A reader of Chinese would be half way through this article by now

Well, the Chinese reader would be halfway through this article if it were written in Chinese anyway…

That’s the intriguing assertion from hsknotes that came out of hanzi-orthography discussion the other day:

Native speakers read chinese at the speed of light, no joke. It’s not even funny. They devour books written in Chinese in a few hours, or a day, while the equivalent thing in English would take them, or a native-english speaker, a substantially longer amount of time. Any small amount of testing will bear this out to anyone

Hsknotes “assumed it was common knowledge”, but I’d never heard this claim before. The sentiment was echoed by others in comments.

Fascinating hypothesis. Let’s give it a name, maybe, the ZEI hypothesis:

Zhongwen (i.e. 中文, written Chinese)
Enables
Incredible speed

OK, the acronym’s hokey, but it incorporates one of my favorite Beijing dialect words and taxicab conversations (go here and scroll down to 贼快 / zéi kuài if you want background).

To rephrase it, the ZEI hypothesis is saying, simply, that among equivalently proficient readers of Chinese and English, the former will be faster at reading the same content in their native language. And to state the obvious: we’d have to show that comprehension is about the same, or else speed by itself is pretty meaningless.

How to test the claim

The best way to test the ZEI hypothesis is probably to borrow someone else’s research, of course. It’s hard to imagine that no one has ever tried to test this. More importantly, what I’m about to propose might be too full of holes to constitute a reasonable test. So if you know of any work in this area, please comment.

In lieu of that for the time being, I’m wondering if the following test might be plausible and executable.

  1. Create Chinese and English versions of a text
  2. Show the text on-screen for a limited time
  3. Ask comprehension questions with the text no longer available for reference
  4. Score the comprehension question answers and compare native English to native Chinese speakers

That’s the short version. Here are some of the details I think would be important:

Create Chinese and English versions of a text. The text should be

  • Technical, but accessible to the lay person without background in the subject
  • About an obscure topic, something that most people wouldn’t have had occasion to come across
  • Long enough so that it would take, maybe, 4-5 minutes to read and understand thoroughly

Show the text on-screen for a limited time

  • Not enough time for most people to finish reading thoroughly
  • Maybe 2-3 minutes

Ask comprehension questions with the text no longer available for reference

  • Questions cover the entire length of the text
  • Idea is that readers would not know answers to later questions if they didn’t have time to read that far

Score the comprehension question answers and compare native English to native Chinese speakers

  • Idea is that Chinese speakers should have read further, according to the ZEI hypothesis
  • So if the ZEI hypothesis is true, Chinese will have higher comprehension scores

Would it work? The whole idea of putting this up here is to get feedback on design. If we can get this refined to a reasonable degree of satisfaction, maybe we can figure out a way to run it.

Why this matters

The ZEI hypothesis, if true, is incredibly important, at least to me. Why? Because of my general stance towards writing Mandarin with Chinese characters. Although my rhetoric is occasionally misunderstood as “anti” Chinese character, I will continue to say that this is not my position. I am no more “anti” character than one could be “anti” gravity. Characters are the writing system of China. Call it fatalism, but that’s how things are. I’m not interested in changing that.

In looking at Chinese characters (as well as English spelling) as a writing system, I simply want to separate out the facts from the myths. Here’s my simple version of it:

Facts unfavorable to Chinese characters (hanzi)

  1. It takes longer to learn to read hanzi than a phonemic script like Pinyin
  2. It takes much longer to learn to write hanzi than a phonemic script

Facts favorable to Chinese characters, that knowing modern hanzi

  1. Enables folks to read back through hundreds of years of written history
  2. Makes it easier to read back through many hundreds, even thousands, of years of written history. (Read: you still need significant education, far beyond just knowing modern characters, to be able to read classical works)
  3. Makes it possible to communicate with older generations that know no Pinyin.

In my haste I’ve probably missed many. Feel free to comment and we can add the ones we agree on to the list. The point is to separate out facts such as “X takes longer” from opinions such as “it’s absolutely vital for everyone in the next generation to be able to read 红楼梦 in the original Chinese characters”.

