NotesNudge the notes you never go back to

What We Still Don't Know

A site that spends its time correcting overstated claims owes an account of where its own confidence runs out.

Almost everything specific to note-taking is unstudied. The memory research is strong and it is about laboratory materials and deliberate learning. The gap between that and four thousand personal notes is enormous, and it is bridged mostly by assertion — including, at points, by this site.

For a workplace-oriented comparison with this kind of workflow, this guide provides additional context.

Reviewed August 9, 2026.

The six

Whether resurfacing does anything. The central question here and there is essentially no direct research. Rediscovery and recombination are plausible, consistently reported, and unmeasured. Nobody has run the study.

Whether note-taking practices differ in outcome. Zettelkasten, PARA, folders, flat archives — no comparative evidence exists. Every strong recommendation you have read, here included, is reasoning from adjacent findings and personal experience.

For an independent external reference related to this topic, see Science.

What the right interval is for ideas. Spacing research gives intervals for facts with retention targets. An idea has no retention target and no correct answer to be graded against, so the parameters do not transfer and nobody has established replacements.

Whether linked archives produce insight. The central promise of a decade of tooling and the least examined claim in the field.

What the long-term effects of offloading are. People offload readily and it changes what they remember. Whether that is good, bad or neutral over years is open.

And whether any of this matters more than writing at all. The encoding benefit arrives at the moment of writing. It is entirely possible that everything downstream — organising, reviewing, resurfacing — is a rounding error against simply writing more, thoughtfully. Nobody has tested that either, and it would be the most useful study in the field.

Why the evidence looks like this

Laboratory materials are studiable and personal archives are not. You can randomise word lists. You cannot randomise someone's nine years of notes, control what they wrote, or measure an outcome that means anything.

The outcomes are undefined. What is the dependent variable for "my notes work"? Retention of what? Ideas produced? Satisfaction? Nobody agrees, so nobody measures.

And the timescale is wrong for research. The interesting effects operate over years. Studies operate over weeks.

None of this is negligence. It is what a genuinely hard research problem looks like, and it means confident claims in this area are confident for reasons other than evidence.

What this site does about it

Label the category. Established, plausible, or asserted. Every claim about memory here should carry its material and its measurement.

Prefer mechanisms to prescriptions. "Attempted recall is what the evidence is about" is defensible. "Review at these intervals" is not.

And distinguish my practice from my evidence. Ten minutes a week on a marked set is what I do and what I recommend. The support for it is that spaced effortful retrieval is well established and that a practice which happens beats one that does not. That is reasoning, not a finding, and it should be read as such.

What would change things

A study of resurfacing with an actual control, measuring rediscovery rather than retention.

A comparison of note methods over a year, with outcomes defined in advance.

And, most usefully, a test of whether reviewing adds anything to writing. If it does not, most of this site is elaborate and unnecessary, and finding that out would be worth more than any of the refinements.

I do not expect any of these soon. They are expensive, unglamorous, and there is no product waiting at the end of them.

The short version