The memory catalogue

Every memory system and technique in meaningful use, with what it claims, an example, an honest analysis, and how widely it is actually practised — which is usually the least-documented and most useful column.

Companion to memory-systems.md, which goes deeper on the classical families and their history. This one goes wider.

How adoption is rated

The gap between fame and adoption is the recurring story here. The memory palace is the most famous technique in the world and is practised by a few thousand people. Chunking is practised by everyone alive and almost nobody can name it.

1. Spatial systems

Method of loci / memory palace

• Principle — spatial memory is unusually robust and effortless. Fix an ordered route through a place you know, put one vivid image at each position, and walk the route to recall.
• Example — the twelve eras of the history strand placed as twelve hours on a clock face; each hour holds one event's image.
• Analysis — the strongest single technique for ordered material, and the only one that lets you detect a missing item. Dresler et al. (2017) trained novices for six weeks and got large durable gains plus athlete-like connectivity changes. Costs real effort to build; reusing a palace for new material causes interference.
• In practice — Niche. Universally famous, rarely practised. Essentially 100% of competitive memory athletes; a small number of medical students; a rounding error of everyone else.

Journey method

• Principle — loci using a real route (a walk to the shops) rather than a building.
• Example — 46 European countries walked as a geographic route from Iceland to Russia.
• Analysis — scales better than a room; a familiar walk supplies dozens of loci with no invention. Dominic O'Brien's preferred form.
• In practice — Niche, and the most-used loci variant among people who actually do this.

Roman room

• Principle — one room, its objects as loci.
• Example — ten items on a shopping list around your kitchen.
• Analysis — the cheapest possible entry point; caps out around 10–20 items.
• **In practice — Common as the thing people try once,** Niche as sustained practice.

Memory theatre

• Principle — Camillo's 16th-century project: an architecture where position encodes meaning, not merely order — a map of knowledge rather than a filing cabinet.
• Example — a house where each room is built in the architectural style it stores, so walking Romanesque → Gothic is the contrast.
• Analysis — intellectually the most ambitious idea in the tradition and practically the least reproducible. Nobody has demonstrated Camillo's actually working.
• In practice — Historical, with the self-exemplifying trick worth stealing.

Songlines and country-as-memory

• Principle — encode knowledge into landscape traversed in song: sequence, ecology, law and genealogy bound to a physical route.
• Example — Australian Aboriginal songlines spanning hundreds of kilometres.
• Analysis — a genuinely continuous oral tradition of enormous capacity and the deepest time depth of anything here. Lynne Kelly's broader claim that Stonehenge and similar monuments served the same function is an interesting hypothesis, not an established finding — treat it separately from the Australian ethnography, which is solid.
• In practice — Living tradition, and not a technique outsiders can adopt; it is inseparable from the law and country it encodes.

2. Phonetic and number systems

Major System

• Principle — digits map to consonant sounds (0 s/z · 1 t/d · 2 n · 3 m · 4 r · 5 l · 6 j/sh/ch · 7 k/g · 8 f/v · 9 p/b); vowels are free filler, so numbers become words become images.
• Example — 34 = m-r = MOWER; 1789 = 17 + 89 = TACK + VIP.
• Analysis — the highest-leverage artifact in the tradition, and it only pays once you have a fixed 00–99 list. Improvising words mid-recall is slow and error-prone. Learn 110 images once, reuse for life.
• In practice — Niche. Standard equipment for anyone serious about numbers; near-zero outside that. Most people who know of it have not built the list, which is the same as not knowing it.

Dominic System

• Principle — two digits → two letters → a person, plus a characteristic action. People are more memorable than objects.
• Example — 43 = DC = David Copperfield, vanishing.
• Analysis — easier than Major for many people because faces and actions encode more richly than arbitrary nouns. Invented by an eight-time World Memory Champion, which is decent evidence.
• In practice — Niche/Specialist, and the main rival to Major.

PAO (Person–Action–Object)

• Principle — every 00–99 gets a fixed person, action and object; three pairs compress into one scene. Six digits per image.
• Example — 09-42-71 = Einstein (09) juggling (42) a teapot (71).
• Analysis — the standard competition rig. Several hundred hours of setup, and it only pays on long digit strings under time pressure.
• In practice — Specialist. Essentially memory sport only.

