A [class.] zoo

« previous post | next post »

In English, if we want to say something about a place where a lot of different kinds of animals are kept for viewing by the public, we just refer to it as "a zoo".  Ditto for other quantifiable or specifiable nouns.  But in Chinese, you usually have to put a measure word [m.w.] or classifier [class.] between the quantifier or demonstrative and the noun.  (In this post, I won't go into the subtle distinction between measure word and classifier.)

yī + class. + dòngwùyuán 动物园 ("zoo")

But what classifier / measure word should we use with dòngwùyuán 动物园 ("zoo")?  Here are the google hits for various possibilities, followed by the meaning of the measure word / classifier, which should not be translated when you render a text into English, otherwise you will be producing Chinglish):

yī chǎng dòngwùyuán 一场动物园 553,000 ("broad, flat, open space / place")
yī suǒ dòngwùyuán 一所动物园 231,000 ("place; spot; location")
yī gè dòngwùyuán 一个动物园 158,000 ("piece")
yī jiā dòngwùyuán 一家动物园 115,000 ("home")
yī zuò dòngwùyuán 一座动物园 24,200 ("seat; base; pedestal")

Taking into account that some of these are false hits (e.g., when the measure word is separated from the noun by punctuation), this gives us a rough idea of people's preferences in writing.  My impression, however, is that in casual speech people would generally resort to the all purpose yī gè 一个, and that to use one of the others would sound pretentious.

A good dictionary, such as ABC, will specify which m.w. / class. goes with which noun (ABC gives zuò 座 for dòngwùyuán 动物园), and sticklers will try to use the right one.  It seems to me, however, that more and more people tend to use the general or default gè 个 / 個 for many nouns that in the past would have merited their own special classifier.  When I started learning Mandarin, I prided myself in knowing dozens of classifiers and measure words and which nouns to use them with.  Forty or more years ago back in Taiwan, people would praise me for using the correct m.w. / class.  But as the decades passed, and especially when I went to the mainland, even my friends started to give me amused looks when I pulled out an obscure m.w. / class. in speech, as though I were trying too hard or showing off.  The effect was similar to that elicited by the overuse of chéngyǔ 成语 ("set phrases", aka "idioms").  Forty-five years ago when I peppered my speech with them I was praised, but now if I use too many, I will receive bemused stares as though I were being peculiar or pedantic.

Despite what the google numbers tell us, I'd wager that, if you say yī zuò dòngwùyuán 一座动物园 on the mainland, people will think you're a bit weird, though in Taiwan people might well be impressed by your erudition.  Since I haven't used zuò 座 very much in recent years, I wouldn't want to attempt to list all the nouns with which it can be used, though I do remember that I was taught to use it for banks, thus yīzuò yínháng 一座银行 ("a bank"), which garners 43,800 ghits.  I wonder if that's what they teach students to say on the mainland.

APPENDIX (for specialists and advanced learners)

Context makes a great difference.  If we insert adjectives or adjectival phrases between the classifier and dòngwùyuán 动物园, the search results come out quite differently:

m.w. + xīn de 新的 ("new") + dòngwùyuán 动物园 ("zoo")

一个新的动物园 56,800
一座新的动物园 1,520
一所新的动物园 9
一家新的动物园 6
一场新的动物园 no results found

m.w. + hěn dà de 很大的 ("big") + dòngwùyuán 动物园 ("zoo")

一个很大的动物园 112,000
一座很大的动物园 1,170
一所很大的动物园 1
一家很大的动物园 6
一场很大的动物园 no results found

m.w. + shìjiè jí de 世界级的 ("world class") + dòngwùyuán 动物园 ("zoo")

一个世界级的动物园 2,650
一座世界级的动物园 964
一所世界级的动物园 no results found
一家世界级的动物园 1
一场世界级的动物园 no results found

We also need to take into account the fact that the search results include many strings in which dòngwùyuán 动物园 is modifying another noun (or noun phrase) that happens to use, for example, chǎng 场 as its measure word. Here are some of the strings returned in the Google search results:

yī chǎng dòngwùyuán èmèng 一场动物园噩梦 ("a nightmare about a zoo")
yī chǎng dòngwùyuán sàichē bǐsài 一场动物园赛车比赛 ("a zoo race")
yī chǎng dòngwùyuán fēngbào 一场动物园风暴 ("a zoo storm")
yī chǎng dòngwùyuán dàzhàn 一场动物园大战 ("a zoo war")
yī chǎng dòngwùyuán biǎoyǎn 一场动物园表演 ("a zoo show")

In these cases, we need to be thinking about why chǎng 场 was chosen as the measure word for "nightmare", "race", "storm", "war", and "show", not for "zoo".

