Naming
Why “it sounds like your name” is a weak rule for Chinese names
Most Chinese names handed to non-Chinese speakers are chosen by ear. The method is simple, the result is serviceable, and it quietly throws away the only part of a Chinese name that Chinese speakers actually hear.
Ask a Chinese speaker to give an English speaker a Chinese name and they will usually do the same thing: find the characters whose readings land closest to the syllables of the original. Michael becomes 迈克尔 (Mài-kè-ěr). John becomes 约翰 (Yuē-hàn). Anna becomes 安娜 (Ān-nà).
This is not a hack. It is a real convention — the same one newspapers use for foreign heads of state and the same one behind every official list of translated surnames. It is consistent, it is reversible, and it is what most people mean when they say a Chinese name “sounds like you”.
It is also, as a way of choosing a personal name, close to the least informative option available.
What transliteration keeps, and what it drops
A Chinese personal name normally does two jobs at once. It sounds like something, and it means something. The characters are selected so that the meaning contributes — a wish, an image, a quality the family hopes will hold. Chinese speakers do not treat the two as separable; when they hear a name they hear the meaning whether they want to or not.
Transliteration keeps the sound and lets the meaning fall where it may. 迈克尔 does have character meanings if you look them up — 迈 is “to stride”, 克 is “to overcome”, 尔 is a classical “you” — but nobody selected those characters for those meanings. They were selected because the syllables fit. The result reads, to a Chinese ear, roughly the way “Banana” would read to an English ear if it had been chosen because it fit the shape of a foreign name: the parts are words, but no one chose them for their sense.
That is the first cost, and it is the one people notice: a transliterated name has no answer to “what does it mean?” The honest answer is “it sounds like my name”, which is an answer about phonetics, not about the name.
It also creates collisions you cannot hear
The second cost is worse, because it is invisible until someone laughs.
Mandarin has a small syllable inventory compared with English — on the order of 400 base syllables, or roughly 1,300 once you count tones separately. English has more than 10,000 distinct syllables. A language with that few slots produces a lot of homophones, and Chinese speakers are extremely sensitive to them, because the written language disambiguates what the spoken one does not.
When you pick characters by sound alone, you are rolling dice on which homophone you land on. Two entries from the rule library our scoring engine runs against:
| Name | Sounds like | Category | Level |
|---|---|---|---|
| 杜子腾 | 肚子疼 — “stomach ache” | disease | severe |
| 范统 | 饭桶 — “rice bucket”, a fool | insult | severe |
Both are in the library because they are the classic examples Chinese people use to explain the problem to each other. Neither is exotic. Both are the kind of name that gets chosen by exactly the method described above — syllables that fit, characters that were never checked against the phrase they spell.
The library is not complete, and we do not present it as complete. It holds 621 patterns: 222 severe, covering insults, disease terms, vulgarity, body parts and excretion; and 399 weaker ones that depend on context. Matching runs on the flattened full name with the surname included, which is the only way either of the examples above gets caught — in both of them the surname is doing half the work.
| Category | Entries |
|---|---|
| Insult | 66 |
| Disease | 45 |
| Vulgar | 42 |
| Body part | 31 |
| Excretion | 28 |
| Negative word | 10 |
Sound is still tested. It is just tested last.
None of this argues for ignoring sound. It argues for not letting sound be the only input.
The order matters. If you start from the sound, meaning is decoration you add afterwards, and the search space is tiny: there are only so many characters that read “mi”, “ke”, “er”. If you start from meaning, the sound becomes a filter — one of several checks applied to candidates that already have something to say.
In the studio, the sequence is: cast the birth chart, derive which elements the chart is short of, raise characters carrying those elements, filter out characters feeding the unfavourable ones, then screen every surviving candidate against the homophone library and a 1.1-million-name corpus, and finally score. The sound check still runs on every name. It just runs on a set that was built for a reason other than its sound.
A transliterated name answers the question “what does it sound like”. A chosen name answers “what does it mean” — and can still answer the first question at the same time.
The honest limit
A meaning-first name is harder to explain than a transliteration, and that is a real trade-off, not a marketing line. A Chinese speaker who meets 迈克尔 immediately knows what happened: someone transliterated Michael. A Chinese speaker who meets a name built from a chart and a character library will ask you where it came from — and you will need an answer.
The point of a name you can explain is not that the explanation impresses anyone. It is that the name survives the question. Most transliterated names do not.
Six names from your surname, free, each with the character glosses and the reason for every point of the score.
Generate a name →