When Taking Notes Makes You Forget

In 2013, a psychologist at Fairfield University named Linda Henkel led visitors through a museum and asked some of them to photograph the objects they saw and others to simply observe them. The next day, she tested their memories. The people who took photos remembered less — not just in general, but specifically about the things they photographed. They couldn't recall details. They recognized fewer objects. The camera had done something unexpected: it gave their brains permission to stop paying attention.

Henkel called this the photo-taking impairment effect. The mechanism she proposed is cognitive offloading: when you know an external system is capturing something, your brain doesn't bother encoding it deeply. The record exists. You don't need to remember. So you don't.

The obvious extension is to note-taking. And the obvious implication is uncomfortable for anyone who has built a capture-first productivity system.

The collector's excuse

There's a style of knowledge management that treats capture as the primary virtue. Clip everything. Save everything. Dump it all into Notion, Obsidian, or your notes app of choice. Process it later. The idea is that no useful thing should escape the net.

The problem is that "later" usually doesn't come. But even when it does, something more fundamental needs reckoning with: passive capture might be reducing your understanding of what you captured.

When you clip a web article with one tap, your brain sees the transaction complete. The information is "saved." The anxiety resolves. What the tap doesn't do is make you understand the article better, connect it to anything you already know, or decide what you actually think about it. It removes the cognitive pressure that would have made encoding happen.

This isn't an argument against capturing. It's an argument against the fiction that capturing is the same as knowing.

The zoom exception

Henkel found something important that often gets dropped when this research is cited. When museum visitors zoomed in to photograph a specific detail of an object — not just snapping the whole thing, but deliberately attending to a part — their memory for the whole object was actually good. Better, in some cases, than the non-photographers.

The act of directed attention during capture preserved the memory. Passive documentation did not.

This distinction matters more than the headline finding. It's not that recording causes forgetting. It's that recording without attending causes forgetting. The photograph that requires you to decide what to frame, where to focus, and what matters is doing cognitive work. The point-and-shoot is not.

The same split exists in note-taking. A note that forces you to categorize, summarize, or connect is doing cognitive work. A note that is just a paste is not.

What structure does that capture doesn't

Most note-taking tools optimize for the first kind of capture. The friction is in retrieval, not writing. You clip first, organize later, and "later" is where the value is supposed to emerge.

Structured capture works differently. When you're forced to decide what something is — a person, a preference, a decision, a task — you're doing the zooming-in. You're making an interpretive choice. That choice requires engagement. And that engagement is what makes things stick.

I've noticed this in my own use of Harbor. Writing a plain document about someone I met takes thirty seconds and produces a record I'll struggle to use later. Deciding that the person is an entity — typing in their name, their role, what we discussed, what I committed to following up on — takes ninety seconds and produces something that feels qualitatively different when I read it back. I had to think to write it. The thinking is in the note.

This isn't about making capture harder for its own sake. It's about the fact that deciding what something is forces a minimum of processing. Processing is what makes knowledge durable rather than just stored.

The quiet cost of low-friction capture

There's a popular pitch in the personal knowledge management space: capture everything, and your future self will thank you. It's not wrong exactly, but it skips over the real problem. Your future self will find the captured thing. Whether they understand it — whether it means anything to them — depends entirely on the quality of attention that went into recording it.

The tools that make capture most frictionless are often the ones that produce the least durable knowledge. One tap and it's "saved." Nothing in your understanding has changed. You have a record, not a thought.

Henkel's museum visitors thought they were building something they could return to. They were, technically. What they weren't building was a memory. The two are not the same thing, and conflating them is how you end up with a knowledge base full of notes you don't understand and can't use.

The question worth asking before you reach for the save button isn't whether this is worth keeping. It's whether you know what it is.

---

Asgeir Albretsen is the founder of Harbor.