Pinyin for Traditional Chinese Text

One pronunciation, two scripts

Traditional and simplified Chinese are not two languages, and they are not two pronunciations. 學習 and 学习 are both xuéxí; 臺灣 and 台湾 are both Táiwān. The character reforms that produced simplified Chinese changed how characters are written, and in some cases merged several traditional characters into one — but they did not change how the language sounds. For the purpose of annotation this is the good news: pinyin is entirely indifferent to the script. A reading attached to a traditional form is the same reading you would attach to its simplified counterpart, and the tone rules, including third-tone sandhi and the changing tones of 一 and 不, behave identically in both.

What does differ is everything around the characters. Vocabulary choices diverge between mainland China, Taiwan and Hong Kong, character variants are rendered differently by different fonts, and the mergers introduced by simplification create genuine ambiguity when you convert in either direction. An annotated traditional text is therefore not a mechanical transformation of an annotated simplified text. It is the same reading layer over a different set of characters, with a handful of places where you have to make a decision yourself.

Why the order of operations matters

Convert first, annotate second. Annotating before converting produces readings attached to characters that will be replaced, and if the replacement is not one-to-one — which it frequently is not — you end up reconciling two mismatched layers by hand. Converting first also lets you review the readings on the text your reader will actually see, which is where a wrong reading does its damage.

The tool handles both in one pass: the Conversion dropdown tells it which script to produce, and the annotation is laid over the result. That is convenient, but it also means a conversion mistake and a reading mistake arrive together. Correcting a reading is a click on a dashed-underlined character; correcting a conversion mistake means editing the source text. For trap cases, edit the source before you paste it, then set conversion to No conversion — the next section shows where that is worth doing.

Which direction should you choose?

Choose Simplified → Traditional when your source text is written in simplified characters and your reader works in traditional: a mainland news article prepared for a reader in Taiwan, a story you wrote yourself and want to publish in traditional for a heritage-language family, or a passage whose simplified version is easier for you to obtain. Choose Traditional → Simplified when the source is a Taiwanese or Hong Kong book, a scanned page, or a classical text published in traditional form and the reader is studying simplified characters. Choose No conversion whenever the source and the reader already match — introducing a converter into a text that does not need converting can only add errors.

One question deserves an answer before any of this: does the reader use pinyin at all? Pinyin is a Mandarin romanisation system. A family in Hong Kong whose home language is Cantonese may study Chinese through Cantonese romanisation such as Jyutping, in which case Mandarin pinyin is a second, unrelated system. That is not necessarily wrong — plenty of Cantonese-speaking learners study Mandarin pinyin deliberately — but it should be a decision rather than a default. If the reader's pronunciation system is Cantonese, this tool's pinyin will not match the readings they use.

Worked conversion traps

Automatic conversion works at the level of words, and the cases where it fails are the cases where simplification merged distinct characters into one. Three are worth learning by name.

著 and 着

Simplified Chinese keeps 著 for zhù (著名, famous) and uses 着 for the continuing aspect and related readings (zhe, zháo, zhuó). Taiwan convention writes 著 for both, so 看着 becomes 看著 — same reading, zhe, different character. The trap appears in the other direction: converting 著名 from traditional to simplified must leave it as 著名, not 着名, and a converter that has learned "著 becomes 着" from the majority case will get this wrong. The reading is unchanged in both, which makes the error invisible if you only check the pinyin and not the characters.

里 and 裡 / 裏

Simplified 里 covers two unrelated words: 里 meaning inside, and 里 as a unit of distance and a common element in place names. Traditional keeps them apart — 裡 (the usual Taiwan form) or 裏 (the usual Hong Kong and older form) for "inside", and 里 for the unit. So 这里 becomes 這裡 or 這裏, both lǐ, but 公里 must stay 公里: 公裡 is a straightforward error produced by a converter that has learned "里 means inside". This is the trap to check for every time, because distance and measurement words appear constantly in practical texts.

发 / 發 and 髮

Simplified 发 merges two characters with two different tones: 發 (fā, to send, to emit, to develop) and 髮 (fà, hair). Traditional separates them completely, so 发现 becomes 發現 (fā, first tone) and 头发 becomes 頭髮 (fà, fourth tone). A converter working from a word list gets most of these right, but it will slip on compounds and on names. Because the two readings differ in tone rather than only in character, this is the trap where a wrong conversion can also produce a wrong reading — check 髮-related words explicitly.

