Type designers add ligatures because two letters keep colliding. Brand designers turn them on because they look elegant. The right question is: why are those two letters colliding in the first place?
Open the OpenType panel in any modern type tool and you'll find a checkbox labelled Standard Ligatures, usually on by default, and another labelled Discretionary Ligatures, usually off. The first one swaps in fancier joined-up forms for letter pairs like fi, fl, ffi, ffl. The second one does the same for showier combinations — ct, st, sp, Th, Qu — that some typefaces include for editorial flair.
Designers tend to feel virtuous when they turn ligatures on. The thinking goes: the type designer drew these. The type designer is smarter than I am. Therefore I should use them. The result is a paragraph that's nominally more elegant and, if you're being honest, sometimes worse than it would have been with the ligatures off.
The reason ligatures exist at all is illuminating, and once you understand it, you can decide for yourself which to turn on and which to leave alone.
In metal type, the f had a problem. Specifically: in most serif typefaces, the top of the f curves over to the right and ends in a small ball or terminal that hangs out beyond the body of the letter. When the next letter is an i, the dot of the i is sitting right where the f's terminal wants to be. They collide.
The metal-type solution was to cast the f and the i as a single sort — one piece of metal, with the f's terminal and the i's dot merged into a unified shape. That's a ligature. Same thing for fl, ffi, ffl, and a few others. The ligature was a workaround for a physical limitation of metal type: you couldn't kern past zero, you couldn't overlap two sorts, so you had to redraw the pair as one shape.
Digital type doesn't have that limitation. You can kern f against i as much as you want, including negative kerning that makes the two letters overlap. Modern OpenType engines can apply the kern automatically. So the original problem ligatures solved — the physical impossibility of two metal sorts occupying the same horizontal space — doesn't exist anymore.
Type designers continue to draw ligatures, and OpenType continues to substitute them, for two reasons:
Backwards compatibility. Readers are used to the look of the fi ligature. A book set without it looks slightly off to anyone who's been reading books for thirty years. The expectation is so deep that breaking it costs the typeface credibility, even though most readers couldn't articulate what's different.
Quality control. Even with negative kerning, an f-i pair in some typefaces looks awkward. The two shapes are doing different things at the top — the f is leaning right, the i is standing straight — and getting them to compose nicely takes care. The ligature is the type designer doing that care once, in advance, so the layout designer doesn't have to do it every time.
Both reasons are good. They're also both reasons rooted in a specific problem with a specific letter pair in a specific kind of typeface. Standard ligatures (the fi, fl family) are the type designer fixing a known issue. They're almost always worth keeping on.
Discretionary ligatures are a different beast. These are the ct, st, sp, Th, Qu, &c. that some typefaces — usually editorial serifs aiming at a vintage flavor — ship with the OpenType feature off by default. They are not solving a kerning problem. They are decorative. The ct ligature, where the c's terminal swoops up to meet the t's crossbar, doesn't exist because c and t collide. They don't collide. The ligature exists because it looks like 18th-century printing.
Decorative ligatures are fine in their place. The place they're fine is short, display-sized, occasional. A book title page. A pull-quote. A wordmark. A magazine subhead. Any place the reader is going to encounter the pair in isolation and read it as a flourish.
The place they're not fine is body text. If you turn discretionary ligatures on for a 200-page book, the reader is going to see the ct and st combinations dozens of times per page, and they're going to start to feel the typeface is showing off. The flourish that was charming in a chapter title becomes mannered when it's repeating in every line.
When you're about to enable a ligature feature, ask: am I solving a problem or am I adding a flourish?
If you're solving a problem — fi, fl, ffi, ffl in a serif body text — turn it on. The type designer drew the ligature to fix something that would otherwise be wrong. Use the fix.
If you're adding a flourish — ct, st, sp, Th, Qu, ornate alternates — ask whether the flourish belongs in the context. Display? Probably yes. Body? Probably no. Wordmark? Look at the specific pair carefully — sometimes the discretionary ligature reads better as a wordmark detail than the standard pair, sometimes it reads worse.
Here's the part that gets overlooked. A lot of designers reach for ligatures as a way to fix something that's not actually a ligature problem. The pair looks bad — the spacing is wrong, or the typeface doesn't have great kerning, or the size is too small for the typeface's default spacing — and the designer turns on ligatures hoping it'll help. Sometimes it does, by accident. More often, the underlying problem is the spacing, and turning on a ligature is a costume that disguises the issue without fixing it.
If a word looks wrong, look at it before you reach for the ligature toggle. Does it need a kern? A different size? A different cut of the typeface (text vs display)? Try those first. If the pair still wants more help, then bring in the ligature. The ligature is a final-mile fix, not a first-pass diagnostic tool.
The places where ligatures genuinely make a typeface better are quieter than the showy discretionary set:
None of those are decorative. They're all problem-solving. That's the line.
Type tools make every ligature feature feel like a switch — on or off, no consequences. The right way to think about it is one feature at a time, in context, with the question: is this making the reader's life easier, or is this making my work look fancier?
The first is what type designers spent decades drawing for. The second is a perfectly fine instinct in moderation, and a tedious one when overused. Most ligatures are kerning problems wearing a costume. The ones worth turning on are the ones where the costume is structural — where the costume is the thing the type designer drew the costume to be.
For everything else, use restraint. The cleanest body text in the world has the standard ligatures on, the discretionary ligatures off, and the spacing dialed in so well that nothing else has to compensate for it.
Be the first to mark this page