Skip to content
TabShedKitsMy desk

Simplified vs Traditional Chinese: Converting Text for Taiwan and Hong Kong Without Embarrassing Mistakes

Updated 5 min readBy

What changes when you convert Chinese between Simplified and Traditional, why word choice matters as much as characters for Taiwan, and what to check before you publish converted text.

Simplified ⇄ TraditionalConvert Chinese text, Taiwan and Hong Kong
Open Simplified ⇄ Traditional
On this page
  1. Characters: one-to-many
  2. Words: where Taiwan differs from the mainland
  3. Which option to choose
  4. A worked example
  5. A checklist before you publish
  6. What the tool does not convert
  7. Speed, size and privacy
  8. After converting

Converting Chinese between Simplified and Traditional looks like a character swap, and for a short phrase it often is. For a website, a menu or a product description it is not: some Simplified characters have several Traditional forms, and a text written for mainland readers uses words that Taiwan and Hong Kong readers do not. This guide explains what changes, what the options in the Simplified to Traditional converter mean and what you should still check by eye.

Characters: one-to-many

When Simplified characters were introduced, several Traditional characters were merged into one. Going back is the hard direction, because the converter has to guess which Traditional character was meant. Some examples:

SimplifiedTraditional formsExample
发發 (to send, to develop), 髮 (hair)发展 → 發展, 头发 → 頭髮
后後 (behind, after), 后 (queen)后面 → 後面, 皇后 stays 皇后
面面 (face, side), 麵 (noodles, flour)面子 stays 面子, 面条 → 麵條
里裡 (inside), 里 (a unit of distance, a neighbourhood)里面 → 裡面, 公里 stays 公里
干幹, 乾, 干干部 → 幹部, 干净 → 乾淨

A converter that works character by character gets many of these wrong. OpenCC, the open conversion project behind this tool, matches whole words first and single characters second, which resolves most of them correctly. It is not perfect: a person's name, a line of classical poetry or a rare word may come out wrong, so read the result.

Words: where Taiwan differs from the mainland

Even correct characters can read oddly. In Taiwan, a computer program is 軟體, not 軟件; memory is 記憶體, not 內存; a mouse is 滑鼠, not 鼠標; a taxi is 計程車, not 出租車. A text that only had its characters changed is clear to a Taiwanese reader but marks itself as written elsewhere, which matters for a shop, a school or a government form. The option Simplified → Traditional (Taiwan, with local words) also converts this vocabulary. Use characters only when you want to keep the original words, for example when quoting a mainland source.

Which option to choose

Your reader isChoose
In Taiwan, general publicSimplified → Traditional (Taiwan, with local words)
In Taiwan, quoting a mainland source as it isSimplified → Traditional (Taiwan, characters only)
In Hong Kong or MacauSimplified → Traditional (Hong Kong)
Reading a Taiwan page, wanting mainland styleTraditional (Taiwan) → Simplified, with mainland words
Only the script mattersTraditional (Taiwan) → Simplified (characters)

For Hong Kong, the characters follow the Hong Kong standard shape for a handful of characters, but the written language of newspapers and forms is close to Taiwan usage. The Hong Kong option does not turn standard written Chinese into Cantonese: words and spellings used only in spoken Cantonese are not added.

A worked example

Take the line 我在台湾买了一台新电脑,软件和内存都很好。 and convert it with the Taiwan option that includes local words. You get 我在臺灣買了一臺新電腦,軟體和記憶體都很好。 Three kinds of change happened. The characters were changed (买 became 買, 电脑 became 電腦). The Taiwan standard writes 台 as 臺 (臺灣, 一臺), so that changed too. And the words 软件 and 内存 became 軟體 and 記憶體. A second line, 他的头发很长,后面的发展很快。, becomes 他的頭髮很長,後面的發展很快。, with two different Traditional characters for the same 发.

Now run the same line with the other two options. Characters only for Taiwan gives 軟件和內存: the characters are Traditional, but the words are still the mainland ones, which a Taiwanese reader understands and recognises as written elsewhere. The Hong Kong option keeps 台灣 and 一台, because Hong Kong does not use the 臺 form, and it chooses different shapes for a few other characters: 里面 becomes 裡面 for Taiwan but 裏面 for Hong Kong, and 面条 is 麵條 in Taiwan but 麪條 in Hong Kong. These small differences are the reason to choose the variant by your reader, not to take one Traditional text for all.

A checklist before you publish

  1. Read the names. Personal names, brand names and place names are the commonest errors. If a name has a fixed Traditional form, search the result for it.
  2. Search for single risky characters such as 干, 发, 后, 里, 面, 台 and 制 in the converted text and check each.
  3. Check numbers, dates and units. Units such as 公斤 and 公里 are the same in both scripts, but date and number formats in the text (年月日, thousands separators) may follow the old audience.
  4. Keep quotations. A quoted line from a book or a law should come from a source in the right script, not from a converter.
  5. Convert back and compare. Switch the direction and paste the result in. It should return to something very close to the original; a surprising difference marks a word to look at.

What the tool does not convert

  • Japanese shinjitai and Korean hanja forms: they are different systems.
  • Text inside images, PDFs or Word files. Copy the text out and paste it in; for a .txt file, drop it onto the box (UTF-8, GBK and Big5 are tried).
  • Cantonese-only characters used in Hong Kong writing.

Speed, size and privacy

The conversion runs in your browser. The tables are downloaded the first time you choose a direction: about 1.1 MB for Simplified to Traditional and about 110 KB for the other direction. The browser keeps them, so the tool also works offline later. Long text is converted in pieces, so a file of tens of thousands of characters does not freeze the page. Nothing you paste is uploaded.

After converting

If you are producing a study text, add readings with the Chinese to Pinyin converter, which reads both scripts. Converted pages should also carry the right language code (zh-Hant-TW, zh-Hant-HK or zh-Hans) in the HTML, so that browsers and search engines choose the matching font and show the page to the right readers.