• Typefaces
  • Tomorrow
  • Articles
    • Tratex
    • Kleisch
    • Ruder (Part II)
    • Tissi
    • Ruder (Part I)
  • Bespoke
  • Editions
  • Typewriter
  • Timeline
  • Information
  • Account
  • Shop

Kleisch:
Drawing Connections Between Latin Serif and CJK Ming Typefaces

By Chia­chi Chao — The chal­lenge of bi-scrip­tural type­set­ting poses sig­ni­fic­ant dif­fi­culties for design­ers and type­set­ters, who must har­mon­ise vari­ous styles across dif­fer­ent scripts. Sim­ilar tasks have his­tor­ic­ally been a fun­da­mental respons­ib­il­ity of punch­cut­ters and type design­ers, and traces of their efforts can some­times remain vis­ible for cen­tur­ies. In con­tem­por­ary Latin typefaces, for example, upper­case char­ac­ters derive from Roman monu­mental cap­it­als, lower­case from the revived Car­oling­ian minus­cule, numer­als from Arabic numer­a­tion, and ital­ics from the Chan­cery hand. This phe­nomenon extends beyond Latin scripts and is also evid­ent in CJK (Chinese, Japan­ese, Korean) typefaces, which fre­quently incor­por­ate mul­tiple lan­guages and scripts.

Problem Statement

In bi-scrip­tural type­set­ting with CJK Ming and Latin serif typefaces, two prac­tices are most common today: either using the default Latin char­ac­ters that are often included in CJK fonts, or select­ing an exist­ing Latin typeface with sim­ilar visual attrib­utes. Unfor­tu­nately, both meth­ods lead to unsat­is­fact­ory res­ults rather fre­quently. On the one hand, default Latin char­ac­ters are often impaired by rough draw­ing qual­ity and some­times naïve design decisions res­ult­ing from a lack of expert­ise in Latin type design among CJK Ming design­ers. On the other hand, it is dif­fi­cult to find an exist­ing Latin typeface that aligns well with the very dif­fer­ent visual para­met­ers, pro­por­tions, con­trast and grey tone of Ming typefaces.1 Espe­cially in light styles, com­pat­ible options are rare.

Accept­able default pair­ings of Ming with Serif Latin styles.

How­ever, three rather suc­cess­ful examples of default pair­ings allow for a closer under­stand­ing of some of the cri­teria at hand. For example, Apple LiSung incor­por­ates a par­tic­u­larly light weight of Times New Roman which doesn’t exist in the ori­ginal family of this typeface. Des­pite the slight incom­pat­ib­il­ity due to the brack­eted serifs and con­densed coun­ters, this com­bin­a­tion is quite effect­ive thanks to the gen­er­ous x-height and high-con­trast let­ter­forms of Times New Roman. Sim­il­arly, Adobe’s Source Han Serif uses a Latin typeface by Frank Grießhammer, which was based on Pierre Simon Fournier’s model but aligns well with Chinese char­ac­ter pro­por­tions thanks to a larger x-height res­ult­ing from redu­cing the extenders. Finally, the Latin char­ac­ters of Hama Mincho are derived from a new ren­der­ing of Pierre Didot’s cut. The stroke con­trast matches the Ming typeface well, yet the let­ter­forms convey a dis­tinctly dif­fer­ent cul­tural connotation.

Methodology For a New Approach

When cre­at­ing a new Latin serif typeface that would offer a fit­ting and ver­sat­ile coun­ter­part to CJK Ming typefaces, I set out from two basic assump­tions: First, I wanted to avoid unusual shapes ori­gin­at­ing in CJK typefaces that would com­prom­ise legib­il­ity in the Latin. Shapes or styles from one script system – if mind­lessly trans­ferred – can intro­duce visual noise in the other and thus dis­rupt the read­ing pro­cess, as observed by Zuzana Licko: ‘people read best what they are most famil­iar with.’2

Fur­ther­more, imit­at­ing the appear­ance of one cul­ture’s script in another can result in cul­tural inap­pro­pri­ate­ness. For instance, ‘Wonton fonts’ or ‘Bamboo fonts’ are typefaces that mimic the stroke of a Chinese cal­li­graphy brush and are often used for Chinese res­taur­ants or by the early Chinese dia­spora in non-Chinese-speak­ing coun­tries. How­ever, to Chinese speak­ers, these fonts can be seen as ste­reo­typ­ing and even racist. As recently as 2021, an art­icle on the CNN web­site noted that ‘for an older gen­er­a­tion of Asian Amer­ic­ans, spot­ting the imit­ated brush­stroke let­ter­ing can trig­ger past trau­mas.’ 3

Ste­reo­types in cul­tural appro­pri­ation: Latin char­ac­ter shapes mimick Chinese brush strokes.

