1. Introduction

[NOTE: This topic has been treated much more extensively by Thomas A. Fine in his blog beginning at: http://widespacer.blogspot.com/2012/12/hello-world.html. I wasn't aware of his blog when I began writing this (I wish I had been), and hope merely that the evidence here contributes further to discrediting the astonishingly ignorant claim by so-called "typographers" of the late 20th and early 21st century that a single space after a period in typewritten work (who uses a typewriter any more?) is somehow the One True Way.]

The fashions for the set and spacing of type have changed over the decades, yet each generation firmly believes that its fashion is somehow objectively correct. This presents issues not only to the maker of new types, but also to the typefounder working with older matrices.

I'll look at this issue in two ways here: first as mirrored in the conventions of typing, and second as a practical matter in casting type.

(For more data on the evolution of the horizontal spacing of types see History of the Fit of Type.)

2. Spaces Between Sentences

I'll start with an anecdote and its explanation. It concerns the use of type rather than the making of individual types, but it does show quite clearly the direction of the change of fashion in the spacing of type.

Several years ago, a good friend of mine who has had recent training in the graphic arts as they are taught in the 21st century took exception to the fact that I had used two spaces between sentences in a plain ASCII text file. "Took exception" is perhaps overly mild - he said that it actually "hurt" to look at such an abomination.

So it would appear that now younger students in the graphic arts are being taught that it is correct to put one space after a period but wrong - very wrong - to put two. This is interesting on at least two counts.

First, there is an irony here in this emphasis on how many times you tap the spacebar when we have now entered an era where it doesn't matter. The spacing of text is now done by computers, and since the development of markup languages four decades ago by Goldfarb, it simply makes no difference how many spaces you type. The computer will change it for you. The only times you'll ever actually see two spaces if you type them are in (a) plain ASCII text files, or (b) typewritten documents. A young graphic designer is unlikely ever to see either of these.

Second, it would seem that this convention (one space) is being taught as if it were an objective truth. Yet I know exactly why I type two spaces: because it was the way I was taught. It was the way everyone in the US was taught who is old enough to have been taught typing on a manual typewriter. Then, just as now, it was an objective truth. Except the truth has changed.

This means, of course, that the number of spaces one must type between sentences is demonstrably not a matter of objective truth, but one of changing fashion. But why did it change? I wish to suggest that fashions in typing follow fashions in type; that it changed because the fashion for inter-sentence spacing in typeset text was once one thing and is now another.

As I have observed from my own experience as a student, the standard practice in the 20th century was to employ two spaces between sentences. To take a convenient example at random a book of professional typing instruction from 1903 has this to say:

"Space twice after a period, question mark, or exclamation point at the end of a sentence, and after a colon where it immediately precedes a sentence beginning with a capital letter." ( {Musick 1903}, p. 13/ PDF 32)

By 1903, the typewriter had become firmly established in business and typewriting had become a profession. The "correct," professional, inter-sentence spacing was twice that of the inter-word spacing. This is, I would argue, an approximation of the "correct" inter-sentence spacing for typeset text in that period.

Let's put this in the context of recommendations for composing typeset text from the late 19th to the late 20th centuries.

The typefounder Thomas MacKellar (one of the foremost authorities in 19th century practice) wrote:

"... What is commonly called the thick space is the proper separator between words, ...

"The comma requires only a thick space [after it], but the other points [that is, punctuation marks] should have a hair space before and an en quadrate after them, except the full-point, which should have an em quadate, as terminating a sentence." ( {MacKellar 1882}, pp. 130, 131).

MacKellar's "thick space" was a three-to-the-Em space, and of course an En quadrate was defined then just as an En quad is now, as two-to-the-Em ( {MacKellar 1882}, p. 77). It should also be noted that in MacKellar, as in later typesetting texts, the default spacing was always to be modified to result in even spacing for justified lines.

So MacKellar's advice, taken here as characteristic of 19th century American composition, is that the space between sentences is three times that between words. That is of course even more than the standard for typing, but in typing the inter-word space is typically wider, so two spaces between sentences in typed matter is a reasonable approximation of 19th century typesetting.

For a confirmation of MacKellar, we need only to turn to Theodore Low DeVinne, undoubtedly the finest printer in America in the 19th century. His office rulebook survives, and it defines the various practices acceptable to him for inter-word spacing. They are much more nuanced, but generally even wider than MacKellar's ( {DeVinne 1883/1928}, p. 11).