[Hat tip Rachel Kronick; thanks to Mark Liberman, Gene Buckley, Matthew Trueman, Bob Sanders, Zheng-sheng Zhang, and John Rohsenow]


  1. Richard W said,

    January 18, 2015 @ 11:54 pm

    CC-CEDICT suggests 個 (个):
    動物園 [dòngwùyuán] zoo; CL:個[gè]

    In Taiwan Panorama articles, I can find one example of 座 and one of 個.

    His experience with the frogs has given Hsu a deeper understanding of the meaning of "ecology." He says, "If we only reintroduce a bunch of frogs we'll just be building a zoo. What would be the point of that?" Inspired by the "ecological island" concept, in 2002 he founded and became the chairman of LFRA and now devotes all his energy to the ecological restoration of Lichia Forest Road.

    In the open-air museum of Parque La Venta, gigantic trees tower into the sky everywhere. As soon as we go in through the gate, in a sandy enclosure we see several monkeys hanging playfully from exercise bars, for this tropical park doubles as a miniature zoo. It is also one of the places where Mexico's earliest civilization, the Olmec, once flourished.

  2. Mark S. said,

    January 19, 2015 @ 12:16 am

    If one converts the Hanzi to traditional characters and restricts the searches to .tw domains, the results are largely reversed.

    yī chǎng dòngwùyuán 一場動物園 553,000 vs. 165
    yī suǒ dòngwùyuán 一所動物園 231,000 vs. 6,130
    yī gè dòngwùyuán 一個動物園 158,000 vs. 8,800
    yī jiā dòngwùyuán 一家動物園 115,000 vs. 27,600
    yī zuò dòngwùyuán 一座動物園 24,200 vs. 9,570

    Modifying those with xīn de / 新的 never produces more than 6 hits (for yi ge), so I'm omitting those results.

    Restricting the results to .cn sites yields this:
    一场动物园 21,400
    一所动物园 35
    一个动物园 33,200
    一家动物园 26,300
    一座动物园 5,320

  3. Mark Mandel said,

    January 19, 2015 @ 3:15 am

    How do these pattern over time?

  4. Bruce said,

    January 19, 2015 @ 4:58 am

    I remember seeing 座 or a character very much like it in Hong Kong to designate which which numbered building among many in a multi-block housing estate. viz

    香港深井麗都花園3座11樓平4A ==> Hong Kong Sham Tseng Lido Garden Block 3, 11/F Flat 4A

    That's a real address not far from my own once upon a time.

  5. maidhc said,

    January 19, 2015 @ 5:17 am

    So yī by itself can mean "one", thus the pidgin "one piece shirt" would be a literal translation of the Chinese. However "a shirt" would be a more graceful translation. Interesting.

    That's useful information as I have been trying to read simple Chinese phrases using just Pleco and Google Translate. I've noticed some amount of redundancy but until now I didn't realize there is some structure to it. (I know this is a stupid way to learn Chinese, but for me it's more of a pastime.)

    There's a sort of vestigial version of the same structure in Indo-European languages, like in English we say "twenty head of cattle". Head is used in the same way in Irish only more, so for example if you wanted to say "I already have one [of something]" it would be "I already have a head".

    I think there must be some other examples but I can't retrieve them.

  6. Richard W said,

    January 19, 2015 @ 5:44 am

    pairs of scissors
    sticks of dynamite
    sheets of paper
    bottles of milk
    items of laundry
    pieces of candy
    cups of coffee

  7. anonymous coward said,

    January 19, 2015 @ 10:06 am

    > There's a sort of vestigial version of the same structure in Indo-European languages

    Well, no, not vestigial. Both Chinese and IE languages make a distinction between countable and uncountable nouns. For example, in English 'five waters' or 'three rices' is ungrammatical; you'd need to say 'five cups of water' or 'three grains of rice' instead.

    In Chinese, only animate nouns (mostly pronouns), and, I think, nouns related to time, are countable. Objects and animals are uncountable.

  8. Matt said,

    January 19, 2015 @ 10:24 am

    It would be interesting to see how trends towards all-purpose classifiers correlate with the commonness of the noun involved. My hypothesis would be that it'd work like irregular verbs: nouns that appear often with many different numbers will keep their classifiers forever, however uncommon those classifiers themselves may be outside that context, while unique classifiers associated with less frequent nouns will gradually erode away.

