Chinese does not use a word separator. How a novice learner can detect the word's boundaries?

  • You can check for verbs, adjectives, adverbs, nouns, conjunctions etc. that you know, there are some that are very common (eg 去 [v],好 [adj],很 [adv],昨天 [n].) Once you know your sentence structure, you'll know what comes after that verb should be a noun, even if you don't know what that word is. Unfortunately, there's not much else you can do. Hint: words are more often one to two characters. Not as often three or four.
    – Ming
    Jan 29, 2016 at 6:32
  • Classic Chinese didn't even use punctuations.
    – user4072
    Jan 29, 2016 at 8:23
  • 1
    You need to build vocabulary and understand the context.
    – imrek
    Jan 29, 2016 at 11:50
  • @Ming Is space used as a word separator at least to teach the languages to beginner? I'm although curious why they do not use word separator? Is it really useless for native speaker? Jan 29, 2016 at 15:23
  • @PHPst In Chinese script space is not commonly used, but if your textbook has Pinyin transliterations parallel to the original text, then the Pinyin subs are separated along word boundaries. A possible exception are children's books.
    – imrek
    Jan 29, 2016 at 16:54

2 Answers 2


You can only detect word boundaries by the context and your vocabulary. If you don't know words many enough or the situation at hand much enough, there isn't much other way to ensure you understand the things correctly.

The reason that Chinese doesn't have word boundaries is that there are "character boundaries". In English, the basic unit of the lexicon, called a lexeme, is one word separated by spaces or punctuation. In Chinese, the lexeme is one Chinese character, so rather than having separated words, it is easy to distinguish different characters in a sentence. However, the hard part is that many Chinese words are actually multi-character words, much like multi-word terms / phrasal verbs in English. If English were to use multi-word terms as frequently as does Chinese use multi-character words, new English learners will find it equally hard to learn English.

  • English has bazillions of compound items that take on meanings of their own, and, yes, that makes English (and many? all? other languages) difficult to learn. To quote from Lexical Limits? on LanguageLog: "what sorts of things [are words]? Once you have dog, do you treat dogs as a separate word? Presumably not. But how about doggy, dogging, dogged? If you have food as well as dog, is dog food (or dogfood) a separate word? Or hotdog, the food or the skiing style? What about idiomatic phrases like dog's breakfast?"
    – flow
    Jan 30, 2016 at 16:24
  • @flow Yes English also has many compound items, but if you compare, Chinese uses them much more frequently, that is, aside affixes. However I agree that English still suffers from ambiguity, though IMHO the "compound item" factor is less important than in Chinese and the order of modifiers is a more common factor.
    – busukxuan
    Jan 30, 2016 at 17:27
  • It seems wrong to me to claim that lexemes are 1 Chinese character. It may have been true in some distant past (and even then, it was not entirely true anyway, there's plenty of polysyllabic words found in Classical Chinese texts), the average lexeme length in Mandarin is far above 1 syllable. Perhaps you meant morpheme, that would be closer to reality.
    – vermillon
    Jan 31, 2016 at 12:38
  • @vermillon I didn't mean morpheme though. I understand that most characters can be morphemes themselves, yet composing words with more than one character can give new meanings and make the whole thing 1 morpheme, like how English idioms can be considered 1 morpheme. Also I don't see any problem in using most characters alone, as in 1-character words. It might make things harder to understand and probably very ambiguous, but the sentence should still be valid. That said, can you please suggest a term closer to what I wanted to say? I guess I'll really need to learn that.
    – busukxuan
    Jan 31, 2016 at 13:46
  • @busukxuan: A "morpheme" is the smallest unit of meaning. Combining 2 morphemes does not make a new morpheme, by definition. It can make a new word, a radical, whatever the case happens to be. If you take the word "werewolf", it has 2 morphemes, its meaning is different from both morphemes, it's a new word with a specific meaning, but it's not a single-morpheme word.
    – vermillon
    Feb 1, 2016 at 12:05

Today over at LanguageLog, I came across this sentence:


Inadvertently, when scanning that phrase I ran down a gardenpath (yes, that's what linguists call it: a fruitless attempt that turns out to be the wrong choice, so you have to back up). Turns out it's not 開發 中國(家) *"develop China's (state)??", but, rather, 開發中 國家 "in-development states" (i.e. "developing countries").

Such things will happen and happen again to the reader, and if you care to know more, head over to said site and google it for 'gardenpaths', there's lots of examples for a number of languages. Just the other day I realized German "Maisturm" is ambiguous: is it "Mais-Turm" (maize/corn tower) or "Mai-Sturm" (May tempest)? Probably the latter in most cases, but only context can tell.

I remember some of the beginner's textbooks for Chinese using spaces between words, so I guess you'll still be able to find some. As for why spaces are not normally used in Chinese and Japanese (but in Korean), one can only guess that people find them too unsightly, not really needed and maybe too big a change from what they're used to.

Also, while extra spaces might benefit the reader, they do put a burden on the writer (try to add them while staying consistent—it's not easy!).

  • Korean is another case, Korean lexemes can span many blocks, such as Korean native word 다르다 meaning different, whereas Chinese loanword (异, trad:異) of the same meaning is indeed a single-block lexeme. Korean uses spaces as lexeme separators, and Chinese uses characters' block structures as lexeme separators.
    – busukxuan
    Jan 30, 2016 at 11:10
  • 1
    When you say 'lexeme', I take it that you mean 'morpheme'. Yi4 異 is not a 'lexeme'/'word' b/c it's not used by itself (in Ch. at least), only in compounds (異讀, 異國...). There are lots of polysyllabic morphemes in Ch., eg 蝴蝶, 葡萄, 蜘蛛, 蟑螂.... Ch. does have many monosyllabic morphemes & words, but a similar fact is true of English and, I think, Korean. It's hard to tell what a 'word' is, but we can say that in 開發中國家 there are at least 開發, 開發中, 開發中國, 中國, 國家 as candidates for recurring compounds with meanings that can't 100% be predicted from the meanings of their components.
    – flow
    Jan 30, 2016 at 16:18
  • The Chinese 異 can be used by itself, only that it's uncommon nowadays. In your examples, 蝶、蛛、螂 can all stand alone. By the way, I was originally talking about the Hanja 異 rather than the Chinese 異. In Korean spacing is used to separate lexemes(or wrong term again?), not morphemes. Back to the example I gave, 다르다 actually contains 2 morphemes: the root 다르- and suffix -다 that implies either the propositive or the declarative voice, but there is no spacing between them. That's why I said spacing is used to separate lexemes.
    – busukxuan
    Jan 30, 2016 at 17:21
  • "In Korean spacing is used to separate lexemes(or wrong term again?)" =>yes, wrong. A lexeme would be, roughly speaking, the set (meaning you can't call something "a lexeme") of forms you would group under 1 dictionary entry: {kill, kills, killing, killed,...} is 1 lexeme, whose lemma (the form you'll find as the dictionary entry) is "kill". It happens to also be a single morpheme word, but that's a coincidence in this case.
    – vermillon
    Jan 31, 2016 at 12:47
  • @vermillon I know a lexeme is a set of forms, and please excuse my wrong usage of terminology, does this mean I have to say "to separate elements of lexemes"? It sounds weird. Is there a better way to say that?
    – busukxuan
    Jan 31, 2016 at 13:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.