The Should Trap

"I should really be someone who works out." "I should really be someone who gets up earlier." "I should really be more disciplined about this."

Notice the shape of that sentence. It is always about someone you are not yet, someone up ahead, someone you are supposedly working toward. Every morning you fall short of that person becomes a little evidence against yourself.

There is an old idea, older than any productivity book, running through Dale Carnegie, Napoleon Hill, and further back to James Allen's line that a person is literally shaped by their habitual thought. Become the person first through thought and the behavior follows. For a long time that sounded like a nice idea with no real mechanism behind it, the kind of thing that shows up in a hundred years of self-help books with no one quite explaining why it would actually work.

It turns out there is a mechanism, and it is more direct than most people would guess. You do not have to earn the identity first. You can be handed it, or hand it to yourself, before the evidence is anywhere close to solid, and the behavior will often move to catch up with it.

Why the Label Comes First, Not Last

In 1975, researchers Richard Miller, Philip Brickman, and Diana Bolen ran a study across three fifth-grade classrooms with a simple setup. One classroom was told, repeatedly and most importantly unconditionally, over several days, that they were neat and tidy people. A second classroom was told, just as often, that they should be neater and should stop littering. A third classroom got nothing.

Here is the part worth sitting with. Nobody sorted the tidy kids into one classroom first and told them so. The same identity statement went to every student in that room regardless of that particular kid's actual litter habits the week before. The label was not a personalized assessment of who had earned it. It was simply asserted, out loud, again and again, to the whole room at once.

And it worked anyway. The neat identity group littered less and cleaned up after themselves more than either the persuasion group or the control group, and they were still measurably tidier two weeks later. The persuasion group, the one told what they should do, barely moved, and what little progress it made faded fast.

That is a stranger and more useful finding than "reward good behavior when you see it." Nobody waited to see the good behavior first. The researchers simply told a room of ordinary kids who they were, and a meaningful number of them started acting like it was true. The label did not describe reality. It got ahead of it, and behavior spent the next two weeks catching up.

Worth naming the limitation plainly: this was one classroom per condition, not each child individually assigned at random, so some of the difference between rooms could in theory reflect something other than the label itself. It is also one study from 1975, on one age group, in one setting. The direction of the finding, that the identity statement beat the should statement and did it by a wide margin, held up cleanly enough that it is still the reference study people point to fifty years later.

The researchers ran a second version of the same idea in a different setting, this time around math performance in school, to check whether it was a fluke specific to littering. It held up there too. Children told they had real ability in math improved more than children given other kinds of encouragement, again without that ability necessarily being demonstrated first.

This connects to something the psychologist Daryl Bem described in 1972, called self-perception theory. He argued that people often figure out who they are the same way an outside observer would: by watching what they do, especially when their own internal sense of things is not clear-cut. Put a label on a child before the evidence is settled, and their own ambiguous, everyday actions, the ones that could go either way, start getting read through that label instead of against it. The identity does not wait for proof. It recruits the proof afterward, from material that was already sitting there.

Leading with what you should be instead of what you are choosing to already be.

"You should stop leaving your dishes in the sink" and "you are the kind of person who cleans up after himself" are aimed at the same behavior. The classroom study is fairly blunt about which one actually moves the needle. The should-version was the one that barely worked and faded fastest. The is-version was the one that held up two weeks later.

Most people default to the should-version without noticing, on themselves most of all. "I should exercise more." "I should be better about mornings." Every one of those sentences quietly confirms the gap instead of closing it, the same way telling a kid he should stop littering just reminded him he was currently a litterer.

The fix is not waiting for enough proof to honestly claim the identity. The study's own children did not have that proof yet either. The fix is choosing the is-version on purpose, before the evidence is finished, and letting your own actions get read through it the way the researchers found they naturally will. It turns out the old saying "fake it till you make it" seems to have been true all along.