"The spacing of words must be governed by the leading. Solid matter and ten-to-pica leaded material, in ordinary faces of Roman and Italic, should be spaced with three-to-em spaces. ...

"Ordinary-leaded matter and poetry should be spaced with en-quadrats ...

"On double-thick leaded or on white-lined matter, two thick-spaces may be used.

"However widely leaded out any matter in regular roman body type may be, avoid the use of more than two thick spaces. ...

"Thin-faced, lean, and condensed letters call for close spacing; an expanded letter should always be widely spaced and widely leaded.

"Lines of round-faced capital letters, two-lines with little or not shoulder, and the lower case of all extended letter[s?], should have at least two thick spaces between words. The em-quad must be used as the space for very expanded letter[s?]" (p. 11)

Regrettably, his office manual does not specify inter-sentence spacing or spacing after the period, but an examination of books printed by his firm with which he certainly took some care indicates that his standard for inter-sentence spacing exceeded an Em-quad by some measure. (See {DeVinne 1900} for a good example.)

Moving forward into the early/mid 20th century, Polk's The Practice of Printing, still widely used as a basic printing text today, recommends slightly tighter spacing. I'll quote here from my 1945 edition, but I'd be surprised if his advice was different in the original 1926 edition:

"The 3-em space ... is the standard space used between words in a line, in ordinary composition..." (p. 34)

"... the compositor starts out with 3-em spaces between words." (p. 46)

"Use a 3-em space between words, and two 3-em spaces between sentences." (p. 45)

So by the middle of the 20th century, spacing is tightening up a bit: a 3-em space between words rather than an En-quad, and two 3-em spaces between sentences rather than an Em-quad. But the relative width of the two is the same - twice the space between sentences as between words. This matches 20th century typing practice exactly.

Later in the 20th century, spacing tightened up still more. Two examples from different editions of Craig's Designing with Type, separated by several decades, may suffice.

In the first edition in 1971, writing during the transition from metal to photographic type, he says that:

"For average text types - say 10, 11, or 12 point - 4-to-the-em is desirable. Smal type sizes - such as 6, 7, or 8 point - are often easier to read with the more generous 3-to-the-em word spacing." ( {Craig 1971}, p. 127)

Craig says nothing in this edition about inter-sentence spacing, but an examination of his text reveals that it runs between roughly an En-quad and an Em-quad. (It's a bit hard to tell, because his publisher adopted a column measure significantly narrower than the one he recommends in his text.)

So by the third quarter of the 20th century, spacing is definitely getting tighter. Nearly three decades later, almost at the turn of the 21st, things are both tighter and more ambiguous {Craig 1999}. Inter-word spacing is no longer actually defined as something that the typographer controls directly. It has been outsourced to the computer programmer, and can only be specified in vague terms such as "normal" or "tight." Inter-word spacing, on the other hand, has shrunk dramatically:

"Normally, a single wordspace is required after a period. This is fine for the end of a sentence but too generous when periods follow an abbreviation." (p. 123)

But this reduction of the inter-sentence space to no more than the inter-word space happened earlier. One of the better-known typographical authorities of our time, Robert Bringhurst, was quite emphatic about it as early as 1992. He is worth quoting in full because his approach is illustrative not only of current fashion but of the nature of change in fashion:

"In the nineteenth century, which was a dark and inflationary age in typography and type design, many compositors were encouraged to stuff extra space between sentences. Generations of twentieth-century typists were then taught to do the same, by hitting the spacebar twice after every period. Your typing as well as your typesetting will benefit from unlearning this quaint Victorian habit. As a general rule, no more than a single space is required after a period, a colon or any other mark of punctuation. Larger spaces (e.g., en spaces) replace punctuation. ({Bringhurst 1992}, pp. 28, 30)

So here we have the modern position, and the argument for it. Disregarding his interesting claim that an En-space can function as a punctuation mark, this is a relatively straighforward example of how fashions change from one generation to another: by insulting the previous one.

The human arts are not like the sciences. After Kepler (not Copernicus, but Kepler), geocentrism was not only old-fashioned, it was wrong. It became part of the history of science, not science itself. But however much more we know about printing and type today, our knowledge and our works do not make the Gutenberg Bible "wrong." It remains a part of printing and the graphic arts today, even though few now can actually read it.

