The Bamboo Island System
Chinese has thousands of characters. No one memorizes them by brute force. The Bamboo Island System is our way of combining a technique thousands of years old — the memory palace — with Chinese sound components, so each character becomes a picture you can't forget.
The big idea
The human brain is terrible at remembering abstract symbols (like 睡 = "sleep") and extraordinary at remembering places, people, and stories. You remember the layout of youhood home perfectly, decades later. The memory palace turns that superpower toward Chinese.
Here's the key: every Chinese syllable breaks into four ingredients. Bamboo Island gives each ingredient its own cast member. A character is then just a short story combining them.
The four ingredients
| Ingredient | What it is | Becomes… | Why it works |
|---|---|---|---|
| Radicals 部首 | The picture-parts inside a character (hand, water, mouth) | A shape picture — what it looks like | Each radical carries a meaning clue. Recognize 氵 (shuǐ, water) and you can guess that 海 (hǎi, sea), 河 (hé, river), 泪 (lèi, tear), 洗 (xǐ, wash) all involve water — before anyone tells you. |
| Initials 声母 | The starting consonant sound (b, p, m, zh…) | A person — by name | The consonant is brief and easy to confuse (zh vs z, m vs n). A person whose name starts with that sound (m → Mia) makes the sound and the face one memory — every word starting with m begins with Mia: 妈 (mā, mom), 猫 (māo, cat), 米 (mǐ, rice). |
| Finals 韵母 | The ending vowel sound (-an, -ao, -ing…) | A place — a room in a castle | The ending vowel is the longest, loudest part of the syllable. A fixed room in a memory palace (an → Throne Room) gives each vowel a location — the Method of Loci does the remembering for you. E.g. 饭 (fàn, meal) = the f-person walks to the an-room. |
| Tones 声调 | The pitch shape (rising, falling, dipping…) | A family member — by their voice | In Mandarin, tone changes the meaning — mā 妈 (mom), má 麻 (hemp), mǎ 马 (horse), mà 骂 (scold) are four different words. A family member's voice (1st → Mother calling, 4th → Grandpa commanding) makes the pitch part of the story, because real voices you've known your whole life are unforgettable. |
Here are all four libraries — one picture per ingredient:
Take 八 bā — one of the very first HSK 1 characters you meet, meaning "eight". All four of its ingredients become part of one memory:
- A radical — the character's own shape, 八: two strokes spreading apart → an open fan (a picture)
- An initial b → Bella the Brave (a person)
- A final -a → the Apple Tree (a place — red apples in the courtyard)
- A tone (high level, tone 1) → Mother's voice, calling high and steady from afar
The memory: "Bella shakes the Apple Tree so hard that eight giant apples rain down — four to the left, four to the right, landing in the exact splayed shape of 八 like an open fan on the grass — just as Mother calls her name in one long, high, steady note: baaaaa—! (八, bā!)" The weirder the picture, the deeper it sticks. That one absurd image holds the shape (八 = a fan of falling apples), the sound (Bella + Apple Tree = ba), and the tone (Mother's high, held note = tone 1) all at once. The brain holds stories far longer than it holds symbols.
The tones above are the dictionary tones — how each syllable is said in isolation. In real speech, though, a tone can shift depending on the syllable next to it. The commonest case: when two 3rd tones meet, the first is spoken as a 2nd tone. So 你好 is written nǐ hǎo (3rd + 3rd) but actually said ní hǎo (2nd + 3rd).
Bamboo Island shows this directly in the readers: with Pinyin on, a green row gives the dictionary pinyin and a blue row shows the spoken form when it differs — with the changed syllables drawn red. So you learn the rule by seeing it, not just hearing it. (See Stories or Talks.)
Why people, places, and family voices?
This isn't arbitrary. It's the Method of Loci, used since ancient Greece. The trick is to split the slippery thing (a sound) into vivid things (a person, a place, a family member's voice) and weave them into a scene. The finals library is literally built as one coherent castle — Throne Room, Dragon's Cave, Library — so the "places" feel like a real building you walk through. And the tones use family members because the voices of people you've known your whole life — Mother calling, Grandpa commanding — are the most unforgettable sounds there are.
