A common claim about Chinese characters (Hanzi) is that they take less space than alphabetic systems and so using them “saves paper.” After all, there aren’t spaces between words when writing in Chinese characters, and Chinese characters handle entire syllables rather than having to spell them out letter by letter. So this claim would seem to be self-evident. But things don’t always work out as expected.
A few weeks ago I was browsing the shelves of the enormous, wonderful Eslite bookstore near Taipei City Hall. (Nobody seems quite sure how the so-called English name of this chain is supposed to be pronounced, so many foreigners here prefer the Mandarin name: Chéngpǐn (誠品).) In many of the store’s sections, English-language originals and their translations into Mandarin are shelved right next to each other. So, after looking at a science book in English I pulled out the Mandarin Chinese translation of the same work and browsed through it. While I was doing so, I noticed something unexpected: the Mandarin version was longer than the English-language original.
This sparked my interest, so I pulled out some more paired titles, more or less at random, off the shelves for the purpose of comparison.
I did my best to keep the comparisons fair. In almost all of the cases I compared pairs of trade paperbacks: standard trade paperbacks in English with standard trade paperbacks in Mandarin.
Also, I didn’t count the pages taken up by indexes, since none of the translations into Mandarin had indexes. (Alphabets win hands down over Chinese characters when it comes to creating and using indexes, and I saw no reason to penalize the English books for this by counting pages that the ones in Chinese characters didn’t have the equivalent of.)
In addition, I avoided old books, since I wanted to be fairly sure the Mandarin Chinese translations were from the same English text as I was looking at. (I do, however, have one book written in German and translated into English. I didn’t check to see if the Mandarin version was done from the German original or the English translation.)
Of course, comparing across scripts and languages is certainly not the same as comparing simply across scripts (Hanzi vs. Hanyu Pinyin); but one does what one can.
Later, when I was supplementing my survey at the Eslite bookstore on Dunhua South Road when I noticed an error in my original method: I had forgotten to check where in the book page 1 fell. Many (but not all) English-language books mark the first page of the first chapter as page 1; many (but not all) books printed in Taiwan, however, include the front matter in their pagination, which leads to the first page of the first chapter being page 10 or so. So to help compensate for my oversight, it might be fair to subtract 10 pages from the Mandarin versions of those titles below followed by an asterisk. (The ones without an asterisk are those I examined most recently — and more carefully.)
Here are the results of my admittedly brief and unscientific survey:
Chronicles, Vol. 1, by Bob Dylan
English: 291 pp.
Mandarin in Hanzi: 295 pp.Collapse, by Jared Diamond
English: 560 pp.
Mandarin in Hanzi: 609 pp.The Death of Vishnu, by Manil Suri
English: 283 pp.
Mandarin in Hanzi: 287 pp.Deep Simplicity: Bringing Order to Chaos and Complexity*, by John Gribbin
English: 235 pp.
Mandarin in Hanzi: 255 pp.Did Adam and Eve Have Navels?: Debunking Pseudoscience*, by Martin Gardner
English: 310 pp.
Mandarin in Hanzi: 367 pp.The Elegant Universe*, by Brian Greene
English: 428 pp.
Mandarin in Hanzi: 463 pp.The Enigma of Arrival, by V.S. Naipaul
English: 350 pp.
Mandarin in Hanzi: 422 pp.Harry Potter and the Half-Blood Prince, by J.K. Rowling
English: 607 pp. (hardback)
Mandarin in Hanzi: 716 pp.Laboratory Earth*, by Stephen H. Schneider
English: 169 pp.
Mandarin in Hanzi: 227 pp.The Long Tail, by Chris Anderson
English: 226 pp. (hardback, slightly larger than the Mandarin trade paperback)
Mandarin in Hanzi: 313 pp. (written left to right)Perfume*, by Patrick Su?skind
English: 255 pp. (translation from German)
Mandarin in Hanzi: 278 pp.Tough Choices, by Carly Fiorina
English: 309 pp.
Mandarin in Hanzi: 341 pp.Vernon God Little, by D.B.C. Pierre
English: 275 pp. (mass market paperback)
Mandarin in Hanzi: 325 pp.
In every instance, the books in Chinese characters are longer than those in English. Moreover, the pages in the Mandarin-language trade paperbacks are somewhat larger than those in the English-language trade paperbacks. So that’s even more paper consumed by the books written in Chinese characters.
Although I certainly do not believe that all pairs of books in English and Mandarin translation follow this pattern, a pattern this very much appears to be.
