: When designing a latin-based typeface, how are diacritics handled? I was looking into designing a font that could include characters with diacritics but outside of knowing any other language besides

@Hamaas979

Some background on me, so you can estimate how much or little authority I have: My native language (German) uses diacritics (ÄÖÜäöü) as well as non-diacritical special characters (ß) and is in the process of introducing or rejecting a new special character (ẞ, the capital eszett) right now. I did some research on the diacritical characters myself for the purpose of including them into a blackletter font, which could not easily be done by looking at existing fonts, due to the differences between blackletter and roman type.


Since the modification of a diacritic can mean another meaning for the character is there a tolerance allowed or does that come down to having to know how to write that language? Are there some languages that do not allow any alteration to the diacritic?


Most diacritic letters are used by more than one language and in fact many languages borrowed diacritic letters from other languages when they were in need for one. Given the globalisation of typography, it thus makes sense that a given diacritical character looks the same in every language that uses it.¹

Thus many users of languages with diacritical characters will use realisations of these characters designed by somebody not speaking their language on a daily basis. As a native speaker of such a language, I very rarely encounter badly designed special characters – and if I do, they often occur in generally horrible fonts and common sense should have sufficed to avoid the mistake (e.g., do not place the two dots of ä on the same height as those of Ä). I have encountered wrong ogoneks (as in ȩ) though, despite not speaking any language that employs them – I attributed most of them to utter laziness however, as they just recycled the cedilla (as in ç).

I am not aware of any characters or languages where a small change to the character would change the meaning.² Such a “feature” would probably not survive long, as it does not allow for an easy distinction when reading and it would require an excessive effort to distinguish the characters when writing them by hand.³ Most languages have some restrictions and preferences regarding the appearance of their diacritical marks, however.

That being said, it is generally a good idea to inform oneself about the diacritics of the languages one would like to cover and not wildly start designing them. A few examples:


Replacing the umlaut dots (e.g., in ö) with short, tilted strokes, is acceptable in German. In Hungarian, however, there is the double acute accent, which could be confused with this. Compare ö with ő.
When your font is not suitable for all-caps anyway (e.g., a blackletter font), there is no need provide capital variants of certain letters that appear only within words and thus are never capitalised if not for all-caps. Examples include ȩ, ß, ð and ÿ (with the probable exception of some very exotic alphabets). Something similar applies to kerning pairs, e.g., you do not need to worry about fß, as no language uses this letter combination.
While I am aware of no language that employs more than one diacritical mark below the letter, you should not design those diacritical marks identically⁴. There is a difference between the diacritical marks of:


ș, ķ (Romanian, Latvian),
ȩ, ų (Polish, Lithuanian)
ç, ş (French, Turkish, Finnish and many other languges).


In particular for Romanian, things went horribly wrong regarding this in the past.
Diacritical marks may change their appearance upon captialisation. For example, ď → Ď or ģ → Ģ.




¹ The only exception I am aware of is the acute accent in Polish, which should have a certain steepness, whereas there is no such requirement in other languages, as claimed by here and here. However, these sources also state that most or at least some fonts have sufficiently steep acute accents anyway.
² There are some examples in the International Phonetic Alphabet, though.
³ There are alphabets, however, who have strong variations between printed and handwritten characters, such as Cyrillic or Polish for some of its diacritical characters.
⁴ With the possible exception of minimalistic pixel fonts and similar.


Do charts exist for some language characters?



The Diacritics Project aims at providing resources for font designers who want to design diacritical characters “correctly”. It gives extensive informations on what needs to be considered for some diacritics and guidelines on designing them. It does not cover non-diacritical special characters though.
Adam Twardoch gives extensive information on the Polish diacritics and also contributed to the above.
This site gives instructions on how to design ð, Ð, þ and Þ – special characters used by Icelandic and Faroese.
This article on Typography.Guru by a German typographer explains how to design the eszett (ß); this article explains the capital ẞ.
This article provides by the type foundry Underware explains what you need to take into account when typesetting the Dutch IJ and ij.
Vietnamesetypography.com gives an extensive introduction to the special characters of the Vietnamese language.
This article from a Norwegian type designer elaborates on the design of Æ, æ, and a bit of œ.

Vote Up Vote Down