The same pattern of mergers drives other well-known traps: 干 covering 乾 (gān, dry, as in 乾淨 and 乾杯) and 幹 (gàn, to do, as in 幹活); 后 covering 後 (hòu, after) and 后 kept for the empress in 皇后; 面 covering 麵 (miàn, noodles) and 面 (face); 只 covering 隻 (zhī, the measure word for animals) and 只 (zhǐ, only). None of these are rare, and all of them change meaning or reading when converted carelessly.

Regional vocabulary, not just regional characters

Even with the characters converted correctly, the words may be wrong for your reader. Taiwan and Hong Kong share the traditional script and differ sharply in everyday vocabulary, and both differ from the mainland. A taxi is 計程車 in Taiwan, 的士 in Hong Kong and 出租車 in mainland China. The internet is 網路 in Taiwan, 網絡 in Hong Kong and 网络 on the mainland; software is 軟體 in Taiwan but 軟件 in Hong Kong and on the mainland; the metro is 捷運 in Taipei and 地鐵 in Hong Kong. A bicycle is 腳踏車 in Taiwan, 單車 in Hong Kong, 自行车 on the mainland, and a boxed meal is 便當 in Taiwan and 盒飯 on the mainland.

This matters for annotation because unfamiliar vocabulary defeats comprehension just as thoroughly as unfamiliar characters, and the pinyin above an unfamiliar word does not tell your reader what it means. The practical fix is to pre-edit: if you are adapting a mainland text for a Taiwan reader, replace the handful of mainland-only words with their Taiwan equivalents before you annotate, so the vocabulary matches the characters. There is also a smaller class of genuine reading differences, where the same character is standardly read differently in the two regions: 垃圾 is lājī on the mainland and lèsè in Taiwan's standard, and the conjunction 和 is commonly read hàn in Taiwan where a mainland speaker reads hé. A tool built on mainland pronunciation standards will render these the mainland way, so check them by hand if your reader follows Taiwan norms.

Preparing the annotated traditional text

1. Decide the target script and, with it, the target region's vocabulary. Write the final wording in a plain text file before you open the tool, replacing vocabulary that does not match your reader.
2. Paste the text into the online annotation tool, into the box labelled Paste or type Chinese text here.
3. Set Conversion to Simplified → Traditional or Traditional → Simplified as decided; if you have already pasted the final script, leave it on No conversion so nothing is altered twice.
4. Keep Show tones on, and raise Character size by one or two steps compared with a simplified text — traditional forms carry more strokes and blur together at small sizes.
5. Click Annotate, then read the output against the trap list: 里 versus 裡/裏, 發 versus 髮, 乾 versus 幹, and any proper noun containing 后, 面 or 只.
6. Click any dashed-underlined character to correct a reading, exactly as you would for simplified text.
7. Click Export Word to lay the text out or Export PDF to print, and confirm the characters render in a font with traditional coverage before you print the run.

Print and export notes

Traditional glyphs need a font that contains them. The pinyin font you choose in the tool applies to the pinyin only; the Chinese characters are drawn with whatever East Asian font the viewer, word processor or printer has available. If a character appears as an empty box, or as a variant form you did not expect, that is a font substitution problem rather than a wrong reading. Fonts with reliable traditional coverage include Noto Sans TC and Noto Serif TC, Source Han Serif, PingFang TC on macOS and Microsoft JhengHei on Windows. When you export to Word, set the East Asian font explicitly rather than leaving it on automatic, so the document does not fall back silently on another machine.

Two cosmetic points are worth separating from real errors. Traditional Chinese has variant glyph forms — 為 and 爲, 說 and 説, 裡 and 裏 — and different fonts render each differently; the reading is the same and neither is wrong. Likewise 台 and 臺 are both tái and both current in Taiwan. Do not spend time "fixing" these unless a specific house style requires one form. Before sending a file to a print shop, check that the PDF carries the CJK fonts embedded, or ask the shop to confirm that it has traditional-capable fonts installed — otherwise the safest route is to ask for a single proof page. Page setup, paper and binding are covered in the printing and binding guide.

Where to go next

For the underlying readings themselves, see the polyphonic characters reference and the pinyin chart. If you are converting a story into a traditional edition for a reader, making a pinyin picture book covers the rest of the production chain, while graded reading practice covers passage-level work. All guides are collected in the Tutorials, and every step above uses the free online annotation tool.