Looking at this list, generally, the “anti-hanzi” facts seem to be on the side of efficiency while the “pro-hanzi” facts seem to be on the side of social and historical continuity. But here’s where we get to the importance of the ZEI hypothesis.

If the ZEI hypothesis is true, it would be, I think, the first substantial pro-hanzi fact on the side of efficiency.

Now I realize there have been other claims made for the “efficiency” of characters. For example, people say it makes it easier to learn, say, Japanese, if you already know Chinese characters. Or I guess you could argue that there’s an efficiency argument in being able to read ancient Chinese texts more easily if you already know modern Chinese characters. The problem I have with such claims is that they depend on wanting to do something else, something other than obtain modern literacy in your native language. In my mind, then, the ZEI hypothesis is substantially different.

So is it true?

32 responses to “A reader of Chinese would be half way through this article by now”

  1. Max says:

    That’ll be an interesting experiment once it comes to fruition. One thing I’d like to add in favor of Hanzi is that they definitely make texts shorter. I mean, how could they not, with, simplistically speaking, 5000 letters vs 26? So even if it takes the same time to read one page of hanzi and one page of English, then hanzi would still win by a long shot, since one hanzi page packs five times the content of the English page

  2. Aaron says:

    My experience echoes Max’s in that when tweeting in Japanese or Chinese I find that I can pack in much more meaning than when tweeting in English (trying to write the same tweet in English invariably results in more than 140 chars).

  3. Blake says:

    It makes some intuitive sense that you can cram more information into a few Chinese characters than you can into the equivalent number of English characters. What seems less intuitive is the idea that Chinese speakers can process this information faster than English speakers can. On average, it seems like both groups would be able to process the information presented by the text at the same speed.

    If there is a difference at all, maybe it’s because Chinese is more efficient, and takes less words to convey the same thought. If this is true, then English speakers would be wasting more mental energy making sense of all the extra words, while Chinese speakers would already be off processing the information. But even this is hard to believe. In English, we waste hardly any time parsing every single word. I would guess that the bulk of the time spent with a text isn’t spent decoding the letters, but instead trying to understand the substance of the writer’s thoughts.

    I guess what I’m saying is that I can’t quite figure out why this difference would exist.

  4. Alex Whyte says:

    Lots of interesting posts on this blog, especially this one.

    What does it mean to “read faster” though? That’s easy enough to define when you’ve got two people reading the same language, but with the case of Chinese characters vs English, it brought up some questions for me.

    To say that an English and a Chinese reader are reading at the same speed, would it mean that in a given time they read the same gross number of lines/pages, or that they were imparted with the same amount of information? Let’s say that a morpheme in Chinese and a morpheme in English impart the same amount of information.

    I’m still an amateur at Chinese, but it seems to me that by that definition a page of Chinese characters would have more information than a page of English simply because it takes less space to print a Chinese character than it does a typical English morpheme.

    Should we create the sample reading for the test based on the number of morphemes? Say an equal number of Chinese characters and English morphemes. This would surely come out to a longer (page length wise) passage for English, but wouldn’t it be the same amount of information? Analogous, I guess, to comparing the length of a single Chinese passage written in characters with the same passage written in pinyin. One doesn’t contain more info than the other, but the page length is different. I guess that’s one example that proves equal page length doesn’t equal equal amounts of info imparted.(too many equals in that sentence.)

    So then what we would really be testing here is the speed at which readers read morphemes, and if Chinese people were found to read Chinese characters faster than English readers read English morphemes, then I guess that would make them faster readers.

    But if what I said about written Chinese taking up less print space than English is correct, and Chinese people were found to read a page of characters faster than English speakers read a page of English, then this would all be a moot point.

    If however, it was found that the English speakers read a page faster than the Chinese read a page, then I wouldn’t say it necessarily makes the English ones faster readers.

