Field Notes Issue No. 07 August 22, 2026 Est. reading — 6 min

An almanac, not a course

How learning
actually works

Five field notes on attention, memory, and practice — drawn from what cognitive research has actually measured, not what studying feels like it should be.

fig. i — retention across three reviews
i. Attention is the narrowest part of the whole system — everything downstream depends on what gets through it first.
Multitasking doesn't split attention evenly. It switches it, at a cost, every time.

Attention is a bottleneck, not a skill

de attentione

Nothing gets encoded into memory without first passing through attention, and attention can only hold a handful of things at once. What feels like "focus" is really the absence of competing claims on that narrow channel — a quiet room, a single tab, a single task.

This is why background noise, notifications, and task-switching cost more than they seem to. Each switch has a measurable tax: a few seconds of degraded performance while the mind re-loads the discarded context. Ten switches an hour is ten small taxes, not zero.

ii. Rereading feels productive because it increases familiarity, which is easy to confuse with retention.
Highlighting a textbook is one of the least effective study habits measured — it feels active but isn't.

Recognition is not recall

de memoria

Reading a page again makes it feel more familiar, and familiarity is easy to mistake for knowing. But recognizing a fact when you see it and retrieving it unaided are different mental operations, and only one of them is what's needed in an exam, a conversation, or a real task later.

Closing the book and trying to recall the material — even badly, even with gaps — does more for retention than several more rereads. The struggle to retrieve is not a sign that the method is failing; it's the mechanism doing its job.

iii. Spacing out practice feels slower and worse in the moment. It performs better weeks later.
See the review ledger further down this page for a working interval schedule.

Spacing beats cramming, reliably

de intervallo

Massed practice — one long session — produces fast improvement that fades fast. Spaced practice, the same material reviewed with growing gaps in between, produces slower improvement that lasts. This is one of the most consistently replicated findings in learning research, and one of the least followed in practice, because the spaced version feels harder in the room.

The forgetting that happens between sessions isn't wasted time — it's what makes the next retrieval effortful enough to matter.

iv. Mixing topics within a session slows visible progress but builds the ability to tell them apart later.
Blocked practice: aaa bbb ccc. Interleaved: abc bac cab. The second is harder and works better.

Mixed practice builds judgment

de varietate

Practicing one type of problem repeatedly in a block makes each attempt easier than the last, because the strategy stays the same. Mixing problem types forces a decision each time — which approach applies here — which is closer to what real use of a skill actually demands.

The cost is visible immediately: more errors, slower sessions, more friction. The benefit shows up later, in transfer to problems the learner hasn't seen arranged that way before.

v. Confidence and competence are measured separately, and they frequently disagree.
Asking "how would I explain this to someone else?" surfaces gaps that rereading hides.

Feeling confident is not the same test

de iudicio

Fluent material feels known. But fluency is partly a property of how the material was practiced — familiar formatting, a familiar room, a familiar order — and it can be high even when actual recall would fail under different conditions.

Checking understanding by generating an explanation, working a problem cold, or teaching the idea to someone else tends to reveal a gap that simply reviewing notes again does not.

A working ledger

One spacing schedule

There is no single correct interval — this is one commonly used schedule for reviewing the same material over time, expanding the gap each time recall succeeds.

Review Gap since prior What the interval is doing
1st review 1 day Resets the curve while the trace is fresh.
2nd review 3 days The gap most people quit at — recall feels effortful.
3rd review 7 days A week is long enough that recall, not recognition, is being tested.
4th review 14 days The interval most study apps stop at by default.
5th review 30 days What survives a month tends to survive a year.

A short glossary

Encoding
Turning something you perceive into a form memory can hold — the moment information first sticks, or doesn't.
Retrieval practice
Pulling an answer out of memory unaided, rather than recognizing it — the single most reliable study method measured.
Interleaving
Mixing related topics within one sitting instead of blocking them, which slows practice but strengthens later recall.
Desirable difficulty
A level of effort hard enough to force real retrieval, but not so hard that it fails outright.
Transfer
Using a skill in a context different from where it was learned — the actual goal of most education.
Metacognition
Noticing what you do and don't actually know, distinct from how familiar the material feels.