  9. J. W. Brewer said,

    January 19, 2015 @ 11:58 am

    With the exception of "pairs of scissors," which is driven by the lack of ordinary singular/plural distinction for scissors (and "trousers" and "eyeglasses" and other such examples), the other English examples given above deal with mass nouns. ("Cattle" was, I believe, at least anciently a mass noun, although it has drifted somewhat.) Further examples would be "grains of sand" or "drops of water." I wonder if in languages like Chinese where the use of classifiers is obligatory or near-obligatory (and there is also, as I understand it, no morphological marker for singular v. plural on nouns) the whole mass-noun/count-noun distinction simply isn't present? (I do know that even among IE languages, whether a particular noun is count or mass is not always consistent cross-linguistically, e.g. there are lots of things that are mass nouns in Welsh but not in English.)

  10. atif said,

    January 19, 2015 @ 5:51 pm

    It seems like 一間動物園 (traditional chinese) has the most hits on Google at 1,200,000.

  11. Victor Mair said,

    January 19, 2015 @ 6:56 pm


    You have to search for it in quotation marks, thus "一間動物園", not just 一間動物園. Otherwise, you'll pick up any pages that have those five characters, but not necessarily in that collocation. If you search for "一間動物園" in quotation marks, which is the string you're looking for, you'll get 12,100 ghits.

  12. Victor Mair said,

    January 19, 2015 @ 7:38 pm

    @anonymous coward

    "In Chinese, only animate nouns (mostly pronouns), and, I think, nouns related to time, are countable. Objects and animals are uncountable."

    I'm not sure I understand what you're talking about.


    yītóu niú 一头牛 ("a / one cow") (m.w. "head" — like English)

    yī pǐ mǎ 一匹马 ("a / one horse") (m.w. "one of a pair" — because chariot horses usually came in pairs)


    yī zhāng zhuōzi 一张桌子 ("a / one table") (m.w. "flat surface")

    (all of the following mean "a / one house") — I'm not trying to be exhaustive with the measure words

    yī jiàn fángzi 一间房子 (m.w. "space between") 552,000

    yī dòng fángzi 一栋房子 (m.w. "roof beam") 508,000 — this is what I was taught to say in first-year Mandarin

    yī chuáng fángzi 一幢房子 (m.w. "pennant; streamer; tent") 363,000

    yīgè fángzi 一个房子 (m.w. "piece") 521,000

    yīzuò fángzi 一座房子 (m.w. "seat; pedestal; stand") 394,000

    yī suǒ fángzi 一所房子 (m.w. "place; spot; location") 378,000

    It's remarkable how evenly balanced these six different measure words for "house" are.

  13. Matt said,

    January 19, 2015 @ 7:57 pm

    The argument I think is that "yītóu niú" and "yī pǐ mǎ" should be understood as analogous to something like "one head of cow" and "one half-pair of horse", i.e. that the classifier system is really just the result of virtually all nouns in Chinese being the equivalent of uncountable/mass nouns in English like "water" and "rice".

  14. Victor Mair said,

    January 19, 2015 @ 8:17 pm


    That's what the Chad Hansen argued here and elsewhere:

    "Mass nouns and 'a white horse is not a horse'"

    Philosophy East and West 26 (2):189-209 (1976)


    The most famous paradox in chinese philosophy, Kung-Sun lung's "white horse not horse" has been taken as evidence of platonism, Aristotelian essentialism, Class logic, Etc., In ancient chinese thought. I argue that a nominalistic interpretation utilizing the notion of "stuffs" (mass objects) is a more plausible explanation of the dialogue. It is more coherent internally, More consistent with kung-Sun lung's other dialogues, And the tradition of chinese thought which is usually regarded as nominalistic. The interpretation is also strongly suggested by striking parallels between all chinese classificatory nouns and english mass nouns


    Time to call in the philosophers.

  15. tsts said,

    January 19, 2015 @ 10:25 pm

    @anonymous cowards:

    Actually, I would argue that "five waters" is grammatical. Or at least very widely used to the point of being acceptable. Just go to any restaurant with a group of people, and the server will happily bring you "three beers, five water(s), and a coke".

    Now, when ordering sides, "two rice" might be better than "two rices", unless they are different types of rice.

  16. Bryan Van Norden said,

    January 19, 2015 @ 11:30 pm

    Hansen's hypothesis was only intended to apply to Classical Chinese. In any case, his work has been discredited. See, for example, Harbsmeier, Christoph, "Marginalia Sino-logica", in Allinson, Robert E. ed. (1989):Understanding the Chinese Mind: The Philosophical Roots (Oxford University Press), pp. 125-166; and Harbsmeier, "The Mass Noun Hypothesis and the Part-Whole Analysis of the White Horse Dialogue", in Henry Rosemont, Jr. ed. (1991): Chinese Texts and Philosophical Contexts (La Salle, Illinois: Open Court), pp. 49-66.

