Adam Grant Writing Style

Students came to Adam Grant with a plan to sell eyeglasses online, and he passed on the chance to invest. They had kept their day jobs. Grant read that as a shortage of conviction, and he says so on the opening pages of Originals, in a book whose argument is that hedging your bets is what people who change things tend to do. Warby Parker made him look ridiculous. He hands the reader that fact before he asks the reader to give up anything at all.

That order is the method. Will Storr gathers the work on what happens when somebody's belief gets contradicted, and the useful part for a writer is how narrow the danger zone turns out to be. We do not defend trivia. Sarah Gimbel scanned people who were shown evidence against their political convictions and found a response close to what a brain produces on meeting a bear in a forest, and the beliefs that set it off are the ones a person has built an identity on. So the belief has to come off the reader before it can be argued with. Somebody else was wrong first, usually him.

The labels do the second half of the work. Givers, takers and matchers in Give and Take. Preachers, prosecutors and politicians in Think Again, with the scientist held back as the mode the reader has not claimed yet. Each set is built from words a reader already owns, and each arrives before the verdict, so you file yourself under one of them and only then find out what it costs you. A category can be argued with. You cannot argue with a reader about who they are, and Grant rarely tries.

Read him for the sequencing rather than the sentences. Gladwell opens a hole in your knowledge and keeps you reading to fill it, which is a different engine: Grant tells you the answer early and spends the chapter making it survivable to accept. That order lifts off social science cleanly. It works on any subject a reader arrives at with an opinion already formed, which is most of them.

The concession that costs something

A concession works when the reader can price it. Give and Take argues that generosity pays, and the same book reports that givers pile up at the bottom of the productivity measures as well as the top, which is the most expensive sentence in it and arrives early rather than in a footnote. Brené Brown treats disclosure as the subject itself. Grant treats it as a down payment on the reader's turn, so the thing conceded has to be one his argument would rather keep. An admission of perfectionism buys nothing. The reader can tell it was free.

One researcher, named, per turn

Count the studies in one of his chapters and the number comes out lower than the reputation suggests. One finding does the work, and it belongs to a person with a name. Michael Housman looked at which browser customer-service employees used and found that the ones who had installed Firefox or Chrome stayed in the job longer than the ones who opened whatever came with the machine. Originals takes that single result and rebuilds what commitment looks like around it. Naming the researcher is not a courtesy. It keeps the sentence honest, because a finding attached to a person can be checked, argued with and eventually overturned, while a finding attributed to research floats free of anyone who could be wrong.

The argument between two other people

He stages the disagreement between two people who are not the reader. An entrepreneur and an investor. An analyst inside the CIA and the hierarchy she wanted to change. A champion debater and the machine built to beat him. Watching two other people argue is free. A reader who takes a side has committed to a position without having been asked to change one. The move is old and underused, and its risk is that a badly chosen pair turns into a parable, with one figure built out of straw so the other can win.

Asking instead of telling

Think Again spends a chapter on motivational interviewing, the technique where you stop arguing and start asking what would move the other person, and Grant then writes the way the technique prescribes. Whole paragraphs are questions. What would have to be true for you to change your mind here? What evidence would you accept? The questions are not decorative, since a reader who answers one has produced the objection themselves and now has to deal with it, which is a harder thing to walk away from than a claim somebody else made. Used carelessly it goes wrong fast, and the tell is a question the writer answers in the next sentence.

Confident humility

Think Again names the stance and the prose performs it. He states a claim flatly, then says what would change his mind, and the second half is what licenses the first. A writer who qualifies everything sounds like somebody covering themselves. A writer who qualifies nothing sounds like somebody who has stopped listening. Grant sits between the two by staying certain about the claim and open about the test, which reads as strength rather than caution because he has told you where he could be beaten. The habit has a side effect worth knowing about. Readers start bringing him counterexamples instead of complaints.

The exit that keeps the reader's face

The wrong belief arrives as the sensible default, held by capable people for good reasons. Then its reach turns out to be shorter than they thought. Changing your mind costs status. A reader is doing that arithmetic whether or not the writer notices, and Grant pays the bill in advance. The mistake belongs to a category of thoughtful people, and leaving it is offered as an upgrade rather than a retraction. Overdone, this curdles into flattery, and the line between the two is whether the position the reader arrives at is harder to hold than the one they left.

