Multiscript Language Support
The product of a four-year research project into Greek and Cyrillic typography, Whitney’s Latin character set now extends to sixty additional languages, reaching another quarter of a billion readers worldwide.
Whitney Extended Latin
Most fonts contain the accents necessary to accommodate the major Western European languages, at the expense of the rest of the world. Type foundries have traditionally addressed these absences on an ad hoc basis, creating one-off variants like “CE” (Central European) which might handle Polish, Czech, Hungarian, or Romanian — but not necessarily all of them. Often ignored were major languages like Turkish, with fifty million speakers, and “minor” ones like Catalan — a language commonly neglected, yet more widely spoken than Danish.
Unlike legacy formats such as PostScript and TrueType, OpenType fonts are virtually limitless in the number of characters they can contain. In preparation for moving its library into OpenType, Hoefler & Co. established the Language Research Program in 2005, in order to develop updated specifications for its Latin-based character sets. The initial product of this research is our Latin-X™ character set, which expands the reach of a typeface to an additional 200,000,000 readers worldwide. This character set reflects not only a more accurate awareness of the political landscape of language, but the most up-to-date understanding of cultural norms.
The Latin-X™ character set covers Afaan Oromo, Afrikaans, Albanian, Alsatian, Aragonese, Arapaho, Arrernte, Asturian, Aymara, Basque, Bislama, Blackfoot, Bosnian, Breton, Catalan, Cebuano, Chamorro, Cheyenne, Cimbrian, Corsican, Croatian, Czech, Dalecarlian, Danish, Dutch, English, Estonian, Faroese, Fijian, Filipino, Finnish, French, Frisian, Friulian, Galician, Gallegan, Genoese, German, Glosa, Guarani, Haitian Creole, Hawaiian, Hiligaynon, Hmong Daw, Hopi, Hungarian, Ibanag, Icelandic, Ilokano, Indonesian, Interlingua, Irish, Irish Gaelic, Istro-Romanian, Italian, Japanese transliteration, Jèrriais, Kapampangan, Kashubian, Kiribati, Koongo, Korean transliteration, Kurdish, Ladin (Gardena), Ladin (Valle di Badia), Languedocien, Latin, Latvian, Lithuanian, Lojban, Lombard, Low Saxon, Lower Sorbian, Luxembourgeois, Macedo-Romanian, Malagasy, Malay (Latinized), Maltese, Manx, Maori, Marshallese, Megleno-Romanian, Míkmaq, Mohawk, Montenegrin (Latinized), Náhuatl, Naxi (Latinized), Norfolk/Pitcairnese, Norwegian, Nyanja, Occitan, Oromo, Pangasinan, Papiamento, Pedi, Piedmontese, Polish, Portuguese, Potawatomi, Quechua, Romanian, Romansch, Rotokas, Rundi, Saint Lucia Creole, Inari Sami, Lule Sami, Samoan, Sardinian, Scots Gaelic, Serbian (Latinized), Seychelles Creole, Shona, Sicilian, Slovak, Slovene, Somali, Sorbian, South Ndebele, Southern Sotho, Spanish, Swahili, Swati, Swedish, Tagalog, Tahitian, Tausug, Tetum, Tok Pisin, Tongan, Tswana, Turkish, Turkmen (Latinized), Tuvalu, Upper Sorbian, Uyghur (Latinized), Veps, Volapük, Votic, Walloon, Walpiri, Welsh, Xhosa, Zhuang, and Zulu.
The written forms of Greek and Latin make plain the differences between these two alphabets. Perhaps because they lived in an age of written language, Northern European typefounders of the sixteenth century seldom confused the two: their Greek typefaces, still some of typography’s best, were not adaptations of Latin designs, but rather autonomous designs with only an incidental relationship to their Latin cousins.
The idea that one alphabet could be imagined as an extension of the other is a modern conceit, a convenience that produces designs both impractical and offensive. It begins with the observation that Greek and Latin share more than half of their capitals, and imposes this relationship onto the lowercase as well, where the forms of these alphabets should naturally diverge. Hasty adaptations of Latin designs for the Greek market exhibit a variety of thoughtless shortcuts: some unnecessarily force the structure of the Latin a or d onto the Greek alpha (α), or offer up the Latin s in place of the sigma (ς); especially lazy designers balk at the challenge of drawing the zeta (ζ), and instead use the Latin z — a decision akin to using a backwards S in place of a figure 2. Western type manufacturers of the early twentieth century were responsible for the first slapdash adaptations, creations that eroded the integrity of their Latin originals, diluted the distinctiveness of Greek typography, and posed genuine problems for typesetting the Greek language. Yet today, designers can be found in many countries producing questionable adaptations, leaning too heavily on the ease with which digital font outlines can be copied and pasted. Whatever the visual, cultural, and practical flaws of such shortcut fonts, their greatest drawback may be their legal standing, if they are produced without the permission of the original copyright holder.
Whitney Greek was a four-year project, designed by Hoefler & Co. in consultation with Gerry Leonidas, Senior Lecturer in Typography at the University of Reading. For Whitney, Leonidas advocated the adoption of two distinct formal vocabularies for the Latin and Greek ends of the type family, an approach supported by the unique characteristics of the Greek language. Where the subtle variation of repeated forms lends cohesion to a Latin lowercase — b, d, p and q share the same basic ingredients — a similar consistency in Greek would make its three most common letters, α, ο, and σ, difficult to distinguish. Rooting the design of Whitney Greek in written forms rather than printed ones opened the design to new gestures, and helped alleviate the hypnotic rhythm of circle-shaped counters so common in rote Greek adaptations. The result is a family of types with lively and organic forms, and a texture appropriate to the requirements of the Greek language.