  17. DMT said,

    January 20, 2015 @ 12:50 am

    The differences between English and Chinese ways of handling measure words can be a problem for learners. It took a long time for me to break the habit of saying "yi shuang yanjing" 一雙眼鏡 for "a pair of glasses" and "yi shuang kuzi" 一雙褲子 for "a pair of trousers" (but oddly enough, I don't think I was ever tempted to say "yi shuang jiandao" 一雙剪刀 for "a pair of scissors). Going in the other direction, many Chinese speakers have difficulty remembering which English nouns are countable and which are uncountable. (A recent personal anecdote: my three-year old daughter, whose English is otherwise pretty good, recently came up with the sentence "I only wear one piece of clothes, because I am afraid of hot" – a direct translation of 我只穿一件衣服,因為我怕熱.)

  18. DMT said,

    January 20, 2015 @ 1:00 am

    @anonymous coward, tsts: "rice" is one of the words that often trips up Chinese learners of English – there are obviously lots of small objects in the bowl, so why don't we say "a bowl of rices" in the same way we say "a bowl of peas"? (Of course, the English word "pea" itself was derived as a back-formation from the uncountable "pease", which could be reinterpreted as the plural of a countable noun because of its final /z/.)

  19. Victor Mair said,

    January 20, 2015 @ 7:10 am

    From a historian of Chinese thought:

    I think that the confusion among the commentators might arise from the skipping over by others of what is implicit in your comments (esp. that in response to anonymous coward), which is that Chinese nouns are neither necessarily mass nor particulate — they are both. That is, they inherently embrace the potential of both mass and particularity, and, thus, when one wishes to specify to what specifically s/he intends to denote with the use of this nominal signifier (the noun), s/he must delineate which particular member(s) of that noun class to which s/he intends to refer. If one just says ma 馬, for instance, this could refer to either the idea of a horse, a thousand horses grazing in a large meadow, or one or more particularly denoted horses. It could also refer to a horse depicted in a painting, which leads us back to the idea of (as a representation, but not truly) a horse.

    The measure word, I think, is perhaps less a "measure" term (for the specifying numerator provides the actual "measure") than it is an additional signifier that helps to identify to which, among tens of characters pronounced ma, the speaker intends to refer with clarity. The use of such "signifier terms" may derive from a time in which the often monosyllabically oriented classical written language was spoken in formal situations and was, as you indicated of your own experiences using the terms in Taiwan 40+ years ago, a means of achieving referential clarity (as well as demonstrating erudition or acceptable educated usage, since the use of such terminology would have been necessary among those who spoke with one another the formal classical written language in order to be understood and to understand without written prompts).

    There are, of course, true unit-measure terms in Chinese, as well, such as sheng 升 as a unit of measure of dry grain. But even here, I would argue, the term serves a double function as a signifier, helping to identify, for instance, which character mi from among tens (and I'm of course referring with this example to 米, harvested rice) to which the speaker refers.

  20. Matt said,

    January 20, 2015 @ 9:19 am

    Bryan, I don't suppose you could summarize the arguments which were used to disprove Hansen's hypothesis? I couldn't find either of the chapters you mentioned online and my local libraries (here in Japan) aren't very good at ILL-ing English-language materials, but having read Hansen's article I'd be interested in the opposing view.

  21. Victor Mair said,

    January 20, 2015 @ 10:14 am


    And yet it seems so natural in Chinese to say yīshuāng kuàizi 一双筷子 ("a pair of chopsticks") and yīshuāng xiùhuāxié 一双绣花鞋 ("a pair of embroidered shoes"). From the Chinese point of view, the chopsticks and shoes are not joined, so it's all right to use shuāng 双 / 雙 ("double; twofold; twin; pair"), whereas scissors, eyeglasses, and pants are joined, so they all have their special measure words:

    yī bǎ jiǎndāo 一把剪刀 ("a pair of scissors") (m.w. "handle")

    yī tiáo kùzi 一条裤子 ("a pair of pants") (m.w. "strip; slip; item; long narrow piece")

    yī fù yǎnjìng 一副眼镜 ("a pair of glasses") (m.w. "assistant; deputy; auxiliary; subsidiary") and yī fú yǎnjìng 一幅眼镜 ("a pair of glasses" — receives nearly ten times as many ghits as 一副眼镜) (m.w. "width; breadth; amplitude")

    And then there is duì 对 ("pair; couple"), as in yī duì fūfù 一对夫婦 ("a couple [husband and wife]") and fù 副 see above for "glasses", but it can also be used more broadly as m.w. in the sense of "pair; set; deck; pack", as in yī fù shǒutào 一副手套 ("a pair of gloves") (although that is more often written as yī fù shǒutào 一付手套) (m.w. "pay; hand over", but I think it's just being used as an easier way to write the syllable than 副) and yī fù pūkè pái 一副扑克牌 ("a deck / pack of cards").

