Friday, September 30, 2011

The Em-Dash and You

John McIntyre, newspaper copyeditor and proprietor of You Don't Say1 recently had some much needed advice for all of us writing online: Know when to use the em-dash.

"Em-dash" is not, as you may believe, a short-distance foot race of some kind or another. It's the long horizontal mark we use to indicate a parenthetical remark — like this one — in our text. It's also used at times to indicate dialogue:

— You mean like this?

— Yes, exactly.

Note that it is a long dash. It is not the minus-sign we use for compound words (like in "minus-sign"). Why not use the minus-sign? Because the text looks cleaner and more professional, and is easier to read with the wide dash. There really is no reason not to use it. But how do you write it when it doesn't seem to actually be on our keyboards? fear not: it's easy to write. The Wikipedia page has instructions for many systems and programs, so check it out.

Under Linux you press the Compose key, then '---'. "Compose" key? On full-size keyboards it is often the right-hand "Win" key or right-hand Alt, but you can change which key is "Compose" in your keyboard layout settings. I always set Compose to be my Caps Lock key; I never, ever actually need Caps Lock, and hitting it by mistake is annoying, so I much prefer it to work as Compose.

While you're at it, it's time you also started using ellipsis '…' rather than simply write three full stops '...'. Depending on what font you use they may look identical, but trust me, they're not. The true ellipsis is one character while the dots are three. This matters particularly when you use a fixed-width font, such as in text editors or when writing Japanese. You write it by pressing Compose, then '..'.

A couple of other useful Compose characters:

  • Degree symbol '°' as in "His fever was 39.2°": press Compose then 'oo'.

  • Cross product or cross mark '×', as in "1024×768": Compose then 'xx'.

  • Long o 'ō' as when you write long-o Japanese words with alphabet, as in "Ōsaka": press Compose then '_' (that is shift+'-'), then 'o' or 'O'.

  • Copyright symbol '©' and trademark symbol '™', as in "Navelicious© — The Go-to Website™ For Your Bellybutton Needs!": Compose then 'oc', and Compose then 'tm'.

  • In case you have no Swedish vowels in your keyboard, you can still write 'å' 'ä' and 'ö': Compose then 'o' then 'a', and Compose then '"' then 'a' or 'o'.

Those are not the only characters of course, and you have many, many more if you use the "AltGr" key (may be something else on a laptop keyboard), such as '·', '÷', 'µ' and so on. It's worth looking up these key combinations for your keyboard and system; it's surprisingly useful.

And use Em-dash, OK? Give the poor minus-sign a rest.

#1 Sadly, while this is the personal blog of McIntyre it is hosted with his employer The Baltimore Sun, who in turn has decided to lock all its content behind a paywall beginning October 10. Since local news is of no interest for people like me living on a separate continent most readers are faced with the prospect of paying over 8000 yen per year for a single blog. Which, as you can expect, few to no people are ready to do no matter how good the content.

So unfortunately one of my favourite blogs will effectively disappear from the net beginning early next month. Do take the chance to browse through his archives while you have the chance; there's a lot of good stuff there to read.


Jonas said...

There's also the en-dash of course. This great article on typographic etiquette at Smashing Magazine recommends to use that instead of em-dash. I'm not sure which is better etiquette, but personally I find the en-dash more appealing.

That article also mentions that hyphen and minus are actually two different characters -- so we have a total of 4 different dashes.

By the way, I'm envious of your Compose function in Linux -- I didn't know about that before. I've trained myself to remember the Windows Alt-codes for the characters I use the most, such as Alt-248 for °, Alt-224 for α and so on. Pretty ridiculous in comparison with Compose!

Jonas said...

Okay, perhaps I should have included the link to the article:

Janne Morén said...

Yes, I skipped en-dash; the normal rules mostly have them for date or page ranges (like 2008–2010). Also, I find that in most common typefaces they are so similar in length to the minus-sign/hyphen that they can be hard to even distinguish. I much prefer the cleaner break of using em-dash, with spaces (tends to look cramped without spaces) over the en-dash.

If you do want to write en-dash under Linux, you press Compose, then '--.', sort of indicating a shorter-length dash than the em-dashes three hyphens. And yes, Compose is really, really useful and I use it quite a lot. Surprised that something similar is not available — even as an add-on or plugin — for Windows.

Richard said...

The use of the em-dash---like this, without spaces around the dashes---is very much an American thing, and is in my opinion a typographical train wreck. As the article that Jonas linked to said, it is too disruptive. I feel that it ruins the balance of the text and leads your eyes to focus on the wrong places.

The European tradition is to use an en-dash -- which is usually the same character as minus in most fonts, but not the same as a hyphen, which is shorter -- surrounded by space. In Swedish typography, a minus used to indicate the start of a quote is known as "pratminus", and as separator in text (as used in this paragraph), "tankstreck". A hyphen is called "divis".

Another ugly typographical idiom that the use of TeX has spread over the globe is the insertion of additional space after a period to separate sentences. ("English spacing".) This is no longer considered good form, as the linked article also mentions, and you can use the command \frenchspacing to disable it.

