Decolonizing Chinese fonts by probing the past

« previous post | next post »

New article by Brian Ng in Rest of World (9/6/21):

"Revolutionary type: Meet the designer decolonizing Chinese fonts

Julius Hui, who has done custom work for companies like Tencent, wants to radically rethink Chinese fonts."

I find this article to be curiously counterintuitive:  Julius Hui, the font designer, wants to revolutionize Chinese typography by hearkening back to a time before modern (say, the last four or five hundred years) fonts for typesetting.  That would be like telling designers of modern fonts for northern European languages to go back to the 4th-century pre-Gothic script of Ulfilas (or Wulfila) to develop a "revolutionary" new script for English or for designers of modern fonts for southern European languages to go back to the uncial majuscule script of roughly the same time period that was used for Greek and Latin.

I personally would like to see radically new fonts created for Sinographs, but — in my estimation — Hui's Ku Mincho project is barking up the wrong tree.  He is so fixated on traditional calligraphic excellence that he seems conceptually unable to advance to a new level of reenvisioning the shapes and forms of Chinese characters to suit the new media for which they are now intended, especially computers and phones.  We have often touched upon the difficulty of deciphering the strokes of characters used in such digital media, especially those characters that have more than 12 strokes, which is roughly the average number of strokes for the 13,000 commonly-used Chinese characters that Hui wants to put in the Ku Mincho font.

Aside from the sheer multiplicity of strokes in Sinographs, one thing that makes them hard to distinguish are the calligraphic flourishes (hooks, turns, elongated dots, etc.) that tend to fill up what white space they may have between strokes.

The mother of one of my graduate students spends hours each day writing out Buddhist sutras.  She could do this with a brush, which is what most people would do, making it a calligraphic exercise, but my student's mother chooses not to do so.  Instead, she writes the characters with a ball point pen, which I would consider a purer form of writing characters than doing so with a brush, where so much of one's attention is devoted to ornamentation and embellishment, whereas my student's mother is content to concentrate on the abstract essence of writing itself, for which I have great respect.


Selected readings


[h.t. Ben Zimmer]


  1. S Frankel said,

    September 11, 2021 @ 7:31 am

    Here's an article that explains the esthetics and has some decent examples:

    To my eyes, it looks very beautiful, but I don't know Chinese, so who cares?

  2. Victor Mair said,

    September 11, 2021 @ 7:54 am

    @S Frankel

    Thanks very much for finding that. Seeing the actual forms of Ku Mincho illustrated here, I have to say that it does represent an innovation in Sinographic font design, but not a truly creative breakthrough.

    I keep thinking about what the equivalent of sans serif would be for Chinese.

  3. Cervantes said,

    September 11, 2021 @ 8:31 am

    I guess this is semi on topic. I've just read that woodblock printing became common in China in around the 9th or 10th Century. It must have been very laborious to cut the woodblocks, I would think. Does anyone know how it was done? Just a guy carving away with a little chisel, or were there ways to make it more efficient?

  4. Antonio L. Banderas said,

    September 11, 2021 @ 9:18 am

    Beautifying Font: Effective Handwriting Template for Mastering Expression of Chinese Calligraphy

    Effects of Font Size, Stroke Width, and Character Complexity on the Legibility of Chinese Characters

  5. Antonio L. Banderas said,

    September 11, 2021 @ 9:54 am

  6. Twill said,

    September 11, 2021 @ 9:58 am

    It's a shame the linked article promoting the great accomplishment in creating any new Sinitic typeface leans on the old chestnuts of blaming the Qing, Westerners, and the Japanese (and, on their side of the strait, the CCP) for all of China's ails, and rather galling that it was wrapped up in "revolutionary, decolonial" language. The narrative doesn't even make sense: if mingti is a soulless imitation by Qing scholars that marked the beginning of calligraphic decline then why not, uh, make an imitation Song or brush typeface? If mass printing was problematic then why chose a family whose major distinctive is printed serifs? The interview S Frankel linked is much better, as you might expect from the mouth of the designer himself, and actually talks about the typeface's selling points: breaking from the conventional proportioning of characters (which would seem to me lie more in the laboriousness of weighting characters crying out for geometric formalizations) and its distinctive serifs.

  7. Rodger C said,

    September 11, 2021 @ 10:21 am

    for designers of modern fonts for southern European languages to go back to the uncial majuscule script of roughly the same time period that was used for Greek and Latin.

    Well, the font this blog is in was adapted from the Carolingian minuscule in Quattrocento Italy, or else we'd all be using Fraktur.

  8. David Marjanović said,

    September 11, 2021 @ 11:06 am

    I keep thinking about what the equivalent of sans serif would be for Chinese.