    Also, you say that the sample text should take about 4-5 minutes to read, but for who? English? Chinese? If you choose a passage that takes equal amounts of time to read in both Chinese and English, then what are you testing? Maybe I misunderstood…

  5. Syz says:

    @Max and @Aaron: I’m with @Blake here. In my mind, denser would just imply slower.

    @Alex (and maybe others were unclear about this too): The readings I have in mind would be the same content in English and Chinese, not the same “length” in any physical sense. Of course there’d be lots of quibbling over translations, and whether one language’s version of the passage is longer than the other language’s. But I think we could make it close enough.

    As for the “should take 4-5 minutes to read” question — yeah, I was kind of waving my hands. We’d have to do a bunch of calibration, but the idea would be to have much more than any normal reader (in English or Chinese) would be able to get through in the set period of time.

  6. Carl says:

    I have heard it said that for simple sentences, some languages can express certain things faster than others, but when you get to complicated stuff like, “She ate the birthday cake that I sent her two hours before the party started,” it takes about the same number of seconds no matter what language you use. So maybe compare something semi-complicated that takes an average speaker ten minutes to say in either language?

  7. Max says:

    syz, then I guess we need to differentiate between two kinds of speed:
    First, the speed to mentally process a character, be it hanzi or roman letter. Second, the speed to process content, e.g. one line of thought.
    This means you could be slower at processing characters, but still be overall faster in reading the meaning, if Chinese is more dense than English (which I’d argue it is)

  8. Julen says:

    Syz, I think the test is fine. I like the idea of forcing a limited time and then checking comprehension, it is more measurable than the other way round (ie. asking to read comprehensively and then measure their time).

    Regarding the question of more information in the same space: this is not what Syz is measuring. Since paper (or screen scrolling) is not a scarce resource nowadays, there is little real advantage in being able to cram more hanzi per square inch. Time, on the other hand, IS a scarce commodity. Therefore, what Syz is measuring (ie. content communicated/unit of time) is IMO much more important.

    Finally, I have seen a lot of misunderstanding regarding the concentration of hanzi. First of all, things like Twitter or other microblogs are not a good comparison. Of course each hanzi gives much more info than a single alphabet letter, but it also takes more bits. You could write 140 character letters in much less bits/pixels than 140 hanzi, comparing this is absurd, because a letter is not comparable to a character.

    In general, it is true that Chinese books and sites usually cram more info per page than English ones. But that does not necessarily imply an advantage of Hanzi, it might just be that Chinese editors like things less spaced. To do a fair comparison you would need to have the same level of reading comfort, which is very difficult to measure. But I am pretty sure you can read a much smaller font in English than you can in hanzi. Therefore, if the objective was really to cram information, it is not sure at all the hanzi would win.

    PS. SYZ, I have written a heated rant against your (outrageous) hanzi debate in the post. However, I have refrained from posting it here because I don’t want to distract attention from the main point. I will take a cold shower this evening and perhaps post it then if it still looks like a good idea.

  9. Julen says:

    @Max: It should be a comparison content to content. Making sure that both texts give the same amount of information, even if it is not 100% identical word to word (for example, English grammar might require repeating a pronoun like “they” where Chinese can omit it, etc.)

    Of course you could make a point that some of the speed would be due not to “hanzi” themselves, but to a more efficient Chinese grammar / vocabulary. But I am afraid both things are impossible to separate, otherwise the text would read very weird and it would not be a fair comparison.

    IMO, the texts should be written naturally, not forcing them for the purpose of the test. They should be taken from existing sources, and just check the translation for accuracy. Literary texts should be avoided as it is difficult to evaluate how much info is carried by a chengyu or a metaphor. A technical or a neutral journalistic piece should be good enough for the job.

  10. Syz says:

    @everybody concerned about information per square inch or whatever: I agree with what Julen said.