I strongly recommend Christer Hellmark's little book "Typografisk Handbok" if you can get hold of it; it's excellent.

Janne Morén said...

Richard, I very much agree that dashes need spaces around them. I disagree with using en-dash, however, as it will look the same as a hyphen in online typography and have too little weight compared to the characters it surrounds.

And I believe you're wrong about the extra space and TeX. Note how Jonas' linked article (and pretty much any other source discussing this) argues against an extra _full_ space after a period; it completely agrees with having a wider space after a full stop than between any two words. This is not a TeX invention, but something you see in typeset material everywhere, from long before computerized typesetting.

It's also easy to forget — typography, like grammar, tends to bring out the strongest of opinions out of people — but these things are all but gentle recommendations, subject to the whims and tastes of the author and publisher, and liable to change with the medium, the typeface, the content and the desired visual impact.

Richard said...

Not to start a war here, but i think combining em-dashes with spacing is overkill -- those long lines stand out too much and draw too much attention, disturbing the reader's flow. And in most fonts, en-dashes really are distinct enough from hyphens. What McIntyre said about using en-dashes to link ranges is correct, but then he says "you probably won't have much call for it", which might make sense to an American who uses Strunk & White as guideline, but not to anyone else. (And don't get me started on what I think about S&W...)

I didn't mean that TeX invented the extra spacing after a sentence, just that it took a specifically English tradition (which was already getting deprecated among typographers) and put it in the hands of many thousands of computer enthusiasts as the default and preferred way. is a good read. In general, you can say that with modern proportional fonts, the space created by the period itself followed by a normal word space is quite enough. Adding more space makes the text look full of holes. (Which is what the article that Jonas linked to says.)

But these are mere quibbles compared to the worst crime against readability that the web has made people consider to be the normal state of things: line width. <petpeeve>For instance, the Baltimore Sun article you linked to uses a fixed line width in terms of characters per line. No matter how you resize your browser window, or increase the font size with Ctrl-+, the lines contain about 100 characters each! In order to be easily readable, lines should not be more than 70 characters long, and preferably even shorter, even as short as 50 characters. Above 70 characters, you get serious eye tracking problems when moving from the end of one line to the start of the next. But these days, we seem to have accepted that online, we get lines that stretch across our entire screen. (And even typical scientific papers in LaTeX are quite awful in this respect, unless they use 2-column mode.)</petpeeve> Is there any hope for web typography to recover?

Jonas said...

Typography certainly seems to stir some feelings :D

So, one thing LaTeX did get right in the defaults was the text block width at about 65 characters per line. Unfortunately that looks really terrible when printed on A4 paper which is way too wide – so often we go ahead and widen the text block anyway, in order not to waste too much forest.

> Is there any hope for web typography to recover?
It may be just a bias due to my increased interest in typography, but it seems to me that lately a lot of blogs and more design-oriented web sites have been trying to promote good layout and typography. And, needless to say, the web evolves at lightning speed, and as people consume more and more text here, I am quite sure that the trend will be towards easier reading. Just compare to the average web site 10 years ago, with full-width paragraphs set in Times New Roman on a strongly coloured background…

(I went and looked up the alt-code for ellipsis: 0133. Not completely sure if it was correct usage there, though… My punctuation has certainly been influenced by instant messenging!)

Janne Morén said...

It is a matter of taste to a large degree, as I said, but I do prefer the longer lines and spaces, precisely because it disrupts the flow to some extent*. It is meant to disrupt the reading after all, to indicate an aside or sidetrack, just like parentheses do. And how much it impacts on the visual impression of the text depends, of course, a lot on the precise font you're reading it in; something I as a content creator for this blog really can't and won't impose on the reader online. You could well say that if your choice of default typeface and weight makes em-dash and spaces too visually disruptive, then you may want to pick a new default typeface.

For all that I love TeX/LaTeX, I doubt it has had much of a global impact on any typographical conventions. It may be commonplace among people like us, but it is rare outside some fairly narrow niches. If you want to put the blame for the spread of any typographical peculiarities anywhere it is Word, followed by InDesign and friends that will have had any real-world impact on printed conventions; and Netscape and IE that's determined what online content should look like for many people.

The problem is, it's hard to get the line width right online. With todays wide screens, if you make a line around 65-70 characters wide it will be a thin string snaking down the middle of an empty page (or through a forest of banner ads). Increase the character size (if you even can do so) and people will complain it's too big. Like it or not, but the preferred default for online text seems to be closer to 90-100 characters today. And I find I certainly prefer to write and edit my own texts in much wider lines than what would be acceptable in printed formats.

I don't see A4 as a big problem in practice; you do want generous margins for many types of text, and you can use the extra room to increase the size a bit if you expect a fair number of elderly or otherwise visually impaired readers. And when space really is at a premium, the two-column format works really well for technical and scientific papers.

* By the way, McIntyre has expressed plenty of hate for S&W in multiple posts over the years.