The Greek character set covers Modern (monotonic) Greek.
Whitney Extended Cyrillic
A curious challenge when designing a Cyrillic typeface is the matter of correctly counting how many letters that alphabet contains. Originally developed by Byzantine missionaries in the ninth century, the Cyrillic alphabet shed eight of its original 45 letters in the 1708–10 reforms of Peter the Great, and lost four additional letters under the Bolshevik purge that followed the October Revolution of 1917. The remaining 33 letters satisfied the Russian language, but during the expansion of the Soviet Union, Cyrillics were pressed into service to render an ever-broadening collection of unrelated tongues. Unlike the Latin alphabet, which serves Germanic, Romance, and West Slavic languages chiefly through the addition of accent marks, the application of Cyrillics to Slavic, Turkic, Uralic, Tungusic, Caucasian, and Mongolic languages was principally achieved through the invention of new letters, which were tacked onto the core alphabet as needed. As a result, the alphabets of neighboring Russia and Ukraine differ by four letters, Bulgaria borrows thirty letters from the Russian alphabet (but draws sixteen of them differently), and the Azeri language of Azerbaijan — spoken by as many as thirty million people, and codified five different ways since 1918 — includes one letter used nowhere else in the world. A typeface might include “the Latin alphabet,” but it can only include “a” Cyrillic alphabet, as there are nearly as many Cyrillics as there are languages.
The development of Whitney Cyrillic began in 2005, with a year-long research project to survey the effects of language, culture, politics, and technology upon Cyrillic typography. Beyond including the characters needed to render the more widely supported Balto-Slavic languages (Russian, Ukrainian, Serbian, Belarusian, Bulgarian and Macedonian), a primary goal of the project was to reach large populations that remain underserved by high-quality digital fonts. In some cases, small adjustments to the character set helped reach significant communities (adding just two characters adds Uzbek to the list of supported languages, with its twenty million speakers), though many of the languages representing these large constituencies were poorly documented, and many sources offered conflicting opinions. In studying examples of Cyrillic typography and lettering, preference was given to those designs created by native speakers, who were among those least likely to succumb to the many “false correspondences” in the Cyrillic alphabet. Typefaces in which the Cyrillic letters be (б) and ze (З) resemble the figures 6 and 3 reveal an insensitivity to Cyrillic form, and are nearly always the work of a Western hand.
Consulting with Hoefler & Co. on the project were two Cyrillists: Maxim Zhukov, former Typographic Coordinator to the United Nations, and Ilya Ruderman, creator of the Type & Typography program at the British Higher School of Art and Design in Moscow. Each brought a culturally enlightened perspective to the critique, and helped us to evaluate our work in the types of linguistic settings in which designers would be most likely to use the fonts.
The Cyrillic-X™ character set covers Abaza, Adyghe, Aghul, Archi, Avar, Azeri, Bashkir, Belarusian, Bosnian (Cyrillized), Bulgarian, Russia Buriat, Chechen (Cyrillized), Chuvash, Crimean Turkish, Dargwa, Dungan, Erzya, Ingush, Juhuri, Kabardian, Kalmyk, Karachay-Balkar, Kazakh, Khalkha (Cyrillized), Kirghiz, Komi, Komi-Permyak, Kumyk, Lak, Lezgi, Macedonian, Hill Mari, Meadow Mari, Moksha, Moldovan (Cyrillized), Mongolian, Montenegrin (Cyrillized), Nanai, Nogai, Old Russian, Ossetian, Russian, Rusyn, Rutul, Serbian (Cyrillized), Southern Altai, Tabassaran, Tajik, Tatar (Cyrillized), Tati, Tsakhur, Tundra Yukaghir, Turkmen (Cyrillized), Tuvin, Udi, Udmurt, Ukrainian, Uyghur (Cyrillized), and Uzbek.
Three of the most widely-spoken languages that use the Cyrillic alphabet exhibit regional differences in the shapes of their letters. Not merely stylistic variations, these are significant morphological alterations without which their host languages can look foreign.
Nearly a quarter of the Russian Cyrillic alphabet consists of roman and italic letters that are visually dissimilar. Much as the Latin letter a traditionally follows two different forms (two-storey in the roman, one-storey in the italic), sixteen Cyrillic letters have roman and italic shapes that seem unrelated: the Russian Cyrillic т is italicized as m, the и as u. Many Bulgarian readers prefer greater consistency among their romans and italics, and choose to import a number of cursive italic constructions back into the upright roman. It remains an open question whether this regional style will survive the proliferation of International Cyrillic, but with Bulgaria’s admission to the EU in 2007, the preferences of Bulgarian readers have a renewed significance.
Like Bulgarian, the Cyrillics used in Serbia and Macedonia feature a handful of localized characters in both their roman and italic alphabets. (An interesting challenge in the design of Serbian Cyrillic is to maintain its distance from the Latin alphabet, as Serbian is one of the few languages that can use either script, a phenomenon linguists call “syncrhonic digraphia.”) To automate these localizations, Whitney employs an OpenType property known as Language System Tags. In applications that allow a language to be specified for text, Bulgarian text (tagged BGR) will use variants of the roman characters Л, Ф, в, г, д, ж, и, й, к, л, п, т, ц, ш, щ, and ю, and the italic characters Л, Ф, в, д, ж, к, and ю. Serbian or Macedonian text (tagged SRB or MKD) will use variants of the roman character б, and the italic characters б, г, д, п, т and ш.