Ben System / Shadow System

• Principle — one image per number 000–999. A thousand images cold.
• Analysis — world-record equipment. Frontier territory with a brutal setup cost and no application outside the sport.
• In practice — Specialist, perhaps a few hundred people worldwide.

Number-shape and number-rhyme pegs

• Principle — a fixed image per digit from its glyph (1 candle, 2 swan) or its rhyme (1 bun, 2 shoe).
• Analysis — instant to learn, caps at ten, interferes badly if reused for two lists at once. The correct first technique to teach a child.
• In practice — Common as taught (widely present in school and popular books), Niche as sustained adult practice.

Katapayadi

• Principle — Sanskrit consonant groups encode digits, letting numbers hide inside meaningful words and verse.
• Example — the names of the 72 Carnatic melakarta ragas encode their own scale, so a trained musician derives the notes from the name.
• Analysis — arguably more elegant than the Major System because the carrier text stays meaningful. Predates Hérigone by centuries, independently invented.
• In practice — Living, and genuinely used — every Carnatic musician relies on it, mostly without calling it a mnemonic system.

Acronyms and acrostics

• Principle — first letters form a word or sentence.
• Example — ROYGBIV; "Every Good Boy Deserves Fruit"; SOHCAHTOA.
• Analysis — trivially cheap, and genuinely effective for short ordered lists. Fails badly beyond about seven items and gives you no way to recover a forgotten element from its letter.
• In practice — Ubiquitous. The single most-used deliberate mnemonic technique on earth, in every classroom in every country.

Rhyme, rhythm and song

• Principle — metre and melody constrain what words can fit, so errors become audible and the text self-corrects.
• Example — "Thirty days hath September"; the alphabet song; "In fourteen hundred and ninety-two".
• Analysis — hugely underrated. Rhythm is an error-detecting code, which is why oral traditions are metrical. A tune carries word order for decades with no rehearsal.
• In practice — Ubiquitous, and the first mnemonic every human encounters.

Keyword method

• Principle — for a foreign word, find a native-language sound-alike and image-link it to the meaning.
• Example — Spanish pato (duck) → "pot" → a duck in a saucepan.
• Analysis — one of the best-evidenced techniques in the education literature for vocabulary acquisition specifically, with the caveat that benefit fades if the keyword is not eventually dropped.
• In practice — Common in research and language-teaching materials, Niche among actual learners, who mostly reach for flashcards instead.

3. Association and elaboration

Link / chain method

• Principle — image item one colliding with item two, two with three.
• Analysis — the workhorse. Fragile in exactly one way: lose a link and everything downstream goes. No random access, and no way to know you have finished.
• In practice — Common. The default technique of anyone who has read one memory book.

Story method

• Principle — the link method woven into a narrative.
• Analysis — more robust than a bare chain because narrative supplies causality, which is itself a retrieval cue.
• In practice — Common.

Bizarre and vivid imagery

• Principle — unusual images are better recalled than mundane ones.
• Analysis — real but smaller than the popular literature claims, and it largely disappears when everything is bizarre. Vividness, motion and personal relevance do more work than strangeness.
• In practice — Ubiquitous within mnemonic practice, and over-sold everywhere.

Elaborative interrogation and self-explanation

• Principle — ask "why is this true?" and answer it, forcing integration with what you already know.
• Analysis — among the highest-rated techniques in Dunlosky et al.'s (2013) review of learning strategies, alongside practice testing and distributed practice. Cheap, general, and boring, which is why nobody talks about it.
• In practice — Common in instructional design, under-used by learners.

The generation and self-reference effects

• Principle — material you produce yourself, or relate to yourself, is retained better than material you are given.
• Analysis — Slamecka and Graf (1978) and a large subsequent literature. This is the reason a peg you write beats a peg you are handed.
• In practice — Ubiquitous as an effect, rarely exploited deliberately.