    @Julen

    a heated rant against your (outrageous) hanzi debate

    I’m chortling just anticipating it. Well, take your cold shower and remember what I said about gravity. It is what it is. I, for one, am not trying to change it — just trying to state some facts about its nature and have some fun in the process 😀

  11. pc says:

    Perhaps this would be helpful:
    Handbook of East Asian Psycholinguistics

    In particular “Processing of characters by native Chinese readers” by Marcus Taft.

    It’s a good and thorough work and it might answer some other questions about the more scientific aspects (outside of syntax or semantics) that you might have.

  12. Zev Handel says:

    The real test would be hanzi vs. pinyin, rather than vs. English. That way content would be identical, and you wouldn’t have to worry about the different morphologies of the languages. The problem is you’re not going to find a native reader of pinyin to do the experiment on.

    I’d add to your list of advantages to Chinese characters that they are pan-dialectal, in the sense that they have no inherent phonological system. Figuring out zh ch sh vs. z c s, or -in vs. -ing, or l- vs. n- is a big hassle for many Mandarin dialect speakers.

  13. Syz says:

    @pc: looks interesting, but are you sure you linked to the right pub? It’s hard to tell since it’s only a partial view, but the title you mention doesn’t seem to be in the table of contents…

    @Zev: the real test of what? Not that I wouldn’t like to test hanzi vs pinyin, but the assertion here is specifically about Chinese vs English.

    Are characters really an advantage for different pronunciations? I mean, I get that the speaker who says zong for 中 would have to learn to write zhong instead of zong. But I’m not convinced that learning to insert the extra h is inherently harder than learning to write 中 in the first place. Maybe I’m being overly simplistic, but it seems to me that it’s just another mapping of one form to another, whether it’s from putonghua-pinyin to dialect pronunciation or it’s hanzi to dialect pronunciation.

  14. Zev Handel says:

    @Syz: ah, I see, I didn’t realize the test was inherently Chinese vs. English, I thought it was characters vs. alphabet. So I retract my statement — testing characters vs. pinyin wouldn’t tell us anything about English!

    The way I see it, zhong and zong are equally plausible pronunciations associated with the written form . But there is an inherent mismatch between the pronunciation of as zong. Or go farther: there’s no inherent mismatch between Cantonese luk and , but there is between Cantonese luk and . A Cantonese speaker doesn’t need to learn Mandarin to write . A Cantonese speaker is going to have to learn Mandarin pronunciation to write .

    Now I’m perfectly open to the possibility that that is still easier than learning all those characters, but at that point the writing system is no longer dialect-neutral in terms of pronunciation. Whether you think there is any social good in a dialect-neutral writing system is another matter. As you know, one thing I really like about English spelling is that it is dialect neutral. You pronounce ‘whine’ differently from ‘wine’ but ‘marry’ the same as ‘merry’? The writing system is no more or less favorable to you than the speaker who does just the opposite.

  15. fremen says:

    @Julen
    About Twitter and SMS, there is a meaningful way to compare: an SMS contain 140 bytes; you can compare how much information you can cram in those bytes.
    If all the characters you send are ASCII (that can be encoded with 7 bits) they use the remaining bit in each byte to get a little more space and that’s how we get the at-most 160 (latin) characters.
    For Chinese, you encode the characters using two bytes per character, so at most you can use 70 hanzi.
    True, 160 latin-characters to 70 hanzi is comparing pears to apples. But if you compare how much information you can put in those bytes, my personal experience is, Chinese wins by and large.

  16. Katie says:

    So digging around in my memory a bit, some of the research on this person’s page is likely to be helpful. He’s a Taiwanese professor and definitely advocates romanization–but it’s been so long since I’ve heard him talk that I no longer remember why. I’d imagine that some of his papers would have literature reviews that would give you some information about studies on processing, acquisition, and so forth that support his view, and presumably also the ones that don’t that he’s arguing against.

    @pc–interesting looking book!

  17. Riley G says:

    Some research was done for Search engine results a few years ago with eye heat mapping, looking at results of their .com site and .cn site. They said this:

    One other amazing comparison was the amount of time spent on the page before the first click. In our North American studies, we saw an average interaction with google lasted about 8 to 10 seconds. In our study, an average interaction with google.cn lasted about 30 seconds, and with Baidu over 55 seconds.

