You only notice what you have a name for
The categories your knowledge tool provides shape what you notice and capture — not just how you organize things afterward.
In 2007, Jonathan Winawer gave English and Russian speakers a simple task: identify the odd-colored square in a row of three. The colors were various shades of blue. English speakers performed roughly the same regardless of which shade differed. Russian speakers were faster on one specific condition — the case where the odd square crossed the boundary between siniy (dark blue) and goluboy (light blue). Two different words in Russian. One word in English.
Same photons. Different perception. The only intervening variable was the category.
The result has been replicated, refined, and disputed for nearly twenty years. But even skeptics agree on the core finding: having a name for something makes you faster at seeing it. The category isn't just expression. It's a seeing instrument.
Your knowledge tool works the same way.
The note is not a neutral container
Most personal knowledge tools give you one primitive: a note. A title, a body, maybe a tag. This looks like flexibility. It's actually a strong default.
When everything is a note, you notice note-shaped things. Quotes you want to keep. Ideas you want to develop. Summaries of things you read. The note format selects for thoughts that are already complete enough to write down — things you're already consciously processing. Everything else slides past.
Dan Bricklin noticed something similar in 1979 when he built VisiCalc, the first electronic spreadsheet. His insight wasn't that people needed faster arithmetic. It was that "what-if analysis" — changing one assumption and watching it ripple through a model — existed as a business practice in theory but was impractical enough that almost nobody did it. The calculation took too long. When VisiCalc made recalculation instant, it didn't automate something people were already doing. It made a new kind of thinking feel natural enough to have.
A tool doesn't just capture what you already know you want. It shapes what you notice as worth capturing at all.
What typed categories teach you to see
Imagine you start keeping structured records about the people in your life. Not an address book — something closer to a live document for each person: their context, things they've mentioned, what they're working on, when you last spoke.
Within a few weeks, something shifts. You're in a meeting and someone mentions offhand that their daughter just started college. You used to let that go. Now you notice it — not because you're trying harder to pay attention, but because you have somewhere to put it. The category exists, so the observation becomes legible.
The same thing happens with decision records. If your knowledge tool has a structure for decisions — context, options considered, what you chose and why — you start noticing when a decision moment is happening, not just what you eventually decided. Before, decisions happened to you. After, you see them arriving.
Eleanor Rosch showed in her 1973 work on natural categories that people learn category-organized information faster than information without category structure. Her more specific finding was about recognition speed: having a category makes you faster at identifying new instances of it. The category isn't just a filing system applied after the fact. It runs at input time.
What the schema asks of you
This matters practically because of what slips through without it.
The most valuable information you encounter doesn't always arrive in an obviously important package. Sometimes it's a preference someone mentions once and never again. A decision that got made implicitly, without anyone naming it as one. A pattern you'd notice if you were looking for it — and you aren't, because there's nowhere for it to land.
Generic notes are optimized for things you already know are significant. Typed structures — person records, decision records, preference records — are optimized for things that would be significant if you had a name for them. That's a different optimization target, and most knowledge tools have never considered it.
There's a design implication that rarely gets discussed: a knowledge tool's schema is a claim about what kinds of things in the world are worth noticing. The generic note says everything worth knowing can be expressed as a prose fragment with a title. It isn't wrong. But it leaves a lot unnamed, and unnamed things don't make it in.
People evaluate knowledge tools by how fast search is, whether the sync works, how easy export is. Those things matter. But they're all downstream of a prior question: what does this tool teach me to see? The retrieval problem is real, but it's about things you already captured. The harder problem is everything you didn't capture because you didn't have a category for it. You weren't inattentive. You were just using a tool with one word for blue.
Asgeir Albretsen is the founder of Harbor.