4. Structure and organisation

Chunking

• Principle — recode many units into fewer meaningful ones. Working memory holds a small number of chunks, and a chunk is whatever you have made into one.
• Example — a phone number as three groups; a chess position as five known patterns rather than twenty-five pieces.
• Analysis — the mechanism underneath most expertise. Chase and Ericsson's runner "SF" reached 80+ digits by recoding strings as running times he already knew, then organising those into a hierarchy — a retrieval structure. Expertise is not more memory, it is better addresses.
• In practice — Ubiquitous, and almost never named by the people doing it.

Hierarchies, outlining and retrieval structures

• Principle — two-stage retrieval: find the category, then the item.
• Example — era 08, then hour four.
• Analysis — the least glamorous idea here and arguably the most powerful. Two easy lookups beat one hard one, and the coarse stage survives long after the fine one has decayed.
• In practice — Ubiquitous implicitly, Niche as deliberate design.

Mind mapping

• Principle — radial diagram from a central concept, with colour and imagery.
• Analysis — Buzan's specific claims about brain function are not supported. The general practice of spatial, visual organisation is moderately useful, mostly because it forces you to decide what relates to what. Evidence for recall gains over good linear notes is weak.
• In practice — Common, especially in business and school settings, on weaker evidence than its popularity suggests.

Concept mapping

• Principle — Novak's version: nodes joined by labelled relationships, forming explicit propositions.
• Analysis — better evidenced than mind mapping precisely because labelling the edges forces you to state the relationship rather than gesture at it.
• In practice — Common in education research, Niche elsewhere.

Cornell notes, SQ3R and their relatives

• Principle — structured note formats reserving space for cues and summary, and reading protocols that force questioning and recitation.
• Analysis — their benefit comes almost entirely from the retrieval steps they smuggle in (the cue column, the "recite" stage), not the format.
• In practice — Common, taught in schools worldwide; adherence collapses after the class that taught it.

Zettelkasten and PKM

• Principle — atomic notes, densely cross-linked, so structure emerges from connections rather than a filing hierarchy.
• Example — Luhmann's paper slip-box; Obsidian, Roam, Logseq today.
• Analysis — this is external memory, not internal. It makes retrieval and connection-finding cheap; it does nothing to put anything in your head, and can substitute for learning while feeling like learning.
• In practice — Niche but conspicuous, with a large gap between people who set one up and people who still use it a year later.

5. Practice schedules

Spaced repetition

• Principle — review at expanding intervals timed just before predicted forgetting. Each successful retrieval flattens the curve.
• Example — Leitner boxes; SM-2 (Woźniak, 1987); FSRS's difficulty / stability / retrievability model; Anki, SuperMemo, Quizlet, Duolingo.
• Analysis — the best-evidenced intervention in this entire document, and orthogonal to every encoding technique above: it decides when, not how. Ebbinghaus measured the curve in 1885; everything since is scheduling on top.
• In practice — Common and growing. Genuinely widespread in medical education and language learning; Anki is close to standard for some medical exams. Duolingo and Quizlet put a form of it in front of tens of millions who have never heard the term. The most-adopted deliberate memory technology ever built.

Retrieval practice (the testing effect)

• Principle — being tested is not measurement, it is the learning event.
• Analysis — Roediger and Karpicke (2006). Retrieval beats re-reading by a wide margin, and the gap grows with delay. Learners consistently predict the opposite, which is why re-reading persists.
• In practice — Common and rising in instructional design; still fighting the intuition that re-reading feels more productive.

Interleaving and varied practice

• Principle — mix problem types rather than blocking them.
• Analysis — strong evidence, and it feels worse during study, which is why it is not adopted. A "desirable difficulty" in Bjork's sense. Directly relevant to confusable material: interleaving is what trains discrimination.
• In practice — Niche, despite the evidence.

Overlearning

• Principle — continue past first correct recall.
• Analysis — real short-term gains, poor cost-effectiveness compared to spacing the same minutes out.
• In practice — Ubiquitous as cramming, which is overlearning without spacing.

6. External and distributed memory

Writing, notes, calendars, reminders

• Principle — offload to a medium that does not decay.
• Analysis — the largest single upgrade to human memory ever made, and every subsequent one is a variation. Worth stating plainly because discussions of "memory technique" often ignore that literacy is the technique.
• In practice — Ubiquitous.