    Not sure how much this compares considering that it’s a search engine but I thought it was interesting to note the difference in time it takes a user to search the results. This small data set points to comprehension faster comprehension of googles English results. In my opinion it seems to refute your ZEI theory.

    I know that google was designed in English first, but even Baidu was slower, so what does that mean?

  18. Syz says:

    @Riley,
    That’s a pretty interesting study, but there are way too many factors for me to think this has much to do with Hanzi. Chief among those are easy things like how many years the respondents have used the internet, and especially for how many hours. My guess is that the 18-25 yr olds they tested from China had not, as their US counterparts would have, grown up on hours a day of internet usage. But even if you could somehow control for that, there could still be cultural reasons.

    The “analysis” of the study is like something from the hanzismatter dark side though. Did you read on down?

    Because Chinese is presented as symbols, where concepts take their final meaning from a group of combined symbols…

    Yeah, well, ‘nuf said.

  19. Julen says:

    @fremen

    I see your point, and I agree 70 hanzi > 160 characters (including spaces). This explains the origin of all those comments on the internet: yes, you can type more info in a Chinese sms than in a english one.

    And yet, it is not really fair in terms of memory usage. This is because the ascii system is packed with things that are not really necessary for writing. Strictly speaking, you could get away with a 6bit (64 symbols) for alphabet, and about a 13bit (8000 symbols) for characters. This would make the comparison at:

    140 *6/13 = 64 hanzi per 140 characters.

    In raw mathematical terms, ignoring the byte system, the formula would be:

    X hanzi = ln64/ln8000* N alphabet. That is a 0.46 ratio, so 140 alphabet characters transmit about the same amount of raw information as 64 hanzi.

    Now, you probably think that you could say more things with 64 hanzi than with 14o characters, and I agree. But I am not sure what this tells us, really. And I am not sure this can bear any relation with the efficiency (speed) of hanzi writing in transmitting information. I have to think of this a bit more.

  20. Julen says:

    @Syz, sorry I have been away from language this week, doing a completely different kind of research that I just finished today.

    I will be back with characters as soon as I can, my speculation is far from finished. In the meantime, my ears are beeping at the mention of another romanizationist. I think it is time I bring up that rant I was talking about.

    Here we go, please don’t take it personally, I am NOT referring to you in the last paragraph :)
    ————————————————-

    Syz,

    I completely disagree with the whole hanzi debate you add in the second part of the post. You do pros and cons as if you were deciding which brand of smartphone to buy, with all respect I think this perspective is superficial. Regardless of the results of your test, there is a better reason why Chinese should use hanzi: because it is their written language, as simple as that.

    All of your points are based on efficiency assumptions, as if being able to read Tang poetry in its original script could be quantified in these terms. But nobody really NEEDS to read Li Bai, and for those who want to do it, they will always be free to study characters anyway, nobody is speaking of banning it.

    My point is simple: efficiency does not and cannot rule everything we do in life, particularly language and culture.

    If it did, then why the hell do Chinese babies learn Chinese culture at school, and why even bother developing a standard mandarin for all China? You could just treat mandarin as a language for use at home (like Shanghainese today) and get the kids to learn written English instead. Besides, instead of wasting time studying history and literature, they could start from a young age learning C++ and HTML5.

    There you go. Much more efficient, the GDP will fly in 30 years. Chinese people will not write any more silly Chinglish signs, and foreign sinologists will finally breathe triumphantly as the dreaded hanzi lie defeated.

    I really have to wonder why some people ever went into Sinology. They would have been better off studying Italian, the challenge was probably more adapted to their intellect, and they would be now happily sipping Amarettos, instead of hating themselves and their subject of study.

