Ernest Hemingway gave the same piece of advice to anyone who asked how he wrote so steadily. Stop for the day while you still know what comes next. Never write yourself empty. Leave a sentence half-finished, a scene unresolved, a thread you already know how to pull. Then walk away.
It sounds like a recipe for losing momentum. In practice it does the opposite. The next morning he didn't sit down to a blank page and the awful question of where to begin. He sat down to a sentence that was already moving, and he simply caught up to it. The hardest part of the work — starting — had quietly been solved the night before.
There's a name for the force Hemingway was exploiting, and it comes from a psychology lab in 1920s Berlin.
The unfinished task that won't let go
Maria Ovsiankina was a student of Kurt Lewin, part of a remarkable group studying how unmet goals shape behavior. In her experiments, she gave people simple tasks — assembling a figure, completing a small puzzle — and then interrupted them partway through, often on a pretext, before they could finish.
What she found is now called the Ovsiankina effect: people show a strong, spontaneous tendency to return to an interrupted task and complete it, even when they're no longer asked to, even when there's no reward, even when they could easily move on to something else. Some participants resumed the task the moment the interruption ended. Others picked it back up after being handed a different activity entirely. The pull to close the loop was remarkably persistent.
This is a close cousin of the better-known Zeigarnik effect, which describes how we remember unfinished tasks more sharply than completed ones. But the two aren't the same thing. Zeigarnik is about memory — the open task stays loud in your mind. Ovsiankina is about action — the open task actively draws you back toward it. One keeps the file open; the other reaches over and reopens it for you.
The underlying mechanism, in Lewin's framework, is a kind of psychological tension. Starting a task creates an intention that behaves like a charged system. Finishing discharges it. Leave it unfinished, and the tension persists, gently insisting on resolution. Your mind treats an interrupted task as a small debt it would prefer to pay.
Why starting is the part that hurts
Most of us experience the cost of work backwards. We imagine the hard part is the doing — the long middle stretch of effort. But anyone who has stared at an untouched project for an hour knows the truth. The expensive moment is the first one. The transition from not-working to working is where the resistance lives.
There's a reason for that. When a task hasn't begun, it exists in your head as a vague, undefined whole — every sub-step, every unknown, every possible way it could go wrong, all bundled into one heavy lump labeled "the report" or "the redesign." Beginning means confronting that entire mass at once and deciding where the first cut goes. The ambiguity itself is exhausting, and ambiguity is what we avoid.
A task already in motion has none of that weight. The first cut has been made. The shape is visible. You're no longer deciding whether and how to start — you're continuing, which is a fundamentally easier action. This is why the blank page is terrifying and the half-written page is merely work.
The Ovsiankina effect tells us something useful here: an interrupted task doesn't just feel easier to resume, it generates its own pull toward resumption. So the question becomes practical. What if you could engineer that pull on purpose?
Stop on a downhill, not at the bottom
The usual instinct is to push each work session to a clean stopping point. Finish the section. Close the task. Reach the natural end, then quit. It feels disciplined and tidy.
It also hands your future self a blank page. A completed task discharges the tension Ovsiankina identified. There's nothing pulling you back, so tomorrow you start cold, from zero, against the full weight of beginning.
The alternative is to stop deliberately while the task is still in motion — and, crucially, while you still know exactly what the next move is. Not stuck. Not depleted. Mid-stride, on a part you understand, with the path forward clearly in view. You're not abandoning the work; you're setting a hook.
The difference between the two is the difference between stopping at the bottom of a hill and stopping on the slope. From the bottom, every restart is a climb. From the slope, you just let go of the brake.
How to leave a hook on purpose
This is a habit you can build into the seams of your day, and it costs nothing but a small act of restraint.
Quit one step early. When you sense a natural stopping point approaching, don't take it. Push slightly past it into the next piece, do just enough to make that piece concrete, and stop there. Now the open loop is fresh and specific instead of finished and closed.
Write down the next move, not the status. Before you step away, leave yourself a single line that names the very next action — "draft the second example for the pricing section," not "work on pricing." A precise next step is a hook with a point on it. A vague one slides right out.
Stop where it's easy, not where it's hard. This is the part people get wrong. Don't quit at the wall you're stuck against — that just hands your future self the wall. Stop where the way forward is obvious, so resuming feels like coasting. Save the easy part to start on, not to finish on.
Resist the urge to tidy. Closing every tab, resolving every thread, and ending the day at a clean zero feels good in the moment. But you're discharging exactly the tension that would have carried you back in. A little productive untidiness is the whole point.
The through-line is the same in each case: leave the system charged. Hemingway leaving a sentence unfinished, a programmer stopping in the middle of a function they already understand, a student closing the book mid-paragraph rather than at the chapter's end — all of them are exploiting the same century-old finding. An open loop reaches back for you. A closed one lets you drift.
The cost of always finishing
There's a quiet irony in all of this. The advice to "always finish what you start" is good for the task in front of you and bad for the task after it. Relentless closure leaves you, again and again, at the foot of the next hill with nothing pulling you forward.
The more durable rhythm isn't finish-then-rest. It's stop-while-moving, then rest. You end each stretch of work not at an exhausted zero but at a charged, specific, slightly-unfinished point — and you let the Ovsiankina effect do the hardest part of tomorrow's work for you tonight, for free, while you sleep.
Where Zenith fits
This only works if the hook survives the gap. A next step you can clearly picture at five o'clock is useless if it's evaporated by nine the next morning — which is exactly what happens when the only place it lives is your tired memory. Zenith is built to hold that hook for you: a place to set down the precise next action the moment before you step away, so the open loop stays open on paper instead of fading in your head. You return not to a wall of vague intentions but to a single, concrete first move already waiting — the tension preserved, the climb already half-done. If you've ever lost a morning trying to remember where you left off, that's the gap worth closing. You can see how it works at https://zenith.lumenlabs.works.