Search engines and the Google effect

• Principle — remember where to find, not what.
• Analysis — Sparrow et al. (2011) found people recall the location of information rather than its content when they expect it to remain available. This is transactive memory working correctly, not decline — but it fails when you need the knowledge in the room, in conversation, or to notice that something you are reading is wrong.
• In practice — Ubiquitous. The main reason deliberate memorisation now needs an explicit justification, and the honest one is: fluency and judgement require material to be in you, not reachable.

Tally sticks, quipu, notched bones

• Principle — physical objects as records.
• Analysis — quipu certainly encoded numeric and administrative data; the claim that they carried narrative is contested and unresolved.
• In practice — Historical.

Commonplace books

• Principle — a personal, indexed compilation of extracts.
• Analysis — the direct ancestor of the Zettelkasten, and Locke's indexing method is a genuine information-retrieval invention.
• In practice — Historical, reborn as PKM.

7. Physiological

Sleep consolidation

• Principle — memories are stabilised and integrated during sleep, particularly slow-wave sleep.
• Analysis — robust, large, and the cheapest intervention available. Sleeping after study reliably beats an equal waking interval.
• In practice — Ubiquitous passively, Niche deliberately. Almost nobody schedules study around sleep, and it would probably outperform any technique in this document.

Targeted memory reactivation

• Principle — replay a cue (an odour, a tone) present at encoding during slow-wave sleep to bias which memories consolidate.
• Analysis — real, replicated, effect sizes modest. Not the same thing as sleep-learning, which does not work.
• In practice — Experimental. Lab only.

Exercise, caffeine, nootropics

• Analysis — aerobic exercise has decent evidence for general cognitive and hippocampal benefit. Caffeine helps alertness, not consolidation. The nootropics market is largely unevidenced.
• In practice — Common, evidence quality descending sharply left to right.

8. Clinical and assistive

Errorless learning

• Principle — prevent mistakes during acquisition rather than correcting them, because amnesic patients retain the error implicitly.
• Analysis — Baddeley and Wilson. A genuine reversal of normal advice: desirable difficulties are for intact memory, not impaired memory.
• In practice — Specialist, standard in neurorehabilitation.

Spaced retrieval therapy

• Principle — spaced repetition adapted for dementia, using preserved procedural learning.
• Analysis — modest but real gains for specific functional targets (names, routines, safety information).
• In practice — Specialist, common within its clinical niche.

External aids in rehabilitation

• Analysis — alarms, labelled environments, smartphone prompting. Boring, effective, and consistently outperform internal strategies for impaired memory.
• In practice — Common clinically, and the default recommendation.

9. Oral tradition

Vedic recitation

• Principle — recite the text in permuted orders (jata, ghana) so that each word appears in multiple contexts.
• Analysis — this is an error-detecting code implemented in ritual. The permutations make a corruption in one recitation detectable against the others, which is why the Rigveda transmitted orally with extraordinary fidelity for millennia. The most sophisticated pure-memory technology any culture has produced.
• In practice — Living, within a small and shrinking specialist community.

Oral-formulaic composition

• Principle — build in metrical formulae assembled live rather than recalling a fixed text.
• Example — Parry and Lord's fieldwork on South Slavic singers, applied to Homer.
• Analysis — reframes "memorising the Iliad" as composition in performance from a stock of ready-made units. Closer to chunking than to recitation.
• In practice — Historical in Europe, Living in parts of the Balkans, West Africa and Central Asia.

Prayer beads and counted sequences

• Principle — a physical loop tracks position in a long repeated sequence, freeing attention.
• Example — rosary, mala, misbaha.
• Analysis — an external counter for a task where losing your place is the main failure. Simple and near-universal across unrelated religions.
• In practice — Ubiquitous within religious practice.

10. Chronology and dating

Everything above is about remembering things. This section is about remembering when, which behaves differently enough to be worth separating — and it is the family the history strand lives in.