  22. Victor Mair said,

    January 20, 2015 @ 10:27 am

    From a specialist on early Chinese thought:

    I only have two small points to offer:

    First, I was thinking about a similar phenomenon in English last week when I ran across two separate instances at Berkeley of people using "trainings" as a countable noun. It strikes me that the loss of the word "session" is perhaps a related simplification/obscuration, like one can say "a water" if the context determines that all choices come in plastic bottles.

    The other point concerns methodology. I wonder if the large number of machine-generated translations on the web affects the distributions you cite. My experience is that sometimes google translate will provide a generic 個, and there are a large number of commercial sites that try to generate hits by machine-translating English-language text found elsewhere on the web. Maybe I'm being optimistic, but I wonder if the results would not be so stark if one were able to distinguish human-entered text from machine-translated text?

  23. Dan Lufkin said,

    January 20, 2015 @ 10:59 am

    English also has brace for a word to make a singular unit out of a couple of things. I have no idea why we say a brace of pistols and a brace of quail but not a *brace of gloves. Perhaps the lack of a marked plural for game birds and the custom of pistols coming in identical pairs for dueling has something to do with it.

  24. Brett said,

    January 20, 2015 @ 11:41 am

    @Dan Lufkin: Is that sense of "brace" part of your active idiolect? It isn't part of mine, but I have no problem with "brace of pistols" or "brace of quail," because I've seen other people use them. While "brace of gloves" doesn't work for me, I wonder if somebody who would actually say "brace of quail" would have the same problem with it. In the case of gloves, the use of "brace" may be blocked by the ubiquity of "pair," but would something else—like "brace of computers"—work for a "brace" user?

    And are there other words (besides pistols and game fowl) for which "brace" is acceptable (if affected) for all educated English speakers? I've seen "brace of kinsmen" once or twice, but I took that be a specific allusion to Romeo and Juliet.

  25. Ethan said,

    January 20, 2015 @ 11:58 am

    @Brett: A brace of goals is the usual way of describing two goals scored by the same player in a soccer game.

  26. J. W. Brewer said,

    January 20, 2015 @ 12:08 pm

    "You say well ; look you, sir, there's a brace of angels, besides much drink of free-cost, if it be lik'd." — from the now-obscure play "The Hog Hath Lost His Pearl," which had its controversial stage debut in 1613. You can do your own research on "a brace of," but looking at a few pages of recent google books hits (where "recent" mostly got recent reprints of much older works) suggests that uses outside a hunting-and-guns ("brace of shotguns" seems to me to be as idiomatic as "brace of pistols") context are highly archaic. Whether people who are themselves deep into the hunting-and-guns subculture might use it more broadly would be an interesting question, however.

  27. J. W. Brewer said,

    January 20, 2015 @ 12:26 pm

    Hmm. "brace of goals" sounds 100% unidiomatic to me, but it's Out There. I expect it may be mostly a BrEngism or at least a non-North-Americanism. (My quick-and-dirty test was to look for it in an ice hockey context and find that the majority of the first two pages of hits were from outside North America despite the fact that only a teensy percentage of English-language writing about ice hockey is generated outside the U.S. and Canada.) There is definitely a little bit of US usage, though — it must have leaked into sportswriters' dialect through some vector or other. E.g., "Muffin Thitaparum got a brace of goals" in a description of a victory by the girls' soccer team of a private high school in New England.

  28. J. W. Brewer said,

    January 20, 2015 @ 12:37 pm

    BTW, googling "a brace of gloves" gets some fascinating hits, although they appear overwhelmingly not to have been generated by native speakers of English and, indeed, in many instances probably not by human beings. This point perhaps underscores the caution made above that some of the variability in Chinese usage might be an artifact of bad machine translation — one would ideally dip into the data and see if there are any systematic differences in general level of fluency/coherence in the contexts different variants appear in.

  29. cameron said,

    January 20, 2015 @ 12:59 pm

    "Brace of goals" isn't how I'd describe two goals. I would, however, say that someone "scored a brace", or "netted a brace" or "bagged a brace", with the goals being implied. A brace is one less than a hat-trick.

  30. Rachel said,

    January 20, 2015 @ 2:06 pm

    Interesting discussion! Especially interesting to see that there may be a regional/topolectal variation.