In my attempt to find a new method for align­ing Latin serif let­ter­forms with Ming, I set out from the obser­va­tion that both scripts under­went his­tor­ical pro­cesses of ration­al­isa­tion when they trans­formed from cal­li­graphy to more geo­met­ric mod­u­la­tion. Look­ing at these trans­form­a­tions in par­al­lel, I assumed, might help to detect some fea­tures that became common in both scripts at a cer­tain time. High­light­ing such fea­tures would then allow to develop Latin let­ter­forms that share with Ming some of their visual para­met­ers and parts of their visual atmo­sphere, includ­ing cul­tural connotations.

Trans­ition­ing from cal­li­graphy to geo­met­ric mod­u­la­tion: Ming and serif Latin styles both went through vari­ous his­tor­ical pro­cesses of rationalisation.

Rationalisation in Ming Typefaces

The earli­est ver­sion of the Ming style resem­bling modern designs appeared in the book Mozi during the Song Dyn­asty (960–1279), but the style only reached matur­ity during the Ming Dyn­asty (1368–1644), when it was intro­duced for mov­able type. It was inten­ded to mimic the style of Kaiti, also known as Reg­u­lar Script, which is the most basic and common upright script in Chinese calligraphy.

Kaiti style typeface (left) and earli­est recog­nis­able Ming style from the book Mozi (right).

Over time, the Ming typeface gradu­ally devi­ated from Kaiti to facil­it­ate the adapt­a­tion of hand­writ­ing to print­ing. To ensure that each char­ac­ter fits evenly within a square, the pro­por­tions of the Ming typeface became much more rigid and squar­ish than those of Kaiti. Addi­tion­ally, the slanted hori­zontal strokes of Kaiti became per­pen­dic­u­lar, and the cal­li­graphic ter­min­als were ration­al­ised into tri­an­gu­lar shapes to sim­plify cut­ting. Known as uroko in Japan­ese, the tri­an­gu­lar ter­min­als reflect the slight dot cre­ated by the subtle pause in brush cal­li­graphy, where the brush moment­ar­ily lifts. Yet they can also be com­pared to tri­an­gu­lar serifs in cer­tain Latin typefaces. Des­pite retain­ing some traces of hand­writ­ing, then, the Ming style even­tu­ally evolved to rather strongly diverge from the struc­ture of human hands and its pre­de­cessor, Kaiti.

Kaiti style (left); later Ming style adap­ted for print­ing (right).

The stand­ard­ised ver­sion of Ming has become one of the most widely used styles for CJK typefaces. While it is often referred to as Ming and some­times as Song in Taiwan and Hong Kong, it is com­monly called Song in China. It is known as Mincho in Japan; and as Myeong­joche in Korea. Ming and Song gen­er­ally refer to Chinese char­ac­ters, yet Japan­ese Mincho includes Hiragana and Katakana, and Korean Myeong­joche com­prises both Hangul and Hanja.

Rationalisation in Latin Typefaces

In Europe, the move­ment of ration­al­ising Latin type star­ted during the 17th cen­tury, when punch­cut­ters sought to refine Renais­sance types (such as those by Gara­mond, Gran­jon, and Le Bé) to make them more com­pat­ible with print­ing tech­no­logy. They also tried to align their shapes with the rational spirit of the Enlight­en­ment, such as in Romain du Roi (1692), which marked a cru­cial shift in West­ern typo­graphy, intro­du­cing let­ter­forms based on geo­met­ric prin­ciples, pro­por­tional struc­tures, and sys­tem­at­ised met­rics. Even­tu­ally this would lead to the devel­op­ment of modern models of serif types, such as in the ‘Didone’ styles (includ­ing Didot and Bodoni), which gen­er­ally exhibit greater con­trast and devi­ate fur­ther from the cal­li­graphic origin of Latin letters.

