Punctuation: one space or two, sir?

(...and the optical scaling of fonts)

Well, I wasn't planning to write anything on this topic! But a short discussion at a recent meeting of the Marrickville Writers Corner writing group alerted me to the fact that I'm out of step with accepted practice (as far as published style guides go), on the topic of how much space to leave between sentences. I use “two spaces”; but some time between the last 30 to 50 years the modern convention changed to use a single space.

The especially interesting thing is that the common explanation and justification for this change is based on a myth: the claim is that the use of two spaces was a convention adopted due to the rise of the typewriter and mono-spaced fonts, and that with proportional fonts this archaic practice is no longer necessary. People often also claim that typefaces have space built in to provide the correct spacing after periods, question marks, exclamation marks, and quotation marks. (This is clearly false because that simply can't work for many common uses of periods such as in abbreviations, nor for the use of single-quote characters that are being used as an apostrophe.) Nor does it actually make sense: the typeface can't know whether a punctuation mark is being used at the end of a sentence. Another argument offered up is that the additional space is disturbing and introduces ugly extra gaps. To my mind, this flies in the face of cognitive science. There is a bigger semantic gap between sentences than between words: it's the end of one idea and the start of another. Words within a sentence, in contrast, are strongly connected and dependent: the meaning accumulates from one word to the next, like steps on a stairway. This is reflected when you read aloud: you pause longer between sentences than you do between words. And the white space on the page should reflect that aspect of the semantics. Which is why the use of a bigger gap was the preferred style until something like 30 – 50 years ago.

Anyway, all these myths are soundly debunked in this well-researched article here: Why two spaces after a period isn't wrong (or, the lies typographers tell about history). (If that link doesn't work, copy and paste this: https://web.archive.org/web/20171207185025/http://www.heracliteanriver.com/?p=324)

In fact, the most convincing explanation I've seen for how the practice developed is described in the article Sentence Spacing, which describes how linotype machines used “spacing bands” to define spaces, so the typesetter could adjust all the spaces within a line by adjusting the spacing control for the whole line. But a mechanical limitation of the linotype machine meant that if the typesetter put two spacing bands side-by-side, the system failed, with results which varied from having to remove the extra spacing band(s), to physically damaging the machine.

Prior to this, and for hundreds of years, people routinely used, recommended and preferred wider spaces between sentences, specifically to ease the reader's burden in seeing the ends of sentences. This practice continued while photo-typesetters were in common use, which was during the same period as the heyday of typewriters .

Another argument has been that one space between sentences is a cost-saving measure, making the text more compact to save on space, and therefore reduce printing costs. I just did a little experiment: in my 528 page book, if I replace all double spaces by single spaces, it shrinks to 524 pages. When you consider that books tend to be bound in sheaves, not only is that a mere 0.75% saving, but in real terms there's a 3 in 4 chance it will translate to no saving at all. So that argument would only apply if publishers really are trying to cut costs to the bone.

You'll notice that the argument also makes no sense whatsoever for epublishing.

I'm also quite sure that the line length would affect the desirability of more spaces between sentences, too, and whether flush-justification or ragged-right justification is being used, as well as depending on the font itself. In a multi-column format where the word spacing varies more from line to line, and especially if you're using flush-justification where inter-word spacing will vary from line to line (shudder), or in fonts where the letters are loosely spaced rather than tightly packed, the desirability of more space between sentences increases, to help make them visible.

Several articles I read while looking into this topic also mentioned, in essence, “Modern typographers prefer a single space between sentences, and they should know.” Well, modern typographers are wonderfully skilled people, but as for holding them up as the arbiters of good typographic style: in my view, on this topic they've been swallowed up by a fad that started as a result of the enormous pain caused to a typographer who accidentally put in double spaces in early linotype machines. I strongly believe it became part of the indoctrination that older members of the “guild” passed on to younger members, until the reason for doing so became lost in history and it simply became accepted as gospel.

Now, for those who may be reading this and averring that such a fad is unlikely to take hold, and that we should simply trust typographers to the arbiters of aesthetic judgements about beauty in typography, let me share a little-known fact I learned in late 2014. I offer it up as further evidence...

Most modern typographers are perfectly satisfied with scalable vector fonts, unaware that their pre-computer ancestors would have been horrified at the loss of the optical scaling of fonts. The idea that you can just geometrically scale a font up to whatever size you need, without taking account of how the letter shapes and stems should change to preserve the essence of the font design and maintain both readability and aesthetics, is a horror story that we're all, unconsciously suffering through. In the good old days, fonts were designed for maximum beauty and legibility for each specific font size. The old-time craftsman would be shocked that their art has been trodden so far into the dirt. If you can find a copy of Harry Carter's 1937 article on optical scaling, it's well worth the read. (An article that discusses Carter's article is called From the Optical Scale to Optical Scaling.)

Yet the number of modern typographers who are even aware of this topic is minuscule; and no computer operating system and no font technology vendor (especially Adobe), have algorithms and software in place to address the problem and achieve the quality that the old-timers routinely achieved. Here is a paper on some work done on it 20 years ago, which has since languished: Dynamic optical scaling and variable sized characters, and 28 years ago A model for automatic optical scaling of type designs for conventional and digital technology, but these papers have been largely ignored, and people have simply come to accept the poor quality as all they can expect – exactly as they have come to accept the use of the same gap between sentences as is used between words, I argue. I think what happened is that at the time the problem was noted, the solutions were too costly for the computing power available at the time, and since then people have simply come to accept the lower quality, because the effect is subtle.

It's interesting how perceptions and attitudes shift over the years. It's still my hope, though, that the older and superior quality might yet make a return to fashion. And I feel that this blog post is something small which I can contribute to the topic.