    I've always taken these sorts of typefaces as sans-serif, and now I see they're often called "sans-serif" or "gothic". I'm curious to learn why you seem to disagree.

  9. Victor Mair said,

    September 11, 2021 @ 12:31 pm

    I don't disagree. They're a nice beginning for what I'm contemplating about future developments.

  10. Antonio L. Banderas said,

    September 11, 2021 @ 1:02 pm

    >a nice beginning for what I'm contemplating about future developments.

    Could you elaborate a bit?

  11. Timothy Rowe said,

    September 11, 2021 @ 2:36 pm

    It looks to me as if what it has gone back to is something analogous to proportional spacing. Just as typewriters constrained Western text to fixed spacing and computers liberated is from that, so it seems the Japanese style of printing Sinetic characters constrained them to fixed spacing, and this seeks to escape that. It's not a perfect analogy, of course, because printing in the West, as opposed to typewriting, retained proportional spacing.

  12. Mark said,

    September 11, 2021 @ 2:55 pm

    Very interesting 2 articles (main text and first comment). Personally, they seem a slight aesthetic improvement, but I was slightly disappointed – the big serif bulges, the pointiness at the end of some strokes. If only Chinese characters were generally printed in larger fonts – perhaps then there would be more of a demand for calligraphic beauty as opposed to legibility. Chinese is usually displayed in about the same size as Roman letters, despite having far more visual complexity. A good break from "foreign influence" would be to print them in a larger size more appropriate for their beauty and complexity.

  13. Victor Mair said,

    September 11, 2021 @ 3:41 pm

    One of John DeFrancis' great innovations for the ABC Chinese Dictionary Series at the University of Hawaii Press — and there were many — is that he insisted on making the Chinese characters larger than the surrounding Roman letter text. I was with him when we worked out the protocols for the dictionaries, and I at first thought it would be technically impractical to be able to do that. But John was the kind of man who always tried to make technology serve ideals, not the other way around. So, for almost everything he strove for as an ideal, his perseverance usually made it happen.

  14. Keith said,

    September 11, 2021 @ 4:29 pm

    John was the kind of man who always tried to make technology serve ideals, not the other way around

    Which is how the world should always work…

  15. Alex Turner said,

    September 11, 2021 @ 5:39 pm

    @Victor Mair,

    I don't quite see how "calligraphic flourishes" sacrifice readability. Filling in white space, certainly can decrease distinguishability. But to me (especially from the webpage that @S Frankel shared) Ku Mincho aims to be more "shapely" (the boxes are distinctively more square, and other critical elements like hooks, angles, and dots are less modest). I don't find Hui's font especially revolutionary, but it seems like a step in the right direction.

    As far as you comparison of ball-point to brush calligraphy. At least in Taiwan, this is very common practice to meditate on writing Buddhist sutras (done with a pen not a brush, usually). When I do Chinese calligraphy (be it with brush, pen, pencil, or charcoal), my attention doesn't focus too much on the ornament, but overall shape (the shape of the white space too): I ask, "does the character sit well in its meaning for this context?" For me, even high-cursive (草書) with a brush focuses on "the abstract essence of writing itself", almost to a fault. I agree that it is a good thing to simply meditate on writing itself, but I think the medium has very little to do with it.

    I agree completely with what @Mark suggests. I'd like to see more experimenting with the digital presentation of Chinese script–that seems like the place for more growth. The struggle then becomes how to integrate a larger Chinese script with a usually-smaller Latin script, Arabic script, etc.

  16. Victor Mair said,

    September 11, 2021 @ 6:52 pm

    Your first two sentences directly contradict each other.

  17. Jonathan Smith said,

    September 11, 2021 @ 8:19 pm

    The notion in the eyeofdesign article (not sure to what degree reflecting Hui's view) that "the Japanese kanji were changed to fit almost perfectly inside a square box," imitation of which led Chinese writing to "stray… from its visual ancestry" seems dubious in light of, say, Han-era clerical script. Also if decolonizing, why the Japanese name :D

  18. RfP said,

    September 11, 2021 @ 11:06 pm

    Continuing on the point made by Alex Turner, I'll mention that I was a digital typesetter for quite some time in the 1990s, and I wish I remember more of the details from that period during which I intensely fascinated with typography and its history.

    One thing I will say is that readability and distinguishability are definitely not the same thing!

    And to the best of my recollection, although the grotesques and other sans serif typefaces are highly distinguishable, they are not the most readable typefaces. That is, if you are poring over a few dozen or a few hundred pages of text, you are—generally speaking—much better off with a serif typeface, even if serif faces aren't as distinguishable as sans faces.