Among the first to refine the Renais­sance models, Baroque styles have recently become recog­nised as an inde­pend­ent cat­egory since the 1960s. They are con­sidered to have ori­gin­ated in the south­ern Neth­er­lands in 1550, when Hendrik van den Keere of Ghent first broke with the Gara­mond style, intro­du­cing ‘fat-faced’ and ‘large-height’ types.4 Baroque types are fur­ther­more char­ac­ter­ised by mod­u­lated strokes, large x-heights, vari­able axes, con­trasty serifs, and lach­rymal ter­min­als. These fea­tures can be observed in designs by Nich­olas Kis, Chris­tof­fel van Dijck, John Fell, and Wil­liam Caslon, among others.

Baroque styles: Van Dijck, Ascendon­ica (left); Kis, Clein Canon (right).

Robert Bring­hurst dis­tin­guishes the Baroque from the later Neo­clas­sical models, which start with Romain du Roi, in his study The Ele­ments of Typo­graphic Style.5 While both models share basic con­struc­tion prin­ciples such as mod­u­lated strokes and typ­ical fea­tures like the lach­rymal ter­min­als, the Neo­clas­sical char­ac­ter shapes exhibit an even more rational struc­ture with stress on the ver­tical axis and mod­er­ate aper­tures. Con­trast is often more pro­nounced with strokes already coming close to the expan­sion model later cre­ated with point nibs (see the chapter ‘Match­ing stroke models’ below). The Italic har­mon­ises with the Roman coun­ter­part. This cat­egory of Neo­clas­sical typefaces is often asso­ci­ated with the works of Pierre-Simon Fournier, John Bask­erville, and Joan Michaël Fleischman.

Neoclas­si­cism: Fleisch­mann, Augustyn.

Where to Start?

Con­sid­er­ing the tra­ject­or­ies of ration­al­isa­tion in Ming and Latin serif typefaces in par­al­lel, I star­ted to look for points of mutual exchange and align­ment between the two. While the prop­erly modern ‘Didone’ styles are also fre­quently used along­side Ming, I opted for focus­ing on the earlier Baroque and Neo­clas­sical models. These are more sim­ilar to Ming in that they are also rep­res­ent­ing a syn­thesis of con­trast­ing design approaches, still reflect­ing hand move­ments while also integ­rat­ing ele­ments of char­ac­ter shapes that diverge from tra­di­tional tools such as brushes and nibs. In my view, ‘Didone’ styles also align less well with Ming cul­tur­ally, as they lack human­istic traits and imply luxury.

Examin­ing many typefaces from the Baroque and Neo­clas­sical peri­ods, I found that most of them com­ple­ment well with Ming in cer­tain aspects, yet they also con­trast in other regards. For example, the Baroque model, includ­ing the work of Kis and Van Dijck, aligns more closely with Ming in visual ambi­ence due to its cal­li­graphic resid­uals. Mean­while, the rational con­struc­tion of the Neo­clas­sical model (Fleisch­man, Fournier, and Bask­erville) offers improved pro­por­tions and con­trast when com­pared to Ming.

An unex­pec­ted dis­cov­ery of a typo­graphic com­pos­i­tion by the late Taiwanese artist Huang Hua-Cheng even­tu­ally hinted into a prom­ising dir­ec­tion. It fea­tured a mod­i­fied cut of Caslon with reduced extenders, paired with a medium weight of Ming. The lach­rymal ter­minal of the lower­case ‘a’, the strong con­trast in the serifs, and the move­ment of the ter­minal in ‘e’ cre­ated a har­mo­ni­ous pair­ing along­side Ming.

Poster design by taiwanese artist Huang Hua-Cheng employ­ing a mod­i­fied style of Caslon.

This dis­cov­ery promp­ted fur­ther explor­a­tion of work by Wil­liam Caslon I, who cut Roman types based on Dutch Baroque as late as 1722, aiming to adapt 17th-cen­tury type for the 18th cen­tury. In com­par­ison to Neo­clas­sical cuts by Fournier, Fleisch­man, and Bask­erville from the same time, Caslon’s type exhib­its a rel­at­ively organic qual­ity, resem­bling hand­writ­ing. Inspired by these let­ter­forms com­bin­ing Baroque and Neo­clas­sical models, I decided that my new typeface should also blend the Baroque model with the rational clar­ity of the Neo­clas­sical era, syn­thes­ising key fea­tures from both styles. After a thor­ough ana­lysis of his­tor­ical sources for the two models, I settled on Baroque cuts by Van Dijck and Kis as the primary sources for defin­ing the gen­eral fea­tures of the new letterforms.

Van Dijck design that served as a start­ing point for the devel­op­ment of Kleisch.