The two-stage negation

Before he denies anything, he spends real prose making the wrong answer attractive. Steven Pinker's account of why that is necessary is worth having in front of you: a reader who hears a proposition believes it for a moment, and tagging it false takes an extra step, so a denial only lands when the thing denied was tempting in the first place. Deny a claim with no pull and the reader answers, who ever said that. Give and Take opens on a venture capitalist who behaves generously toward an entrepreneur and watches the deal go elsewhere, and the book has therefore made the case against its own thesis before it states it.

Labels a reader tries on

Giver, taker, matcher. Preacher, prosecutor, politician. Not one of those words needs defining, which is the point, because a coinage that needs a paragraph of setup cannot be used by a reader in the middle of a meeting. He also names states rather than objects. A New York Times essay gave the flat, joyless middle of the pandemic a word. Languishing. Readers recognized themselves in it before they had read the argument. Susan Cain does something similar for introverts, handing over a category people are relieved to join. Grant's are less comfortable, since two of his three are the ones you were hoping not to be.

The condition instead of the hedge

Watch what he does with a limit. A weaker writer softens the claim. Often and in some cases get sprinkled through the sentence until the advice has no edge left and cannot be wrong. Grant keeps the force and moves the boundary instead: the claim holds under stated conditions, for stated people, and outside them it does not. Givers finish last and they also finish first, and the variable that decides which is whether the giver protects their own time. A reader can act on that. A reader cannot act on usually.

The reversal performed by somebody else

His exhibits are people caught in the act of changing their minds, which means the reader watches the move before being asked to make it. Think Again opens at Mann Gulch, where a fire crew ran uphill from a blaze carrying their tools, and the foreman who survived was the one who stopped, dropped what he was holding and lit a fire in front of him. Later chapters follow a blues musician who talks men out of their Klan robes, and a doctor in Quebec who changes a mother's mind about a vaccine by asking questions rather than presenting facts. The pattern is steady. Somebody with every reason to hold their position lets go of it, on the page, at a cost, and by the time Grant turns to the reader the act has stopped looking like surrender.

Arguing with an assumption no reader holds

This genre rewards the reversal, so the quickest way to produce one is to invent a belief and knock it down. Readers know. Steven Pinker's pair of examples is exact: a whale is not a fish gives you something to correct, and a herring is not a mammal produces a shrug and the question of who ever thought it was. Once a reader has asked that question, every later claim reads as manufactured, including the true ones. Test your assumption by finding somebody who acts on it. Name a person, a company or a policy that behaves as though it were true. If you cannot, you do not have a target.

Confessing to something that cost nothing

The imitation opens with a failure, as instructed, and the failure is that the writer used to work too hard. That is a job interview answer. A reader recognizes the shape at once, which leaves the rest of the piece asking to be trusted by somebody who dodged the one moment when honesty would have been cheap. What the real version buys is permission. Will Storr's account of the model-defending brain explains why that permission is needed, since a reader will not revise a belief in a room where revision looks fatal, and the writer going first is the only evidence available that it is survivable.

Handing out the categories and the diagnosis together

Writers get as far as the taxonomy and then finish the job for the reader. In comes the sentence that says which type you probably are. That sentence undoes the taxonomy. A reader who was sorting themselves a moment ago is now being sorted. The defensive machinery Grant spent the chapter avoiding switches on in a clause. Give the reader the categories and then a few paragraphs of quiet. A self-diagnosis somebody arrives at alone is one they will still hold next week, which is the only kind worth engineering.

Inverting a true claim to buy a surprise

Surprise is what gets a piece passed around, and the fastest route to it is to take something true, flip it, and go looking for support afterwards. The support turns up, since support turns up for almost any position a person is determined to hold, and the piece that results is technically defensible and quietly false, which is worse than being wrong in a way somebody can correct. Grant works the other way round, which is why several of his reversals are partial. Givers do better and also worse. A reversal that survives keeps a version of the old answer alive inside it, and if yours has no such version, you have probably inverted rather than investigated.