The spelling system is the small part

You can sound out Japanese once you have the kana, and you can type Mandarin once you have pinyin. Neither of those is reading the characters. The spelling system is small on purpose. The character list is the rest of the job.

“How to read Japanese” and “how to spell Chinese” land on the small system first, which is right. They go wrong when the small system becomes the whole plan.

Two gaps, same shape

Japanese: 46 hiragana in the basic school set, against 2,136 jōyō kanji. Mandarin: 21 pinyin initials and a 35-cell finals table, against 3,500 level-1 characters in the 2013 table. The chart is on a log scale so 21 and 46 stay visible next to the thousands. Do not read the bar lengths as “kanji are 46 times harder than hiragana”. Different units. The only point is that the second number does not collapse into the first.

[[chart:ld-read-bridge]]

Switch between Japanese and Mandarin. The picture is the same shape. Kana and pinyin get you the sound of a word you have met. They do not store the character. Furigana over a kanji, and pinyin under a hanzi, are the same kind of help: a spelling left in the line so you can keep going. Take it away and see what is left.

A line you can actually practise

Japanese: 私は水を飲む. You can read the hiragana は and を on day two. 私, 水, and 飲 are the characters. A learner text prints みず over 水. Cover the furigana with a finger. If you cannot say mizu, that is today’s card, not a reason to restart the kana chart.

Mandarin: 我喝水. Pinyin gets you wǒ hē shuǐ if someone prints it. Cover the pinyin. Write 水 from “water” and say shuǐ from 水. The initials chart, including why y and w are not initials, is Chinese has no alphabet. The kana count is what kanji is not.

Input methods feel like the large number is finished

Typing mizu or shui and tapping the character is lookup, which is allowed, and it is the end of the dictionary step. It is not production. The character that comes from a menu you recognised is the five-minute win in the retrieval study. The week-later win is the character you can write when the menu is gone.

Why this method works

Furigana and pinyin are the same kind of help as an open book in the Swahili study: useful the first time, harmful as a permanent crib. Covering them is the test arm. Leaving them in the line because “I can read it with the help” is the arm that felt fine in the session and collapsed a week later. The log chart is only a reminder that the crib is the small number and the test is the large one.

Run this session

Twenty minutes. One language. Four words you can already sound out with help.

1. 0–5 min. Write the help in full: furigana over the kanji, or pinyin under the hanzi. Read the line once. This is the only open-book pass.
2. 5–15 min. Cover the help with a finger or a card. Say the word. Uncover only on a miss, then cover again immediately and say it a second time. Hits stay covered.
3. 15–20 min. Blank paper. English or kana prompt only. Write the character. An input method is a miss. Those misses are tomorrow’s first four minutes, not a new list.

Common traps

• Finishing kana or pinyin and calling yourself literate. Fix: 46 and 21 are the bridge. 2,136 and 3,500 are the lists.
• Reading the log chart as “kanji are 46 times harder”. Fix: the scale is only there so the small count stays visible. The units differ.
• Typing the reading and tapping the character you half recognise. Fix: that is lookup. Lookup ends when the card exists. See the dictionary session.

Keep the help covered on the second pass even when the first pass was perfect. The papers’ losing schedule was the one that retired a pair after a single success. One clean reading with furigana still showing does not count as that success.

Build the card from the word you just had to uncover. Rate it in Study Lab. Sit JLPT or HSK when a full paper is the point. The paper will not print furigana over every stem, and it will not leave the pinyin under the passage.