    I find it interesting in English that mass nouns with a plural attached tend to imply 'types of', as in "That grocery store has many cheeses on display" or "They have seventeen beers on tap".

  31. Victor Mair said,

    January 20, 2015 @ 3:15 pm

    From 無名 (anonymous):

    The operative comment on the "mass noun" issue is van Norden's. Hansen's hypothesis wasn't "discredited" by Harbsmeier's work, but did have to be weakened by it, and it still serves useful analytic and speculative functions in the opinion of many comparativists. The best brief critique of the matter is probably Graham's reply to the Harbsmeier article (pp. 274-78) in the Festschrift cited by van Norden.

  32. Belial Issimo said,

    January 20, 2015 @ 4:28 pm

    I'd have said that objects that naturally come in twos use 'pair' (gloves, hands, pants, eyes) whereas objects that can come in any quantity (quail, pistols, hounds) use 'brace' for two.

  33. Chau Wu said,

    January 20, 2015 @ 7:39 pm

    @a historian of Chinese thought

    I would like to echo your comments, which give a penetrating view on the subject, “The use of such ‘signifier terms’….was….a means of achieving referential clarity.” In a monosyllabically oriented language the use of pleonastic combinations of two synonyms helped to achieve clarity. Later the synonyms got separated, and one of them began to serve as “classifier”.

    David Branner observes in Problems in Comparative Chinese Dialectology: The Classification of Miin and Hakka (Mouton de Guyter, New York, 2000, p. 163) that, “Medieval and modern Chinese have developed entirely new syntactic particles, including, in some cases, categories wholly unknown in the early language, such as measure words.”

    During the intervening periods between the Han and Tang dynasties, Northern China was in the hands of foreigners. For example, the people of Hou Zhao (後趙) were known to have deep eye-sockets (深目), a prominent nose (高鼻), and lots of facial hair (多鬚). It was inevitable that a large influx of foreign words of the masters got assimilated in Chinese.

    From these three perspectives, we can analyze the use of classifier-noun combinations as two synonyms in pleonastic combination, a lot of which show similarity to foreign words (OE = OE; ON= Old Norse; OHG = Old High German; L = Latin; Gk = Greek):

    一頭牛yī tóu niú: 牛 ‘cow’ in Taiwanese is gû, compared with OE cū, ON kú ‘cow’; 頭 is thâu, which I suspect was borrowed for sound to write L. taurus ‘bull’ (> It. & Sp. toro, Cf. MSM tóu 頭). Therefore, 一頭牛 is literally ‘one cow-cow’, later the first morpheme was re-analyzed as a classifier.

    一匹馬yī pǐ mǎ: 馬 ‘horse’ in Taiwanese is má (lit. reading), compared with ON marr ‘horse’; 匹 is phit, which may be a loanword from (Late) OHG pfarifrit (> German Pferd). Thus, 一匹馬yī pǐ mǎ is ‘a horse-horse’.

    一kha 籃: 籃 ‘basket’ in Taiwanese is nâ. It might originally be tîn-nâ 藤籃 ‘a basket made of rattan or vine’ but became just nâ 籃 (by displacement) and the former may have been a loanword from ON teina ‘basket made of twigs’. The classifier kha (not sure there is a sinograph for it) may have its origin in L. calathus (or Gk. κάλαθος) ‘a wicker basket, a hand-basket’.

    一股涼風 yī gŭ liáng fēng ‘a [class.] cool breeze’: 股 in Taiwanese is kó·, compared with ON gol (= gul) ‘a gentle breeze’ or ON kul ‘a breeze’ (for the latter, it’s known that -ul > Tw. o·, for example, Kyrgyz kul ‘lake’ > Tw. hô· / ô· 湖 ‘lake’, Cf. MSM hú 湖).

    There are many other examples, but I should stop here as my comments are getting too long. Thank you.

  34. Matt said,

    January 21, 2015 @ 2:28 am

    Chau Wu, if that were the case, wouldn't we expect to see words like 頭牛 and 匹馬 used even when no numbers are involved?

  35. John said,

    January 21, 2015 @ 5:23 am

    Well, 馬匹 is used and so is 牛隻 and 人口. I can't think of any other classifiers offhand that are used in this manner though…

  36. DMT said,

    January 21, 2015 @ 7:44 am

    Not that it has any direct relevance on the issue of Old English loans into Minnan, but Cantonese uses classifiers without numbers to indicate a specific object: 頭牛="the/this/that cow", 個人="the/this/that person", etc. (Do any other topolects also do this? Also, what do we know about the history of classifier usage in Korean, Japanese, and Vietnamese?)

