Tiger Rag: Attitudes toward Hyphenation and Rag Settings

Register today for the free course “5 Skills Every Design Needs to Know.”


Paul Shaw investigates the history of justified, flush left, and rag right settings, and takes a look at contemporary hyphenation practices. Plus, he addresses the modern-day aversion to hyphens in the design world.

This essay began as an inquiry into contemporary hyphenation practices in books and magazines, but quickly spread to include an investigation of the history of justified and flush left, rag right settings. In turn, that investigation threatened to lead to others as it became clear that language, typeface choice, point size, line length and paragraphing styles influenced the quality of these two text formatting approaches. Thus, as with anything involving typography, it turns out that even something as humble as the hyphen is inextricably linked to larger issues. Most of those larger issues will be ignored in order to keep this essay from becoming a novel.

The hyphen is an unprepossessing character. Yet, it elicits loathing among many contemporary graphic designers. They do what they can to avoid it. They turn off hyphenation when setting text flush left, rag right. They try to minimize the number of hyphens in a row at the end of lines in justified text setting. Where did this attitude come from?

The hyphen was originally invented by the Greeks to aid the reader to overcome ambiguity in scriptura continua (writing without word spaces). According to Paul Saenger, author of Space between Words: The Origins of Silent Reading (1997), it was introduced, along with other diacritical signs and marks of punctuation, into Latin writing in the late 2nd century or the 3rd century when the Romans adopted the Greek style of scriptura continua. From the 8th century onward the hyphen was used as a consistent sign to correct an inappropriately placed space in separated writing, including words broken across a line.

In producing his Bible, Johann Gutenberg copied medieval scribal practices as well as the textura of local German manuscripts. This meant that his character set included numerous ligatures, abbreviations, and alternate narrow letters that enabled his columns of text to be justified and yet have close and consistent word spacing. And, like the scribes, Gutenberg employed hyphens to link words across lines. His hyphens are angled double lines and, remarkably, they are hung in the margin. Consequently, the pages of the Gutenberg Bible are beautiful. In the design and use of his hyphens Gutenberg had solved some of the problems that confront modern typographers. [Fig. 1]

Fig.1: Detail of page from the B42 Bible of Johannes Gutenberg (1455). Note how the hyphens hung in the margins as well as the use of ligatures (e.g., colatis in line 1) and abbreviations (e.g., Regnate in line 7) to achieve justification with even word spacing.

While Gutenberg used an enormous character set to achieve a consistent texture, subsequent printers dispensed with most of the ligatures, abbreviations and alternate forms as unnecessary. They realized that, unlike scribes, they could achieve justification by distributing space within a line. Yet, hyphens continued to be essential for line breaks, though hanging hyphens disappeared because they made composing metal type more complicated. Jenson, Ratdolt and Aldus used angled but single hyphens, while the latter also introduced the horizontal hyphen we know today. No one seemed to be overly concerned with the number of hyphenated lines in a row. Four in a row can be found not only in the Gutenberg Bible, but in such other celebrated books as the Nuremberg Chronicle (1492) printed by Anton Koberger, the polyglot Bible (1569–1572) of Christopher Plantin, and the Médailles sur les Principaux Evenements du Règne de Louis le Grand (1702) set in the Romain du Roi. [Fig. 2] In the Hypnerotomachia Poliphili printed by Aldus Manutius there is at least one page with five hyphens in succession: two diagonal, two horizontal, and one diagonal. (It should be noted that diagonal hyphens not only fill space better than horizontal ones, but that they take up less space laterally.)

Fig. 2: p. 88 from Médailles sur les Principaux Evenements du Règne de Louis le Grand (1702) set in the Romain du Roi. Note the three hyphens in a row at top right; and the use of an ampersand in place of “et” in several places.

[Click here to see the Liber Chronicarum (the Nuremberg Chronicle), and note the four hyphens in a row midway down the page (f.1r). Click here to see the Biblia Regia (polyglot Bible), noting that on the right-hand page there are four hyphens in a row in the italic column and five in a row in the Greek column. And click here for the Hypnerotomachia Poliphili printed by Aldus Manutius (1499).]

There is no information about what printers thought about such situations. It is not until the late 18th century that hyphenation—or word division as it was styled—first appears in printers’ manuals. John Smith, author of The Printer’s Grammar (1787), approvingly says, “It shews a good judgment in a Compositor, to prevent Divisions, or any other Point, to fall too repeatedly upon one another, at the end of lines, especially where a syllable may be got in, or drove out, without much difficulty.” But exactly how many broken lines in a row is unacceptable he does not say. Nearly one hundred years after Smith, John Southward, in Practical Printing (1884), was more explicit: “Our first advice, however, to those about to divide would be simply, ‘Don’t.’ If you can prevent it by altering the spacing, do so; but the spacing should not be glaringly different to that of preceding and following lines … Two successive lines ending with a divided word are unsightly, but three should never be permitted, except in very narrow measures.” Unfortunately, Southward’s book is visually unsightly, pockmarked with large word spaces—including double spaces after punctuation—and paragraphs indicated by a combination of indents and line spaces. [Fig. 3]


Fig. 3: p. 99 fromPractical Printing by John Southward (1884). Note the large word spaces following a period and the unnecessary line spaces between paragraphs.