My guess would be that books printed in China would have fewer pages than those printed in Taiwan. (Anyone want to check some of the above titles? Or does anyone have pairs of other titles in unexpurgated editions?) In general, books in China simply aren’t designed and printed with the same degrees of competency, attention, and concern for the reader as books in Taiwan — not to mention books in the United States and Britain. (Or have things changed very much in this regard since I lived in China?) So, among other factors, the characters tend to be smaller, along with the leading and the margins.
And then there’s the fact that translations in China sometimes omit sentences or entire sections, especially if they are deemed “sensitive.” (I doubt, however, that the books I examined suffered from Beijing’s censors.)
Also, China’s left-to-right format might have an advantage over Taiwan’s predominant top-to-bottom style in terms of space.
Interesting! Have you compared books which are originally in Chinese with their English translation? I wonder whether the result would be different then.
Hi, Robin:
I would like to see some comparisons of Mandarin originals with English translations, but neutral comparisons are harder to make in those cases. First, the translations into English that are easiest to find are usually from Literary Sinitic (i.e., Classical Chinese), which is much more abbreviated than most modern texts. Also, the English translations of such texts usually have different footnotes than the original editions, throwing off the page counts. And many of these texts are poetry, which doesn’t fill up pages in the same way as prose does, also throwing off page counts.
Another difficulty is that Eslite doesn’t file English translations next to Mandarin originals, so these would take longer to track down.
But suitable items for comparison are surely there. Perhaps Fortress Besieged (Wéi Chéng / ?? / ??), which is one of my favorites.
I’d also like to see how Japanese texts fit into this.
You haven’t mentioned font size, or more precisely, line height. Do Chinese-language books have fewer lines per page than English-language books? It wouldn’t surprise me if Chinese characters are generally printed larger (in terms of height) than Roman letters, in order to keep them from becoming blurred due to the density of strokes. And as Chinese characters get taller, they will also get wider …
Zev: Yes, absolutely. Thanks for bring that up. I remember writing that point; but I must have deleted it when I reorganized the post. D’oh! As for fewer lines per page, yes, that’s so. But in Taiwan most texts are still written vertically rather than horizontally, so comparisons on the page level are tricky.
Another point I should make explicit: Chinese characters are all designed to fit within the same box (same height, same width). This means that legibility in terms of size is subject to the weakest link (i.e., the character with the strokes most likely to blend together). There are thousands of characters that become indistinct at small sizes, even if “simplified” characters are being used. As a result, all of the other characters in the text need to be made larger.
While Hanzi may not save paper in general use, Pinyin would tend to save paper (or breath) in the case of spelling numbers, don’t you think? I’m talking about comparing “seven thousand five hundred” with “q?qia?n wu?ba?i” (or “q?qia?n wu?”).
Well, actually I’m developing an online number-to-pinyin converter (non-commercial). Kind of like Erik Peterson’s site’s “Chinese Numbers”, but instead of tone numbers, I would like it to generate Pinyin with tone marks. Also, I would like it to connect the Pinyin syllables whenever appropriate.
Having practically no other resources, I found your site very helpful. But there are a few cases that have put me in doubt:
100000: shíwa?n or space-separated shí wa?n?
110000: shíy?wa?n or shíy? wa?n?
1000000: y?ba?iwa?n or y?ba?i wa?n?
Your help on this would be very much appreciated.
And do you think people in mainland also follow this rule (regarding the joining/separation of Pinyin syllables)?
I don’t think you need to control for font size. If your question is “which uses more paper?” an alphabetic system of hanzi, you find equivalent books and see which uses more paper, the reasons are irrelevant.
I think a good comparison might be Chinese and Vietnamese. A lot of books are translated from Chinese to Vietnamese (and Vietnamese people tend toward wordy translations IME). So you won’t have indexes to worry about and a lot fewer cultural issues.
Alternately the Taiwanese books in Church Romanization could be compared to the Hanzi versions.
This post was interesting but (as you mention) it should really be called “Does translating into Chinese [written in Chinese characters, not pinyin] save paper?” Which is an interesting question too but doesn’t tell us anything about pinyin, or hanzi vs pinyin.
What would be more interesting is comparing Chinese in hanzi to Chinese in pinyin, and surely if any site has a bunch of texts in both formats and the ability to present them side-by-side for easy comparison, it is this one… You could even throw together some in-page widgets allowing people to change font size and line spacing to get a better feel for the proportions involved.