As a sec­ond­ary source, I used Fleisch­man’s Augustyn Roman as the main ref­er­ence for the pro­por­tions, due to its gen­er­ous x-height and bal­anced counter space. Sev­eral other typefaces from the hands of Caslon and Fell were also sampled to ensure select­ing the most accur­ate indi­vidual fea­tures of each let­ter­form for pair­ing with Ming.

Fleisch­man design used as ref­er­ence for the pro­por­tions of Kleisch.

Matching Stroke Models

To estab­lish a rela­tion­ship between Latin serif and Ming in terms of stroke models, it is help­ful to con­sider Gerrit Noordzij’s theory of strokes, which dis­tin­guishes a trans­la­tion model from an expan­sion model. The trans­la­tion model is pro­duced by broad nibs and is char­ac­ter­ised by the two fixed points of the nib, which Noordzij refers to as coun­ter­points. In con­trast, the expan­sion model is cre­ated by point nibs, with the stroke width determ­ined by the chan­ging expan­sion value of the two sym­met­rical points of the nib.6

Trans­la­tion model of strokes res­ult­ing from use of broad nibs (top), and expan­sion model from point nibs (bottom), accord­ing to Gerrit Noordzij.

To ana­lyse the stroke logic of the Ming style within Noordzij’s theory, one could observe that the strokes exhibit char­ac­ter­ist­ics of both models. This is because the tool used in CJK cal­li­graphy is neither a broad nib nor a point nib, but rather a brush. As a result, the size of the stroke in CJK writ­ing is determ­ined by the pres­sure applied, akin to the expan­sion model from point nibs, while the form of the stroke pos­sesses a unique shape that indic­ates path dir­ec­tion, sim­ilar to the trans­la­tion model.

Stroke model of Ming style, blend­ing trans­la­tion and expan­sion models.

Gen­er­ally, the ‘old-style’ Latin types from the Vene­tian to Renais­sance peri­ods are based on the trans­la­tion model, whereas modern typefaces are developed from the expan­sion model. Baroque and Neo­clas­sical types often blend both models, lead­ing to cer­tain arbit­rary fea­tures. Many Baroque typefaces dis­play traits of both models, par­tic­u­larly those designed by Van Dijck and Kis. For example, while the top ter­minal of the Gara­mond ‘a’ ori­gin­ates from the cut point of the quill, the same letter in Kis’ type has an arbit­rar­ily formed geo­met­ric circlet. Addi­tion­ally, the flat serif (with almost no brack­ets) and the upright stress of the ‘o’ are also indic­at­ive of the expan­sion model. How­ever, the trans­la­tion model still pre­dom­in­ates in the strokes of let­ters, espe­cially in the shoulders of ‘h’, ‘n’, ‘a’, and in the bowls of ‘b’, ‘d’, ‘p’, and ‘q’. In my new typeface, I used the geo­met­ric ter­min­als and trans­la­tion strokes as dis­tinct­ive char­ac­ter­ist­ics ori­gin­at­ing from Kis.

Matching Ductus and Writing Tempo

Let­ter­forms are shaped not only by the tools but also by the way they are writ­ten, often referred to as ductus. Chinese cal­li­graphy fea­tures a dis­tinct ductus in which most strokes ter­min­ate with a return motion, res­ult­ing in the form­a­tion of geo­met­ric fea­tures such as tri­an­gu­lar-shaped ter­min­als (lín), beak-like hooks (gōu), and return­ing strokes (zhé).

Beak-like hooks (gōu) in Chinese calligraphy.

This char­ac­ter­istic ductus, marked by a pause or return in the stroke, is also evid­ent in typefaces cut by Fleisch­man, par­tic­u­larly in the ver­tical serifs seen in let­ters such as ‘E’, ‘F’, and ‘L’, as well as the top ter­min­als of ‘n’, ‘m’, ‘d’, and ‘h’, which bear a resemb­lance to Ming. The ter­min­als and serifs in Fleisch­man’s work have thus been incor­por­ated into the design of my typeface, allow­ing it to reflect the fluid­ity of move­ment in Chinese cal­li­graphy while pre­serving its Latin typo­graphic roots. This approach achieves a har­mo­ni­ous visual unity that com­ple­ments Ming typefaces well.

‘E’ shape from Fleischmann dis­play­ing ver­tical serifs with bear resemb­lance to beak-like hooks in Chinese.