The Bamboo Island System isn't invented from scratch — it stands on two proven foundations. The memory palace (Method of Loci) is ancient. And the idea of assigning people to Chinese sounds comes from the Marilyn Method, a mnemonic system for Mandarin pronunciation described by Serge Gorodish on his blog Country of the Blind in 2012. What Bamboo Island adds is a specific way of splitting each syllable and weaving it into the castle — and an app that builds the whole thing for you automatically. The table below compares the approaches side by side, without judging either.
| Ingredient | Marilyn Method (Gorodish, 2012) | Bamboo Island System |
|---|---|---|
| Initial (consonant) |
A person, but the vowel glide is folded in — so one consonant can have several people, e.g. bi, bu, ju, li. This adds compound initials not found in standard pinyin (there is no initial "li"), growing the table to ~50 slots. | A person, one per real consonant — the standard pinyin initials (b, p, m, f … s, y, w), 23 slots. No invented initials, so it matches the pinyin chart in any textbook. |
| Final (vowel ending) |
12 finals (base endings only — the medial glide was folded into the initial, so finals like -i, -u, -ü, -ia, -uo don't exist here). Each is a place the person occupies. | 36 finals (includes medial compounds — -ia, -uo, -üe, etc. — because the initial stays a plain consonant). Each is a place — a room in a single coherent castle (Apple Tree, Throne Room, …). |
| Tone (pitch contour) |
Position within the place: tone 1 in front, tone 2 just inside the entrance, tone 3 elsewhere inside, tone 4 in the bathroom (neutral = a fifth sub-location). | A family-member voice matching the pitch: Mother high & steady (1), Father rising (2), Grandma dipping (3), Grandpa falling (4), Brother neutral. |
Both methods encode all three parts of a syllable — the difference is the cue used for each. Read the original that started it all: "Mnemonics for Pronouncing Chinese Characters (the Marilyn Method)" by Serge Gorodish →
The standard pinyin chart
When we say Bamboo Island "follows the real pinyin chart," this is the chart — the same one in every Mandarin textbook. Your initials and finals libraries map onto it exactly, one slot per cell:
The 23 initials (声母) — the starting sounds
| Lip / tongue position → | |||||
|---|---|---|---|---|---|
| b | p | m | f | d | t |
| n | l | g | k | h | j |
| q | x | zh | ch | sh | r |
| z | c | s | y | w | ★ (null) |
Each of these is one person in the app. (The ★ is the null initial — syllables that start with no consonant, like 安 ān or 二 èr.)
The 36 finals (韵母) — the ending sounds
Each of these is one place in the app, grouped by type:
| Type | Finals |
|---|---|
| Simple (6) | a · o · e · i · u · ü |
| Compound (9) | ai · ei · ui · ao · ou · iu · ie · üe · er |
| Nasal (9) | an · en · in · un · ün · ang · eng · ing · ong |
| Medial — i glide (5) | ia · iao · ian · iang · iong |
| Medial — u glide (6) | ua · uo · uai · uan · uang · ueng |
| Medial — ü glide (1) | üan |
So a complete syllable is always one initial + one final + one tone — for example b + a + tone 1 = bā (八, "eight"). Every cell above has its own person, place, and voice in the app, so any of the thousands of Mandarin syllables can be built from these pieces.
Each slot is fully customizable. If you picture their own dog when they hear m, they can rename the slot and upload a photo of the dog. Personal images stick hardest of all. (The cast starts with friendly defaults — Cassio the Clever, Albert Einstein, and the rest — but these are just starting points to replace.)
Every person, place, and family voice has a Listen button that speaks a real sample word, so you link the picture to the actual sound, not just a letter. This matters most for the sounds English doesn't have — the retroflex zh/ch/sh and the dental c/z, where the names were deliberately chosen to match the sound, not the spelling.
The cup and the clay
There's one more idea, borrowed from the Tao Te Ching: 有之以为利，无之以为用 — "the material gives a thing its form; the emptiness gives it its use."
Think of a cup. The clay makes the cup — but it's the empty space inside that holds the water. Learning Chinese is the same. You needs both:
- A solid half-hour a day is the clay — focused, uninterrupted, the spine of their learning.
- The empty spaces fill the cup: the five minutes before dinner, the walk to school, the moment before sleep. These reinforce what the half-hour built.
The half-hour gives it form; the gaps give it life. Not in the clay alone, and not in the emptiness alone — but in the cup of the day.
Where to see it in action
Open Remember on the home screen (the 🧩 path). You can browse each library, tap Listen on any slot to hear its sound, and start replacing the default pictures with their own.