(For the record, I’ve noticed that 1 Japanese bunko page equals roughly 1.1-1.2 English-translation paperback page, but note that English paperback pages are usually about 1.2-1.5 times the size of a bunko page. Also, I’ve only done this comparison with literature.)
Matt: The widgetizer is a good idea. I’m a little worried about how well it can be made to work under present browers, though. Too few fonts provide support for all of Pinyin’s diacritics, for example. And switching from horizontal to vertical formats could get tricky. And then there’s the fact that it, um, sounds like work, which I’m unfortunately already waaaaaaaaaay behind on.
But very soon I’ll put up a Hanyu Pinyin version of Y.R. Chao’s Mandarin translation of the Humpty Dumpty episode in Through the Looking Glass. That should be long enough to provide a useful comparison. (I already have up a parallel version in English and the Gwoyeu Romatzyh romanization system. I made the GR version a little wider — 53 percent — so the lines would match the English better. Hanyu Pinyin will take less horizontal space; but the diacritics may necessitate expanding the line-height.)
I’m a bit surprised about the length of Japanese works. Japanese orthography would seem to have the worst of both worlds when it comes to keeping things small and brief: the need to keep Hanzi/kanji fairly large to keep them distinct but also the need for lots of kana and sometimes romaji.
Michael: I wish I’d thought to check the lengths of the various editions of Streams in the Desert (one in a Mandarin translation in Chinese characters; one in a Taiwanese translation in a mixed orthography of mainly Chinese characters, with some romanization; and one in Taiwanese completely in romanization. Now I can’t even find these on the Web site of the press.
I’d also be very interested to see how line height is handled in nicely set Vietnamese editions. In other words: What sort of adjustments, if any, need to be made to line height in texts with lots of diacritics? Hanyu Pinyin, though, doesn’t have as many such marks as Vietnamese. The same examination could be done with, say, Czech. Or even between French and English.
I wonder who the Robert Bringhursts are of the non-English-speaking world.
Lucky: For the answer to your first question, see my latest post: spelling out whole numbers in Hanyu Pinyin.
As for your other question: wordparsinginChinais real ly ter ri ble.
Well, I do not know for paper saving, but when one goes to examine space, the comparison of the original and the translation of various laws and other such documents (as you may find for instance in Beijin Review supplement) would be quite correct since there is no much opportunity for style differences, omissions and other misleading things.
Mario, Croatia
I did a study about this years ago ~~~I dont know how you did that …but Chinese characters (HANZI) do save papers
From personal experience translating my essays from English to Chinese, I have to agree with Marbury that Chinese characters tend to save space. A two-page essay in English was often only three-quarters of a page in Chinese, sadly.
Also, the Mainland Harry Potter is *much* smaller than its American counterpart. I don’t remember exact page numbers, but it is definitely much smaller.
Its bizzare that your measurements showed English books to be more compact since its common knowledge that the reverse is generally true.
Pinyin will only work if the Chinese accept it and unfortunately it seems that the most interested group seems to be illiterate(in hanzi) westerners.
I do have to agree with Danielle above that the mainland left-to-right versions of Harry Potter books are drastically thinner than their English versions. I’m reading the first one right now and it’s about half as thick as the English version, both being about the same height and width. Same goes for the 6th book in simplified characters of which I also have a copy.
Pingback: Weird English, not so weird. « ~~~ASUN~~~~
Actually, Japanese texts are printed much smaller than English: 9 pt in Japanese vs. 11 pt in Latin alphabet.
But then you have to keep in mind that Kanji/Kana are as large as a capital letter, so all in all, it might be the same.
It is no problem to print Japanese in small letters, though, because you can easily guess the few Kanji from the context, and you usually wouldn’t use so complicated Kanji anyway. If there are some complicated Kanji, most people will know them.
By the way, when comparing: Chinese/Japanese has usually larger line spacing, because (if you compare it to latin letters), it is all printed in capital letters… and then you need larger line spacing.
Interesting stuff. I myself had been working as translator and am naturally keen on checking up the ratio between the Chinese character count and the English word count. From years of experience of my own, I got some personal conclusions on this matter that: an average result will be that 1,000 Chinese characters can be translated into about 600-700 English words, or 1,000 English words can be translated into about 1,500-1,700 Chinese characters, variying depending on the natures of the source contents and target writing styles.
In regard to the specific topic about the ratio between English word count and Chinese character count, you may be interested in reading this article (http://www.actranslation.com/chinese/chinese-wordcount.htm) dedicated for it with concrete evidences to support my conclusions.
Maybe its because you didnt take account of font size?
@Tong: Read the comments.