  37. J. W. Brewer said,

    January 21, 2015 @ 1:07 pm

    For those interested, there's a book-length treatment of the usage of classifier-type words in various languages that came out about a decade ago from Prof. Beckwith of IU-Bloomington http://www.amazon.com/Phoronyms-Classifiers-Pseudopartitive-Construction-Linguistics/dp/1433101394/ref=sr_1_1?ie=UTF8&s=books&qid=1205776236&sr=1-1. I got about halfway through it a few years back before finding it a little too dense for my mood at the time. I don't know how it's been reviewed or how it's regarded by others who've studied the phenomenon.

  38. Chau Wu said,

    January 21, 2015 @ 3:39 pm


    You raised a good question. Thanks a lot. My guess for the fact that we don’t see all of them still staying together is that, once the synonyms got separated, they had difficulty getting back together again. Similar to divorce, the partners, even after separating amicably, would be hard put to reconcile with each other again (pardon my crude analogy). Those few we find staying together today do so because they have found their usage in some fixed meaning.

    @John and DMT

    Thanks for the examples. I would like to add a few more.

    一條線 yī tiáo xiàn ‘a [class.] thread/string’: In Taiwanese 條 is tiâu, compared with German Tau ‘rope’, and 線 is siàn (lit. reading), which may have been a loanword from ON seil ‘string, line’ (Cf. German Seil ‘rope’). So, today we have 線條 xiàn tiáo ‘line’.

    一面鏡 yī miàn jìng ‘a [class.] mirror’: In Taiwanese mirror is 鏡 kèng, which may be derived from spec- of L. speculum ‘mirror’ following a pattern of sound correspondence sp- > h-/k-. 面 is pronounced bīn in Taiwanese, which may trace its origin to Romance *mīrātōrium ‘mirror’ (> It. miratore ‘mirror’ and of course Eng. mirror) with the sound change of mir- > *min > (after denasalization) bīn. So, toay we have 鏡面 jìng miàn ‘the surface of a mirror’, here 面 is re-interpreted as ‘surface’.

    一把劍 yī pă jián ‘a [class.] sword’: In Taiwanese 劍 ‘sword’ is pronounced kiàm, which may be related to ON skálm ‘short sword’ (following declustering of the initial sk- to k-). 把 is pá (lit.) / pé (vern.) in Taiwanese, and may trace its origin to L. spatha ‘sword’ (following declustering of sp- to p-). So 一把劍 is ‘one sword-sword’. Today we have 劍把 ‘the hilt of sword’.

    一間房間 yī jiān fáng jiān ‘a [class.] apartment-chamber’: In Taiwanese 房間 pâng-keng ‘room, chamber’ is a pleonastic combination. 房 pâng may be derived from the Latin verb appartare ‘to separate, set apart’, similar to Eng. apartment, but may have gone through aphesis > -part- and sound change of par(t)- to pâng 房. 間 is pronounced keng and can be seen as a derivative of L. cella ‘chamber, small apartment’ with the sound change of cell- > keng (Latin c- is always pronounced [k]). What is most interesting about this example is that 間 is not only staying in partnership with 房, but also functions as a classifier. Of course, one can say 一間房, here 間 is lost from 房間 by displacement. This seems to support the hypothesis of the need for pleonastic combination (for clarity) first, followed by separation to give rise to the classifier.

  39. Eidolon said,

    January 21, 2015 @ 10:24 pm

    Call me a skeptic, but the irregularity of the sound change rules proposed, especially when the reduplicates are at times consistent only with Minnan, doesn't look correct to me. Northern China was indeed taken over by 'foreigners' between the Han and the Tang, but Minnan is a southern Chinese language and it is not at all obvious that the phonetics of those terms in Minnan were derived from a northern source. Removing the Taiwanese specific rules, we only have:

    一條線 yī tiáo xiàn ‘a [class.] thread/string’: In Taiwanese 條 is tiâu, compared with German Tau ‘rope’, and 線 is siàn (lit. reading), which may have been a loanword from ON seil ‘string, line’ (Cf. German Seil ‘rope’). So, today we have 線條 xiàn tiáo ‘line’.

    一匹馬yī pǐ mǎ: 馬 ‘horse’ in Taiwanese is má (lit. reading), compared with ON marr ‘horse’; 匹 is phit, which may be a loanword from (Late) OHG pfarifrit (> German Pferd). Thus, 一匹馬yī pǐ mǎ is ‘a horse-horse’.

