The Third Way

If you’ve been paying attention lately you’ve probably noticed the raging dispute concerning how many spaces are appropriate at the end of a sentence. If you learned to type on an actual typewriter you were taught the hard and fast rule that the correct answer is two. If you’re a typography geek you almost certainly believe—equally strongly—that the correct answer is one.1

The typography geeks claim that using two spaces evolved because of monospaced typewriter script but that it doesn’t make any sense with proportional fonts and that, in fact, typographers have never adhered to the two space rule. Most people appear to accept that explanation even if they insist on two spaces out of a long ingrained habit.

You’d think that would be the end of it but the two-space partisans have pushed back challenging the conventional history and claiming that the one space rule was all about publishers trying to save some paper. According to this telling of the story, the real historical practice was to have slightly larger spaces between sentences and that the one space rule is less than 60 years old.

The current state of the battle is neatly captured by xkcd is this amusing cartoon


As it turns out, the third way actually exists. In the Troff typesetting system it’s traditional in preparing the input to end each sentence with a period followed by a newline. In Groff, this allows one to specify how much space to put between sentences. For example, both my books were typeset with 1.5 spaces between sentences. If the production people had complained too loudly (they really do like one space), I could have simply adjusted a single parameter to get the desired spacing. Thus, at least in some circumstances, the third way is the best way.



Even these conventions are not as rigid as some believe. See the Wikipedia sentence spacing article on the history of sentence spacing practice. The French, for example, favored a single space even in typewritten text and up until the late 1990s many printers used 1.5 spaces. Although partisans on both sides are emphatic that their choice renders text more legible, scientific studies don’t support this.

This entry was posted in General and tagged . Bookmark the permalink.

3 Responses to The Third Way

  1. mbork says:

    Wow. So much to say. (Even before I read the texts you linked to.)

    1. I am strongly in the “two-space” camp – for TeX (which I use), two spaces are (usually) equivalent to one anyway, and using two spaces enables me to use Emacs’ sentence-related commands (M-a, M-e, C-x DEL etc.).

    2. In TeX there’s also a distinction between a space after a sentence and any other space. And you don’t have to use newlines for that; the algorithm TeX uses to determine the required space width is actually quite smart, and can recognize sentence endings like “some remark.)”, with parentheses intervening between the period and the space. Of course, it is not 100% perfect – this would need semantic analysis of the text. Of course (this is TeX, after all!) it is pretty configurable.

    3. And “the third way” has actually quite a few proponents, and for a good reason: version control. If you stick to a one-sentence-per-line (or one-though-per-line) policy, the diffs suddenly make much more sense… (OTOH, M-q’ing the paragraph kills the usefulness of diffs entirely.)

    • jcs jcs says:

      How Emacs determines a sentence is, of course, configurable. I use one space and everything works fine.

      Most typesetting utilities—TeX, Troff, and even HTML—will do the right thing automatically. It’s only brain dead “word processors” that get things wrong. The real point of a new line after each sentence is to unambiguously determine sentence boundaries.

      You’re right about version control; I hadn’t thought of that.

      • mbork says:

        Obviously, you can configure sentence-end to whatever you wish (this is Emacs, after all!) – but I stick with the default and am quite happy with it, especially that – as I mentioned – TeX deals with it in a smart way. (Interestingly enough, it does not use regexen, but something called “space factor”, which is a clever trick.)

        PS. I noticed a typo in my previous comment – it should have been “one-thought-per-line”, of course.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>