The tempo at which let­ters are writ­ten is also a cru­cial factor in determ­in­ing the final let­ter­form. In Chinese brush cal­li­graphy, the writ­ing speed often changes within a single brush stroke. In Latin cal­li­graphy, let­ters are gen­er­ally writ­ten at a rather steady pace. How­ever, some typefaces from the Baroque and Neo­clas­sical peri­ods indic­ate changes in writ­ing tempo as well, par­tic­u­larly at the bottom ter­min­als of the let­ters ‘a’, ‘c’, and ‘e’. This char­ac­ter­istic is espe­cially evid­ent in the Ascendon­ica Romein cut by Van Dijck, and in Kis’ Clein Canon Romein. It was fur­ther high­lighted in Fell’s type, which was heav­ily influ­enced by Van Dijck. Modern reviv­als of Kis’s and Van Dijck’s work tend to elim­in­ate these fea­tures, as can be seen in a Mono­type revival of Von Dijck, for example. Kleisch returns to the ori­ginal shape and dynamic of the ter­min­als, as this helps to achieve har­mony with the Ming style.

Speed changes are vis­ible in ter­min­als drawn by Kis and Van Dijck (left). The shortened ter­min­als in a Van Dijck revival by Mono­type loose the char­ac­ter­istic ductus (right).

Matching Proportion

Next to indi­vidual fea­tures of the let­ter­forms that determ­ine the visual para­met­ers and atmo­sphere of the Ming and Latin typefaces, pro­por­tions are also key for align­ing the two scripts. Unlike Latin char­ac­ters, which have ascend­ers and des­cend­ers, Chinese, Japan­ese, and Korean (CJK) char­ac­ters are con­struc­ted within a square frame­work, giving them a visual sim­il­ar­ity to Latin capitals.

Square frame­work typ­ical of CJK char­ac­ters; Latin cap­ital letter.

In Latin typefaces, visual size is largely determ­ined by the rela­tion­ship between the x-height and the extender height. A Latin typeface with a large x-height gen­er­ally pairs more har­mo­ni­ously with CJK typefaces, as it achieves a more bal­anced visual pro­por­tion. His­tor­ic­ally, extenders were reduced in the shift from Gara­mond to the Baroque period. This first appeared in the work of Hendrik van den Keere in the late 16th cen­tury, and sim­ilar pro­por­tions were adap­ted by Kis and Van Dijck. The x-height was even fur­ther increased in Fleisch­man’s Neo­clas­sical model, which was spe­cific­ally designed for small sizes.

In modern reviv­als, how­ever, the ascend­ers were often lengthened to fit the aes­thet­ics of the early 20th cen­tury. An example for this can be found, again, in the Mono­type revival of Van Dijck. In my typeface, I restored and syn­thes­ised the ori­ginal pro­por­tions found in Fleisch­man’s and Kis’s work, to create a more visu­ally bal­anced appear­ance when set along­side the Ming style.

Typ­ical Baroque extender length in Van Dijck (left), and altered accord­ing to early 20th cen­tury tastes in a Van Dijck revival by Mono­type (right).

Fine-Tuning Individual Letterforms

The ‘a’ in Kleisch syn­thes­ises ele­ments from Van Dijck, Kis and Fleisch­man, fea­tur­ing a top-heavy pro­por­tion and a slightly tilted shoulder of Baroque types. The top ter­minal is rendered with subtle ambi­gu­ity between broad nib and expan­sion nib.

The ‘n’ of Kleisch draws from two models: Fleisch­man’s dynamic top serifs (also in ‘m’, ‘h’, ‘d’, etc.), reflect­ing the flex­ible nib akin to Chinese brush strokes, and Baroque-inspired, high-con­trast, brack­et­less serifs that echo the crisp details in Ming typefaces.

The Baroque ‘s’ is char­ac­ter­ised by a slight rota­tion of the nib on the lower stem, a fea­ture typ­ic­ally softened in the Neo­clas­sical period. The ‘s’ forms in Kis and Van Dijck are marked by asym­met­rical ter­min­als. Baroque-style con­trast is pre­served in Kleisch, while the ter­min­als align with the Neo­clas­sical tradition.

The design of the curved cross­bar in Kleisch’s ‘A’ pays homage to the dis­tinct­ive ‘A’ found in Ehrhardt MT (a renowned Kis revival from Mono­type). This fea­ture likely ori­gin­ates from the ‘A’ in Kis’s Goote Text, allow­ing addi­tional room in the restric­ted counter space lim­ited by the high crossbar.