    一股涼風 yī gŭ liáng fēng ‘a [class.] cool breeze’: 股 in Taiwanese is kó·, compared with ON gol (= gul) ‘a gentle breeze’ or ON kul ‘a breeze’ (for the latter, it’s known that -ul > Tw. o·, for example, Kyrgyz kul ‘lake’ > Tw. hô· / ô· 湖 ‘lake’, Cf. MSM hú 湖).

    I have to say the tau>tiau is interesting, but the rest look to be a huge stretch of sound change rules. Why would pfarifrit/pferd be localized into pi and seil into sien? These are very different sounds. A thunni>thynne>sien derivation looks more plausible to me, but there are other problems. For one, 一股 is found in the Han Dynasty text 礼记, where it was already a classifier. Just the same, 一条 was found in Lunheng 論衡 and Mathematics 算書, both Classical Chinese texts. It's not necessarily the case that none of these are loan words, but I think the reduplicate theory and the timing needs to be changed.

  40. Victor Mair said,

    January 22, 2015 @ 5:04 pm

    From a Chinese philosopher:

    In classical Chinese, we do not have to use a measure word for an object in many cases, though to specify the exact number or to describe some trait a measure word is usually added. In the latter case of description, the choice of a classifier would be meaning- and context-dependent, with no fixed rules of application. For example, we can say yi2 shu4 li2 hua1 ya1 hai3 tang2(一樹梨花壓海棠), where li2 hua1 (梨花 pear flower) is modified by shu4 (樹 tree). We can also say li2 hua1 yi4 zhi1 chun1 dai4 yu3(梨花一枝春帶雨), where the same object (梨花 pear flower) is modified by zhi1 (枝 twig). In modern Chinese we are usually obliged to use such measure word for flower as duo3 (朵:a flower of flower?), zhu1(株:a stem of flower?) and zhi1 (枝:a twig of flower). But curiously in modern Chinese, the last case (枝 as in 一枝花) serves as a simple measure word, meaning just one flower, whereas in classical Chinese, the same word creates a richer meaning, describing in fact a flower twig. In that case, the classifier 枝 is combined with the object (花) and becomes a part of it (花枝). In classical Chinese, when we are not emphasizing the number but simply expressing an abstract singularity, the measure word is generally omitted. We just say, yi1 hua1 (一花a flower), yi4 cao3 (一草a stem of grass), yi2ye4 (一葉a leaf). In these cases, what is expressed is not the specific oneness of a particular object, but a singular universality. All these three (花,草,葉) require a measure word in modern Chinese (一朵花,一棵草,一片葉). But this omission of classifier when expressing universality may be true only of the singular form. When we increase the number it somehow gets erratic. In classical Chinese we seldom encounter such oddity as er4 hua1(二花), or liang3 hua1(兩花), or lia3 hua1 (倆花 two flowers). Yet it is perfectly sonorous and idiomatic to say yi4 hua1 (one flower). When the number gets huge, the rule of omission of classifier seems to return to that of the singular form. We say bai3 hua1 (百花 hundreds of flowers), qian1 hong2 (千紅 thousands of flowers), and wan4 hua1 (萬花 millions of flowers, as in 萬花從中過,片葉不沾身). As in the case of the singular yi4 hua1 (一花), these great numbers are universal and abstract. Since they do not refer to particular flowers, no classifier is required.

    Regarding the examples you have used, there are some subtle differences to be noticed. 一家動物園,for example, is not quite interchangeable with 一座動物園. The former typically refers to a zoo as an institution and the latter as a physical entity or architecture, while 一個動物園 is neutral between the two. This is generally true of nouns that refer to objects with both an institutional and a purely physical or architectural dimension, although the appropriate measure words may vary.

  41. Brian said,

    January 30, 2015 @ 3:35 am

    A mainland perspective: I'm not a native speaker but I've studied Mandarin over ten years; most of my Chinese is from the Mainland.

    When I saw the title of this post, my thought process was "个 — no wait, they wouldn't write a post about 个 — hmm, maybe it's 所, or maybe 家." The other two (座 and 场) didn't even cross my mind.

    I consulted a native Chinese speaker (in Mainland China until age 19). He said 个.
    Me: Anything else you can use?
    Him: 个.
    I asked him about the other four: grudging acceptance for 座 and one of 家 and 所 (I forget which); absolutely no for the other of them and 场.

    Then I realized you can get mainland-specific results by searching Baidu instead of Google.

    “一所动物园” 1,010,000
    “一个动物园” 683,000
    “一家动物园” 329,000
    “一座动物园” 50,400
    “一场动物园” 10,100

    And according someone on the internet, the "correct" word is 所.


RSS feed for comments on this post