  21. Chay Schiller says:

    @ Syz

    Don’t discount what Zev has to say about pinyin vs. hanzi. Although the hypothesis is about Chinese vs. English, the reasoning behind the hypothesis rests on the differences not between the languages but between the languages’ written systems, i.e. phonetic vs. logographic. Isn’t that right? If the difference isn’t ascribed to the writing systems, it would have to be due to some inherent superiority in the Chinese language. If this is the true assertion, I would suggest that the writing system is a more plausible explanation.

    If the difference does come from the writing system, to test the hypothesis it becomes necessary to ensure that differences in reading comprehension arise because of the written system of the content and for no other reason. There are so many unpredictable causes for the difference in reading comprehension between Chinese and English speakers that it would be difficult to ensure the validity of the results. However, if you use the same person to test both a phonetic and logographic writing system, then the control variables are all in place already, except for their skill with each writing system.

    My suggested experiment is to take a native Chinese speaker who is proficient with hanzi and train him/her intensively in pinyin to a similar degree of litaracy. Then, the subject’s reading speed and comprehension would be tested across a body of texts all graded to the same reading level and, as previously suggested, on unfamiliar topics. Half the readings would be in hanzi and the other half pinyin. If the subject has an easier time with the hanzi, then we can infer from the results that Chinese people do have an easier time with comprehension simply because they have the benefit of hanzi.

    That is, unless English makes up for it by being an inherently superior language :)

  22. Syz says:

    @Julen: glad you made it back with your “rant”. I certainly take no offense. This is basically a restatement of your earlier comment on another post. I called it then, and I still believe today, that it was indeed “the most rhetorically sound defense of character conservatism I’ve ever read.”

    And for the record, I think it’s worth restating over and over.

    The trouble with efficiency, as you say, is that it’s only a part of the whole script argument. I may not have said it clearly, but I was trying to make that point in my post as well. What’s interesting to me is the idea that (contrary to what I’ve thought before) there might be an efficiency argument for hanzi.

    But let’s say it doesn’t pan out. Let’s say we get a reasonable test specced* out. We run it and find that, boringly, everybody reads at roughly the same speed.

    What do we do with that? From the standpoint of advocating/not advocating a switch to Pinyin: absolutely nothing. It doesn’t really matter because, as you say and Zev keeps saying and I essentially agree: script change is much more than efficiency.

    BTW: I have a copy of that “when script changes work” paper that I referenced way back when. I will eventually get around to summarizing the framework. I think you’ll like it because it’s much broader than “efficiency”.

    ——–
    *For me, “spec” is a perfectly reasonable verb in the sense of “make specifications for” but the past tense spelling is kind of iffy

  23. Syz says:

    @Chay Shiller: thanks for defending Zev’s pinyin vs hanzi. I didn’t mean to sound dismissive. I actually think that would be very interesting too.

    The trouble I have with it is the “fluent reader” problem you mention. Any educated reader has spent SO much time reading in hanzi that (it seems to me) it’s going to be very tough to get to a comparable level of fluency in pinyin. Or kind of the other way around: any reader of Chinese who’s utterly “fluent” (seems like a weird word for a script) in Pinyin is going to be a weird scriptomaniac outlier that I wouldn’t trust to use for generalizable results.

    Idunno. Maybe these are unreasonable biases…

  24. Julen says:

    @Syz – Will be interested to see that.

    @Chay, Zev – You are right. What the test compares is not the scripts’ efficiency, but rather the written languages as a whole – including script, grammar, vocabulary, etc. The same goes for the comparison we were doing last night with the bits and bytes in the sms.

    This is indeed a problem, because I don’t really suppose the pinyin-hanzi experiment is feasible in any significant way. Bear in mind you would have to find not one, but quite a few people naturally fluent in pinyin to do the experiment reliably…

    This requires some more serious thought. I plan to cut down on my beer this weekend and save the Sunday for thinking. Any ideas?

  25. Chay Schiller says:

    @Syz, Julen: I agree that the final control variable I name in my post, “skill in each writing system”, would be a tough one to put in place. I would counter that the myriad of control variables necessary to ensure the validity of the original test may be just as troublesome.

