs, and keeping decisions corrigible without turning doubt into paralysis.

Reviewed by the Gollius editorial team. Editorial policy

Think Again is useful because it gives confidence a repair mechanism. Adam Grant's book is not a command to distrust every conviction, perform doubt in public, or keep arguing until all action becomes impossible. Its best use is more disciplined: notice when a belief is doing more work than the evidence can support, update the belief, and still decide.

That makes the book especially valuable for people who care about critical thinking and decision making but do not want skepticism to become a lifestyle brand. Rethinking is not the opposite of judgment. It is judgment with an edit button.

The Book's Real Subject

Grant writes about rethinking as a habit of mind: the willingness to revise views, invite disconfirming evidence, and treat identity as larger than being right about one position. The book is often sold in a business-friendly language of innovation and persuasion, but its deeper practical theme is cleaner belief maintenance.

That distinction matters. Someone can agree with rethinking in theory and still use it badly. One person turns it into debate theater. Another uses it to avoid choosing. Another updates beliefs only when the update protects status. The book is strongest when it becomes a private operating rule: before an important decision, ask which assumption would change the action if it proved false.

Intellectual Humility Is Not Self-Erasure

Intellectual humility is sometimes confused with lowered confidence. That is too blunt. The useful version is not "I know nothing." It is "my current view may be incomplete, and I can name what would make me revise it."

That keeps humility distinct from paralysis. It also keeps it distinct from surrendering judgment to the loudest person in the room. Good rethinking preserves agency. It lets someone hold a view provisionally, test it honestly, and then act with an appropriate level of confidence.

The related work on intellectual humility and practical doubt is a helpful companion because it separates openness from gullibility. A corrigible mind does not accept every counterclaim. It asks whether the counterclaim has better contact with reality.

Where The Book Helps Most

The book is most useful in three situations. First, when a repeated conflict shows that a hidden assumption is driving behavior. Second, when a plan keeps failing but the explanation stays the same. Third, when a group has rewarded certainty so strongly that people stop reporting weak signals.

In those moments, rethinking is not decorative. It changes the next move. A hiring decision may need a second criterion. A personal routine may need a smaller test. A disagreement may need a shared definition before another opinion. The page on confirmation bias is relevant here because the easiest evidence to collect is usually evidence that protects the existing story.

A Bounded Assumption Update Test

Use this test for one decision that matters but is not catastrophic. Write one sentence: "I am acting as if..." Complete it with the assumption currently guiding the decision. Then add three fields:

  1. What evidence would make this assumption weaker?
  2. What small action would change if that happened?
  3. What date ends the test?

Run the test for fourteen days. During that period, collect only evidence that can change the action, not evidence that merely fuels commentary. The stopping rule is simple: if the evidence does not change a concrete next step by the review date, close the test and decide with the best current view. If it does change a next step, make one adjustment and stop searching until the adjusted action has produced new feedback.

This keeps rethinking from turning into endless research. It also protects against the opposite error: making a permanent identity out of a temporary assumption.

Disagreement Without Performance

One of the book's practical promises is better disagreement. That promise needs restraint. Debate can reveal weak reasoning, but it can also reward verbal speed, status games, and social pressure. A person with less power may "lose" a discussion while being closer to the truth.

Before turning a disagreement into a debate, ask for the job of the conversation. Is it to understand, decide, repair, negotiate, or test a claim? Each job needs different behavior. If the goal is understanding, repeat the other person's position until they recognize it. If the goal is decision, name the criteria before arguing. If the goal is repair, do not hide behind abstract reasoning when an apology or changed conduct is the real issue.

The broader catalogue of cognitive biases helps here, but bias language should not become a weapon. Calling someone biased is rarely as useful as naming the specific evidence gap.

The Critique: Rethinking Has Conditions

The book can underplay the politics of who gets to rethink safely. In some workplaces or families, changing your mind is praised only when it serves the existing hierarchy. In public settings, admitting uncertainty can be punished more harshly than bluffing. In urgent situations, more analysis may be a way to avoid an uncomfortable but necessary decision.

There is also a commercial boundary. Author pages and publisher pages can verify the book's identity and framing, but they cannot prove that reading the book produces the outcomes the marketing implies. Treat those materials as identity and thesis sources, not as evidence that a method will work for everyone.

The best critique is not to reject the book. It is to use it where rethinking has a fair chance to improve conduct: decisions with feedback, relationships where questions are allowed, teams that can tolerate correction, and personal habits where evidence can be gathered without turning life into a lab.

A Decision Journal Makes It Real

Rethinking becomes concrete when it leaves a trace. Keep a small decision journal for decisions that are repeated, expensive, or emotionally charged. Record the assumption, the alternative explanation, the action chosen, and the review date.

Do not record everything. That creates audit fatigue. Pick one decision per week. At review, score only three things: what was true, what was assumed, and what changed. If nothing changes after several reviews, either the decisions are too small or the journal is becoming theater.

How To Read It With Proportion

Read Think Again as a training manual for corrigibility, not as an ideology of uncertainty. The book pairs well with Adam Grant as an author page because his public work often focuses on work, psychology, and organizational behavior. Still, the task is not to import a public persona. It is to test one habit: can I change my mind before reality has to punish me loudly?

The decision-useful conclusion is this: use the book when a belief is repeatedly shaping action and the feedback is messy. Do not use it to delay decisions that already have enough evidence. Strong rethinking ends in cleaner commitment.

Sources