During the Baroque and Neo­clas­sical peri­ods, a more dynamic ‘C’ emerged in Roman cap­it­als, influ­enced by Italic forms and seen in the work of Fleisch­man and Kis. With its abrupt stroke ter­min­a­tion, this form shares a visual trait found in CJK cal­li­graphy, and it inspired the design of the ‘C’ in Kleisch.

The design of the ver­tical serifs is a key fea­ture of Kleisch, inspired by the spiky serifs of Fleisch­man, which echo the return­ing strokes in Chinese cal­li­graphy. How­ever, Fleisch­man’s style felt too flam­boy­ant along­side Ming type, there­fore the serifs were sim­pli­fied to align more closely with the styles of Kis and Van Dijck.

The ‘M’ is an import­ant indic­ator to reveal the iden­tity of a serif typeface. In Neo­clas­sical types, the stems of the ‘M’ are gen­er­ally ver­tical, while in Baroque types, they tend to tilt inward. The final design decision for Kleisch’s ‘M’ places it closer to its Baroque roots.

Kleisch’s italic was designed in the Baroque fash­ion, with incon­sist­ent stem angles, con­densed pro­por­tions, and a lively, fast writ­ing style. The cross­bar in the italic has a slight slant, adding vibrancy to the over­all texture.

Variable Weight and Contrast:
Matching Grey Tone

Match­ing the grey tone in trans-cul­tural type­set­ting of Ming and Latin is chal­len­ging. Even with a large x-height in the Latin typeface and a slightly reduced char­ac­ter size in the Ming typeface, Latin char­ac­ters often appear lighter in com­par­ison due to the com­plex stroke con­struc­tion in CJK scripts, espe­cially Chinese. Since many Ming char­ac­ters use sig­ni­fic­antly more strokes than Latin ones, they appear darker when the stroke width is even. This effect is less notice­able with larger type sizes, as the human eye tends to shifts focus from over­all grey­ness to stroke thick­ness if the char­ac­ters are larger. For this reason, using slightly thicker Latin stems can help achieve a more bal­anced visual grey­ness for text and cap­tion sizes, while even stroke width is gen­er­ally recom­men­ded at larger sizes.

To enable for con­tinu­ous adjust­ing of the stroke width, I decided to build my typeface with two vari­able axes for weight and con­trast. This allows for a refined adjust­ment of the grey tone, depend­ing on the style of the respect­ive Ming typeface used in the pair­ing. As a con­sequence, Kleisch per­forms very well in pair­ings with a large vari­ety of Ming typefaces in vari­ous weights. One way of assess­ing the achieve­ment is to com­pare Kleisch with the default Latin char­ac­ters cur­rently found in the Ming typefaces.

In another series of tests for the ver­sat­il­ity of Kleisch, I com­bined vari­ous instances in a sample layout with bi-lin­gual text set­ting. In each case, weight and con­trast were fit to match the the grey-scale of the respect­ive typefaces, includ­ing Apple LiSung Light, Toppan Bunkyu Mincho Reg­u­lar, YuMin­cho Demi Bold and Source Han Serif TC Heavy.

  1. These cri­teria are derived from Japan­ese typo­grapher Mariko Takagi’s present­a­tion in Wittner, 284.

  2. Licko.

  3. Quito.

  4. Ver­vliet, 38.

  5. Bring­hurst, 12–13.

  6. Noor­dzij, 7.

BIBLIOGRAPHY

Bring­hurst, Robert. The Ele­ments of Typo­graphic Style, Seattle, WA: Hartley & Marks, 1992.
Licko, Zuzana. ‘Zuzana Licko on Design and Legib­il­ity’. Emigre no. 11 (1992).
Noordzij, Gerrit. The Stroke: theory of writ­ing. London, UK: Hyphen Press, 2006.
Quito, Anne. ‘Karate, Wonton, Chow Fun: the end of “chop suey” fonts’. CNN, https://edi­tion.cnn.com/style/art­icle/chop-suey-fonts-hyphen­ated/index.html. Accessed Decem­ber 2024.
Ver­vliet, Hendrik D. L. French Renais­sance Print­ing Types: a con­spectus. New Castle, Del.: Oak Knoll Press, 2010.
Wittner, Ben, Sascha Thoma, and Timm Hart­mann, eds. Bi-Scrip­tual: typo­graphy and graphic design with mul­tiple script sys­tems. Sulgen, Switzer­land: Niggli, 2018.

Decem­ber 2024. Copy­right © 2024 by Chia­chi Chao & Lineto. All rights reserved.