Temporal landmarks

• Principle — people do not retrieve dates, they reconstruct them by reference to salient events whose position they already know, and interpolate. This is how human dating actually works, not a technique layered on top of it.
• Example — placing an event as "just after the French Revolution" and only then reaching for a year.
• Analysis — the strongest scientific support for anything in this document, and it comes with a sharp caveat. Shum's work and its successors find that spontaneous landmark use improves recall of temporal information and not content. A landmark scheme will tell you when something happened and will not help you remember what happened. Related findings: telescoping (recent events are pulled forward, remote ones pushed back) and the living-in-history effect, where public events become personal landmarks for the generation that lived through them.
• In practice — Ubiquitous, and almost entirely unconscious.

Anchor or benchmark dates

• Principle — memorise a small number of dates cold, then place everything else relative to them. The deliberate, taught version of temporal landmarks.
• Example — 1066, 1492, 1789, 1914 as fixed points, with everything else positioned by "before" and "after".
• Analysis — standard history-teaching practice, and cheap: twenty anchors will place a few hundred events approximately. It degrades gracefully, since losing an anchor costs you a region rather than the whole structure. It gives you ordering and rough distance and never gives you exact years.
• In practice — Common. Taught in most school history curricula, rarely named as a memory technique.

Chronographic charts and the Temple of Time

• Principle — draw time as architecture, so a date becomes a position in a picture you can re-inspect.
• Example — Emma Willard's Temple of Time (1846): a perspective view of a vast hall in which distance from the back wall is time, columns are centuries, and streams of nations run along the floor.
• Analysis — a direct ancestor of the era spine, and by a working teacher solving this exact problem. Willard's stated principles were that facts must be connected to one another to be meaningful, and that only by being made visually manifest are they retained — which is the thesis of this whole project, arrived at in 1846. The nineteenth-century chronographic tradition died with the wall chart; nothing replaced it.
• In practice — Historical. Extinct as live practice, and worth reviving.

Century-per-room palaces

• Principle — one palace room per century, years placed around its walls, floor and ceiling.
• Example — the sixteenth century as your kitchen, 1517 on the third shelf.
• Analysis — the established alternative to the era spine, and the comparison is instructive. Because its unit is a fixed span, it can do arithmetic: two positions tell you how far apart two events are. Because its unit is arbitrary, it carries no causal claim, and a century boundary cuts through the middle of whatever was happening. Practitioners report it is strong for centuries, decades and relative order, and weak for exact years — the same two-stage profile the spine has.
• In practice — Niche, among serious mnemonists.

Historic and Future Dates (memory sport)

• Principle — a competitive discipline: five minutes to bind as many year-and-event pairs as possible, twenty minutes to recall them.
• Example — World Memory Championship competitors using a palace plus a PAO or Major encoding of the year.
• Analysis — included as a warning, not a model. The events are deliberately fictional, so that knowing history confers no advantage. The discipline therefore optimises for arbitrary date-to-event binding under time pressure, which is the exact opposite of exploiting causal structure. Its technique does not transfer to learning real history, and the fact that it looks like it should is the trap.
• In practice — Specialist. Competition only.

What this family says about the era spine

Three things, two supportive and one not:

• Landmark-based dating is how people already work, so the spine is formalising an existing faculty rather than imposing a foreign one. Good.
• Grouping events into narrative blocks rather than isolated dates is standard advice everywhere, and is the same decision as a slot is a bucket, not an event. Good.
• But landmarks improve dating and not content, and most of the spine's cards test content — the organising condition, the causal mechanism. The cards the spine genuinely powers are the span and date cards. Expect it to make you good at placing things and not, by itself, at knowing them.

11. Popular and not working

Included because they occupy the attention that would otherwise go to techniques that do work.

The summary table

What the ratings actually tell you

Three patterns are worth more than any individual entry.

The famous techniques are not the used techniques. The memory palace has the fame; acronyms, rhyme and chunking have the users. If you want a technique that will still be running in a year, bet on the boring one.

The best-evidenced techniques are the ones that feel worst. Retrieval practice, spacing and interleaving all feel less productive during study than re-reading, and all substantially outperform it. Anything that feels efficient while you do it is suspect.

Encoding and scheduling are different problems, and most people only solve one. Memory athletes have extraordinary encoding and mostly do not need retention past the competition. Anki users have excellent scheduling and often terrible encoding, which is why their cards become unanswerable chores. Akasha's bet is that the combination is worth more than either half — which is why every atom carries an image and sits under a scheduler.