Non-joining ligature?

jidoe's picture

I'm wondering if technically speaking a ligature must be two letters that are joined together, or, can you have a ligature in which there are 2 letters designed to work together but don't actually join?

PublishingMojo's picture

If the two letterforms are coded as one character, it's a ligature whether they visually touch or not, just as the original ligatures were two letters cast on one piece of metal. If they're alternate forms of single letters designed to fit together (e.g., a lowercase i with no dot, intended to form a kerning pair after a lowercase f or a capital T), they're not ligatures, technically speaking.

charles ellertson's picture

The word has been used in both contexts. Some called two letters cast together, that did not touch, a logotype, and reserved "ligature" for letter pairs that did touch. For example, Vo, Wo, To etc. were special sorts on the Linotype lincecaster, where kerning was not possible. Some people termed these "logotyopes," not ligatures, in distinction to fi, fl, etc. -- Adrian Wilson was one, IIRC. However, as advertising took on greater and greater significance in the world of type, company logos also became known as "logotypes."

As with so many things, the use of this term seems to have undergone a change, and most of us today would be comfortable with "ligature" even when the components did not touch. Use the terms as you prefer.

hrant's picture

Also note that "logotype" could mean multiple (not just two) letters cast as one sort, not for aesthetic reasons, but to speed up composition. See my first post here: http://typophile.com/node/19376?page=1

hhp

Rob O. Font's picture

Unless sizing, tracking, or H&J are important to the typography, the difference between logos and ligatures is not. In the red zone however, if all three of those things are playing on glyph selection at once, something's going to want to know the difference, trust me.

hrant's picture

H&J: Software should be able to decompose.
Tracking: Indeed relatively tricky. Although a text font shouldn't be anyway.
Sizing: What is that referring to?

hhp

Martin Silvertant's picture

Wouldn't a letter pair which doesn't touch be considered either a digraph or a contextual ligature these days?

charles ellertson's picture

No to both.

A contextual ligature would be a ligature used only in certain situations -- effectively, when following or preceding certain other letters. I suppose the context could be size -- use to be, you didn't use the f-ligatures above about 14-point. Anyway, nothing to say about whether or not the letterforms touch.

From Wiktipedia:

A digraph or digram is a pair of characters used to write one phoneme or a sequence of phonemes that does not correspond to the normal values of the two characters combined

hrant's picture

There are other –and in the context of type design, better– definitions of those.

hhp

Martin Silvertant's picture

I meant either a digraph or a contextual ligature, not both. I can see now that contextual ligatures are too specific and don't really have any relevance here. Still, I would like to clarify that I was thinking of contextual ligatures for improved spacing, not these elaborate ligatures you find in script fonts.

I don't think I'm necessarily wrong about the digraph though. I know what it is, but usually it's a letter pair which doesn't touch. It's certainly not encompassing as I initially foolishly alluded to, but it's one of the things a letter pair can be without being a ligature.

Thomas Phinney's picture

I'm not so sure, Martin. My understanding is that a digraph is a linguistic thing first, and a typographic thing second. It exists regardless of whether it gets special typographic treatment, and when it does get special typographic treatment, that isn't usually modifying one or more of two separate glyphs.

charles ellertson's picture

As long as you want to get technical about it, contextual ligatures (clig) are a GPOS lookup type 8, as opposed to the more common GPOS lookup type 4, used in ligatures (liga). (Maybe that's the "better definitions in the context of type" hhp meant? Hard to say -- often snide, rarely helpful, that hhp...)

In any case, "touching" or "joining" has nothing to do with the difference.

https://www.microsoft.com/typography/otspec/features_ae.htm#clig

BTW -- Check the order of processing if you use cligs...

http://partners.adobe.com/public/developer/opentype/index_table_formats1...

hrant's picture

I'm only snide with the hopelessly obtuse. Hey, it's the only fun I have.

hhp

charles ellertson's picture

Hey, it's the only fun I have.

I can believe that (he said, snidely).

Martin Silvertant's picture

I'm not so sure, Martin. My understanding is that a digraph is a linguistic thing first, and a typographic thing second. It exists regardless of whether it gets special typographic treatment, and when it does get special typographic treatment, that isn't usually modifying one or more of two separate glyphs.

I have to agree.

simonemartijn's picture

This looks a whole lot better to me.

Syndicate content Syndicate content