    Perhaps a compromise: how about finding test subjects who were raised bilingually in Chinese and English, and who all have a similar amount of experience reading both languages? If a trend can be found towards Chinese being the easier to comprehend or even the more preferred reading material among these subjects, I think the results will be meaningful. You could even introduce an element into this test to untangle the scripts’ efficiency from the languages’ inherent differences: find a similar group bilingual in Chinese and a phonetically written language other than English and see if the trend runs more or less strongly towards Chinese.

  26. Julen says:

    Chay, all very good ideas, especially the last paragraph.

    I have been thinking of a different approach, essentially looking at it from a theoretical point of view (since the tests you suggest are beyond my means).

    My idea is still cooking, but it goes in the lines of finding a neutral text, in which Chinese and English grammar/vocabulary cannot justify a difference on speed.

    I am based on the observation that: on legal texts (I see loads of that, I’m working with bilingual contracts all the time) the translation is essentially word-to-word, there are few periphrasis (because all legal words are well defined in both languages) and the order of the sentences is practcally the same.

    It would be feasible to find a text that reads naturally IN BOTH languages, at least for lawyers. If we choose this text for the experiment, the difference in speed could only be due to the script.

    OK, there are still some differences, like the English text has articles such as “the” which the Chinese has not. Also obviously some languages have longer words than others, but we could choose on purpose a text that had the same number of syllables in total and STILL read natural for lawyers.

    What do you think?

  27. Chay Schiller says:

    Julen, I think that’s a good idea, and I actually don’t think maintaining an equal amount of syllables will be necessary. I believe we gave confirmed that English speakers naturally pay little attention to lexically insignifigant words such as “the”, “of”, etc. when reading and pay most attention to the words that carry meaning in a sentence. Given this, I think a word-for-word translation of a legal contract would work perfectly.

    However, I do feel there will be a problem determining whether the differences in comprehension should be attributed to differences in the reading level of the test subjects. Equivalent reading level must be one of the control variables, but the traditional means of assessing this rest on the same methods used in the experimnt itself. In obtaining subjects with equivalent reading level in both languages, we have already confounded the results because the Chinese speaker might have an equivalent level simply because of the ease of reading his language and not because of equialent mental functioning or experience with the language. This is why I suggested using the same test subject for both test in each of my propose experiment: we can be sure that reader in both tests has had the same level of education, mental functioning etc. And so the comprehension differences must reult from the language and not the subjects.

    Perhaps the better test would be to find subjects that demonstrate the exact same level of comprhension in each language and then determine if the Chinese subjects don’t need as much mental functioning, experience reading etc. to get there.

  28. Julen says:

    The thing is: if both languages are read similarly fast, say within a 10% difference in speed, then we have to be very careful with the way the test is done, or it will not be significant.

    On the other hand, if there is really a big difference in speed, then it should be quite easy to see it with a few homemade tests, ignoring the little details.

    Today I did the rough test again with one of my Chinese friends, and the result was I actually read in English faster than him in Chinese, by a good margin.

    I am beginning to doubt that any big difference exists, or even that there is an advantage at all in hanzi. I am running out of Chinese friends willing to come to my place and be experimented upon.

  29. Oi-lin says:

    Re: Native pinyin readers
    Why not simply have native readers of Dungan, as it’s written using the Cyrillic alphabet?

  30. Oi-lin says:

    additionally, I really loathe the new design. The text is terribly difficult to read.

  31. […] (I’m talking about studying languages, of course.) And in another post from Sinoglot, Sinoglot talks about the theory that Chinese people read Chinese really, really fast. I believe they do. I once bought a whole bunch of Konan the Detective comics in Chinese. One […]

  32. Colin Zwanziger says:

    this blog post addresses the question at hand:http://persquaremile.com/2011/12/21/which-reads-faster-chinese-or-english/
    upshot: some research has already been done and it tends towards the conclusion that Chinese and English read at exactly the same speed

Leave a Reply to Julen