    The "rediscovery" of the garalde typefaces—that is, the reaching back beyond the Modern Bodonis and the Transitional Baskervilles to "ancient" forms—led to new advances in readability. Jan Tschichold, in particular, created Sabon, after being a champion of "the new typography"—that is, of the supremacy of sans typefaces for… just about everything.

    Stanley Morrison's development of Times New Roman, a more quotidian face that was not, as I recall, a Garalde, was still at least somewhat rooted in the renaissance of "pre-Modern" text faces.

    When I read the AIGA article, I couldn't help but notice the comparison of Ku Mincho to a an absolutely stunning "a" from a Latin typeface, I believe a garalde, that is clearly rooted in the uncials.

    I was never a type designer, and my memory of these things is flawed, to say the least, but Hui seems to be doing a service to Chinese typography that could help revolutionize the readability of the language.

    I wish someone who was more of a type historian could chime in here…

  19. RfP said,

    September 11, 2021 @ 11:41 pm

    I'd also like to mention that William Morris had an important effect on the development of twentieth-century typography of the kind I mention above. He was a seminal influence!

    And this is in spite of the fact that his types really were pretty much along the lines of what Victor is criticizing in the beginning of this post.

    As Wikipedia points out in the article on his Golden Type (, "Morris's aim in the Kelmscott Press was to revive the style of early printing and medieval manuscripts, and the design accordingly is a profound rejection of the harsh, industrial aesthetic of the contemporary Didone typefaces used at the time in general-purpose printing, and also of the relatively pallid "modernised old style" designs popular in books. Instead, the design has a relatively heavy "colour" on the page."

    In spite of this reaching back to medieval aesthetics, and in spite of the fact that no one in their right mind would want to use the Golden Type for mass-scale printing, Morris profoundly moved the type designers who followed in his wake, most notably, in the beginning Frederic Goudy.

    One key factor was the type color, which had a profound aesthetic impact. Even though the best type designers of the twentieth century backed away from the perhaps excessive heaviness and some of the other outdated characteristics of Morris's typography, they still benefitted from, and often celebrated, his incredible breakthroughs.

    Art seldom develops in a straight line.

  20. Victor Mair said,

    September 12, 2021 @ 5:07 am

    "Art seldom develops in a straight line."

    Amen, brother / sister!

    That's why I'll keep hoping for that big, creative breakthrough.

  21. Platy Hsu said,

    September 12, 2021 @ 6:54 pm

    Mandarin speaker here. I have followed the development of Ku MinCho and the designer himself for a while and found the project a worthy endeavor. I agree with the point of the post that modern typefaces should consider the digital context; but with regard to Chinese typefaces, the main problem is we have few high-quality typeface to use — “modern” iOS and Android still lack a serif font for UI purposes as of today — and the only options are dominated ones that are over-optimized for pixel screens to the extent of dullness and unnatural.

    Indeed, the referenced post addressed this exact topic:

    > The shift to digital media in the 2000s drove another evolution, to what Hui calls “fat and blocky” characters. Characters were enlarged, and the spaces between strokes were increased so they could be read in small print on screens, and printed in newspapers.

  22. Alex Turner said,

    September 13, 2021 @ 1:15 pm


    I think a study of the difference between readability and distinguishability would be very interesting! A rather simplified approach to measuring distinguishability would be to overlay each glyph to a bitmap, unwrap the bitmap to a binary series (using Hilbert's curve, or something similar), and process that with something like a Shannon index (usually used to measure informational entropy). But readability necessarily brings in the human factor (at least by the definition of readability I'm thinking of), so that's always a little trickier.

    I think it's interesting how other people define "flourishes" and "ornaments". Rereading over the comments, it seems most people default "ornament" as being an additive process, which is, perhaps, a more precise definition, but I personally think there are some subtractive processes that can be considered ornamental: a pair of distressed jeans, poet eliding articles, Warhol shooting a hole in one of his prints, a type-designer thinning out the stem of a glyph's punch. Such things can embellish just as much attaching as bells and whistles.

    Seems like this difference in definition of additive "flourishes" and additive-and-subtractive "flourishes" is why the author saw a patent contradiction in my previous comment.

  23. James Wimberley said,

    September 20, 2021 @ 11:53 am

    Only tangentially on topic, but the chart of script samples in the Wikipedia page on uncials suggests that we should give more credit to the despised barbarian Visigoths as cultural innovators. It looks as if they – or rather he monks working under their patronage – were the first to use strong ascenders and descenders. Beyond making many letters and therefore words more distinguishable, they make a line of text more lively. Alcuin was right to take up and generalise the intervention for Charlemagne.

RSS feed for comments on this post