dot-font: The Justification for Hyphenation

Hyphenation is a big subject. We may think of it as only pertaining to grammar or language, but hyphenation operates on an aesthetic level, too, as any typographer can tell you. And how well it’s done affects how easy our text is to read.

A typographer friend of mine was recently wrestling with the fine points of typesetting a document, and he was getting a lot of differing opinions about how to do it. He was designing and formatting this document in the United States, but it was going to be published in the UK and read internationally. All the text was typeset flush left, ragged right. He was taking flak for having hyphenated the word “typography” after the “g” rather than the “o”: “typog-raphy” instead of “typo-graphy.”

Of such fine points is typography made. When they’re handled correctly, no one but a specialist ever notices, and readers blithely and happily read along without a thought for the typesetting. If it’s done wrong — then they notice.

Perhaps some of my reply to my friend will prove useful to you, too. Here’s what I told him.

Dear Typographer:
“Should I adjust my QuarkXPress hyphenation settings,” you ask, “to ‘Auto Hyphenation > Smallest Word: 6; Minimum Before: 3; Minimum After: 3’? That might give me a badly ragged right edge of the column, and maybe (God forbid!) some ugly final paragraph lines (too short, like one-word only) but if that results in better word breaks…”

The quick answer about “typography” is that it differs between the United States and the United Kingdom (and their respective appendages and spheres of linguistic influence): U.S. usage is to break words according to pronunciation, while U.K. usage is to break them by etymology. I’ve always thought “typog-raphy” looked odd, but it’s perfectly consistent with the “standard” of U.S. hyphenation. And it does have the virtue of telling you, before you reach the second line, that the word “typography” isn’t going to turn out to be “typographic.”

As for how many letters can be left hanging, before or after a hyphen, I resist hard-and-fast rules. Yes, it’s generally a good thing to avoid two-letter fragments before a hyphen, but when those two letters constitute a prefix, it’s often the best way to break the word. (This is especially true for words like “reinforce” and “reinvent”. If you break after “rein-” then it looks like the word “rein.” The most important thing about hyphenation — absolutely the most important — is not to be misleading.) I do try never to leave two-letter fragments after the hyphen, even if they make linguistic sense, though in justified newspaper columns I’d ignore that stricture. (Anything to help newspapers not look so damn ugly!)

How Does it Look?
My suggestions for running down various syllables or words to the next line were based on how the lines looked. I’d be perfectly happy with “re-” at the end of a short line (in ragged right text), but I would run it down to the next line if it stuck out beyond the lines above or below. Similarly, I usually try to run the word “I” down if it sticks out alone at the end of a line (especially if it follows a period). This is all a matter of aesthetics, basically. But that’s the only reason for any of the two- and three-letter rules, or the avoidance of hyphen stacks. (Sometimes hyphen stacks are preferable to the alternatives. I always leave the “hyphens in a row” limit set to “unlimited.”)

Short words or word fragments also look better or worse depending on just what the letters are, which determines how long the word or fragment really is. At the end of a paragraph, what looks good depends partly on how wide the indent is for the following paragraph. It’s silly to forbid breaking the final word in a paragraph if that word is “singlemindedness”; you might not want to leave “-ness” all by itself, but who could fault “-mindedness” as the last line of a paragraph? It’s a longish compound word in its own right.

I don’t know QuarkXPress’s hyphenation algorithms as well as I know PageMaker’s, but no matter what the program, you always have to adjust some line breaks by hand. The best you can hope for is a setting that automatically gets you as far as possible along the road to quality. I usually prefer to allow shorter fragments and more hyphens, because it’s easier to force a line break by hand than to re-wrap the text to get more on one line. PageMaker’s “hyphenation zone” works quite differently from Quark’s; this is one area where Quark’s programming makes more sense, but I’m used to PageMaker’s.

Justifying Yourself
One thing that makes PageMaker superior to QuarkXPress in terms of fine typography is that, in justified text, PageMaker’s hyphenation & justification algorithm will not try letterspacing (or squeezing) until all other options are exhausted. Quark tries adjusting the letterspacing much too early in the process, with the result that it’s almost impossible to avoid some changes in letterspacing in narrow columns. In PageMaker, if you tell it never to letterspace, it will never letterspace (unless there’s only a single word on the line).

How Ragged Should the Edge Be?
I noticed that you’re using only a very slight rag in these pages. I tend to prefer a looser rag, but it’s a purely aesthetic judgment. (And I suspect that a tighter rag works better with the large paragraph indents — there’s not so much of the effect of raggedness on both sides of the text column.) One preference I have — which no program takes into consideration, as far as I know — is that I dislike seeing the first line of a paragraph, article, or column of ragged-right text turn out longer than the lines under it. I always try to keep the first line a little short (and the last line, but that usually takes care of itself).

What’s So “Proper” About Names?
Oh yes, one more thing that one of your editors brought up: hyphenating proper names. I think this rule is nonsense — unless the names are unfamiliar. Sure, if it doesn’t create a typographic awkwardness, I’d try not to break someone’s name the first time it appears, but after that it’s fair game. And if the name is something like “Johnson,” I wouldn’t hesitate for a second to break it as “John-son,” since any English-speaking reader will immediately recognize the name. A name like “Cusihuamán” or “Finnbogadottir” or “Ashurbanipal,” on the other hand, I’d try very hard to avoid breaking on first occurrence, because they’re unfamiliar to most English readers; breaking them would make them harder for the reader to assimilate (even though, say, “Finnbogadottir” does have a logical hyphenation point — recognizable to English speakers if they realize that “dottir” = “daughter”).

As for product names and company names — break away! They don’t deserve the same consideration as people. And if they insist on having internal caps (like PageMaker, not to mention the orthographically weird QuarkXPress), then the only logical place to break them is before the internal cap. Yet most of these programs’ own hyphenation algorithms (or dictionaries) don’t break them there! This is loony. Right up there with breaking in the middle of one word of a compound, instead of between the two parts of the compound. (Some programs would hyphenate “pagemaker” as “pagemak-er”!) Automation is tricky.

Rules Are Just Guidelines
This has been a long answer to a simple question. But it’s an issue that I pay a lot of attention to. In the end, the answer is always: Use your common sense. (And in this particular publication, don’t forget — you’re the designer, as well as the typesetter. It’s your call. These are typographic questions, not editorial questions.)

See you in Leipzig?
John

Tags
Posted on: September 14, 2000

4 Comments on dot-font: The Justification for Hyphenation

  1. First, it is ironic that an article on hyphenation and typography contains very bad line breaks due to the total lack of hyphenation. Second, anyone calling themselves a typographer should not have auto hyphenation turned on. This article contained some very poor advice.

  2. In response to VF – re-size your browser window, man – it’s scaleable, so the author clearly has no control of hyphenation in the article. Duh.
    Perhaps I’m not the “true typographer” that VF’s claiming to be, but I found much to like in this article. My only quibble is the inadequate discussion of auto-hyphen algorythms, a subject that I find fascinating (so what does that make me?)

    Maybe next time, Mr. Berry?

    “Long” John Holmes

  3. I did find the historical tidbit interesting. The 2 repsponses, however, comment on the article’s hyphenation. There certainly were “widows.” I thought that if you just wanted to flow text, there is no way around that in HTML (aside from forcing a set column width and breaking each line by hand – which is not how creativepro.com is set up). If there is another way, I’m eagerly awaiting enlightenment.

  4. Dealing with 2 languages at once (French & English) makes hyphenating even more puzzling. Have you looked at the typesetting features of InDesign and what do you think about them??

Leave a comment

Your email address will not be published.


*