In type and printing, fashions change, but they do not invalidate the past. What we have today is new and different from what we had yesterday, but that does not make today right and yesterday wrong (nor does good work from yesterday make the always jarring work of today wrong). Yet it has always been the case that proponents of new fashions have believed, sincerely, that they alone possess access to objective truth. This cannot be argued with any logic (because it isn't true), so it must be argued by insult. Indeed, the insult is an excellent marker to identify a mere change of fashion rather than a reasonable argument.

The insults to 19th century typography are threadbare. They were the products of commercial marketing of type in the 1920s and 1930s. But since the pioneering work of Nicolette Gray in the 1930s, Ruari MacLean in the 1950s, and others, the 19th century in type now stands on its own with a validity equal to that of any other period. Calling it "dark and inflationary" and "quaint" no longer harms it, but these characterizations are little to the credit of those who make them. We are after all talking about the era of the poet-typefounder Thomas MacKellar and of scholar-printers of the caliber of Theodore Low DeVinne.

But in one sense Bringhurst is right. The fashion in type today is for exceedingly tight spacing. To the extent that one might wish to mimic digital lettering with a typewriter (a notion which might indeed be thought of as "quaint"), using a single space after a full stop would now be appropriate.

3. The Tightening of Set

As I hope to have shown above, the fashions of type have changed over the last century or more - changed very strongly in the direction of tight spacing. It is clear to the practicing typefounder of today that this change extends beyond word and sentence spacing to the fit of the types themselves.

In very many cases, we know exactly the intentions of the type maker for the set width of their types; we don't have to guess at it from printed samples. For surviving types, we can simply measure them. For surviving matrices, we can often read their set information as stamped on the matrix. This is particularly the case with the various Monotype matrices which form the core of the present revival of commercial and amateur typecasting.

When we cast today from, for example, a Lanston Monotype display matrix, we know to the eighth of a point what Lanston's intent was for its set. It's coded right there on the matrix. Yet it has been the experience at my place of apprenticeship, Skyline Type Foundry, that in most cases this set is wider that the set expected by today's letterpress printer. If this type is cast today exactly as either of the Monotype firms intended it to be cast, it would look "wrong."

The solution at Skyline is simple: cast "beard-to-beard" to make the narrowest body type upon which the face (with beard) will sit completely.

4. The Future

Both expected set and practices of spacing have been tightening for decades now. What will happen in the future is, obviously, unknown. But by chance I heard an interesting observation just as I was writing this Notebook. My elderly mother, who has vision problems but who remains an avid reader, is on the brink of cancelling here subscription to a popular nationwide printed magazine because the spacing of the letters has changed - it has recently become so tight that she can no longer make out the words. Fashions change, but reading does have physiological aspects. As the eyes of the young graphic artists of today age, the only thing that is certain is that the times will change again.

5. Notes and References

{Craig 1971} Craig, James. Ed. Susan E. Meyer. Designing with Type First edition. NY: Watson-Guptill Publications, 1971.

{Craig 1999} Craig, James and William Bevington. Ed. Susan E. Meyer. Designing with Type Fourth edition. NY: Watson-Guptill Publications, 1999.

{DeVinne 1883/1978} DeVinne, Theodore Low. Manual of Printing Office Practice. (1883)

This was an in-house publication of Theodore L. DeVinne & Co. It was reprinted in 1926 with an introduction by Douglas C. McMurtrie (NY: Press of Ars Typographica, 1926). This, in turn, was reprinted in facsimile in 1978 (Forest Hills, NY: Battery Park Book Company).

{DeVinne 1900} DeVinne, Theodore Low. The Practice of Typography: A Treatise on the Processes of Type-Making, the Point System, the Names, Sizes, Styles and Prices of Plain Printing Types. NY: The Century Co., 1900.

Printed by The DeVinne Press.

{MacKellar 1882} McKellar, Thomas. The American Printer. 13th edition. Philadelphia, PA: MacKellar, Smiths & Jordan, 1882.

{Musick 1903} Musick, William L. Musick's Typewriter Instructor. (St. Louis, MO: W. L. Musick Publishing Co., 1903)

Online via Google Books.

{Polk 1945} Polk, Ralph W. The Practice of Printing Peoria, IL: The Manual Arts Press, 1945.