Theodore Low De Vinne, considered to be the best printer in America in the latter half of the 19th century, wrote a quartet of books on printing practice, including one entitled Correct Composition (1901). Unlike Smith or Southward, he refuses to make any rigid pronouncements on the proper number of hyphens in a row, concluding that, “A strict compliance with all these rules is impracticable in the ordinary measure without the cooperation of an author who is willing to shorten or lengthen the words in a line by substituting synonymous words or expletives that will prevent the objectionable division. There are few authors who will take this trouble. Without doubt, words always appear better unbroken, but the breaking of words may not be so unsightly as the breaking up of a general uniformity of the spacing between words.” Only eight times, though, does De Vinne resort to as many as three hyphens in succession in his 489-page book—two of them in explaining word division in German. [Fig. 4]


Fig. 4: p. 83 from The Practice of Typography: Correct Composition by Theodore Low De Vinne (1901). One of the few instances where three hyphens in a row occur. The double word spaces following punctuation was typical of 19th-century typography, even that of De Vinne.

De Vinne’s and Southward’s books are both marked by loose word spacing, a hallmark of 19th-century typography, instances of the feeble printing that spurred William Morris to establish his Kelmscott Press. Although Morris is known for sparking a revival of robust typefaces, his contribution to typography is often overlooked. He copied Nicolas Jenson’s use of angled hyphens. But more importantly, his emulation of the Venetian printer’s dense and even text blocks triggered an enduring preference for close word spacing in text. [Fig. 5] Geoffrey Dowding, author of Finer Points in the Spacing and Arrangement of Type (1954) praised Jenson, noting that “one of the essentials of well set text matter—a striplike quality of line” is only achievable with consistent close spacing. [Fig. 6]

Fig. 5: p. 3 from The Golden Legend of Master William Caxton (Kelmscott Press, 1892) designed by William Morris. Note the angled hyphens and the use of ampersands in place of “and.”

Fig. 6: p. 2 from Eusebius printed by Nicolas Jenson (1470). Note the close word-spacing and the angled hyphens.

The desire for a dense but even color in blocks of text led Dowding to assert that “it is infinitely preferable to have a number of break lines succeeding each other than to have widely word-spaced lines.” In support of his point he approvingly cited an unnamed book by Plantin that, despite having 10 hyphenated lines in a row on one page, had impeccable word spacing. Jan Tschichold was in agreement with Dowding. In his Penguin Composition Rules (1947), he wrote, “Words may be freely broken whenever necessary to avoid wide spacing, as breaking words is less harmful to the appearance of the page than too much space between words.” [Fig. 7]

Fig. 7: p. 15 from Finer Points in the Spacing and Arrangement of Type by Geoffrey Dowding (1954) showing the close word-spacing he advocated. He was indifferent to the presence of widows in bookwork, remarking that they did not matter unless on the last page of a chapter.

Dowding urged close spacing after punctuation (something taken for granted today but lacking in 19th-century typography and much ordinary printing of the first half of the 20th century), proposed hanging hyphens in the margin, and controversially suggested the use of the ampersand in place of “and” to save space (though not for every instance of the word). The latter idea was not as radical as it seemed because early printers frequently used ampersands to adjust the spacing of lines and achieve a justified block of text. It was also an idea propounded (and demonstrated) by Eric Gill in An Essay on Typography (1931).



In fact, Gill was even more willing to challenge convention than Dowding. Not only did he liberally use ampersands for “and” but he also used contractions (e.g., “tho’”), and superscript letters (e.g., “production”) to achieve even spacing. But most importantly, he advocated that text be set flush left, rag right (though he did not use that phrase) as not only more natural than justified setting, but as the best way to guarantee consistent word spacing. He considered the insistence on justified text to be nothing more than a superstition, remarking that “even spacing is more important typographically than equal length.” In his view justified text existed to satisfy man’s desire for neatness. [Fig. 8]

Fig. 8: p. 106 from An Essay on Typography by Eric Gill (1931) showing his use of ampersands and superscript letters.

Gill’s encouragement of such antiquated practices as the use of ampersands, contractions and superscript type not only reflected his background as a letterer and lettercarver, but also suggests that he too yearned for neatness. Otherwise, his desire for natural line endings should have led him to accept loose rags.

Gill is apparently the first person to advocate flush left, rag right setting for text blocks. In the 1920s the proponents of die neue typographie (“the new typography”) in Germany clung to justified text in their works even as they railed against central axis typography. Their focus was on titles and headings, not on continuous text. [Fig. 9] This limited perspective continued with the Swiss modernists in the 1930s and 1940s. The earliest example of flush left, rag right setting after Gill that I have been able to find is Max Bill’s essay “über typographie” in the April 1946 issue of Schweizer Grafische Mitteilungen. Yet, the texts of later publications designed by Bill are set justified. The first Swiss designer to set his texts flush left, rag right on a consistent basis was Karl Gerstner, beginning with a special issue of Werk in 1955.

Fig. 9: p. 11 from Die neue Typographie by Jan Tschichold (1928). The text is justified with four hyphens in a row.

Bill’s essay is notable for its lack of hyphens, while Gerstner’s texts (e.g., kalte kunst? [1957] and die neue graphik [1959]) use them sparingly. Neither man aimed at achieving the near-justified look that Gill sought. But their rags were tight.

In his retrospective Karl Gerstner: Review of 5 x 10 Years of Graphic Design, Etc., Gerstner emphasized his pioneering use of flush left, rag right text in Werk, but in his influential essay “Integral Typography” (1959) he avoided the topic entirely, including his views on hyphenation. Instead he guardedly observed the following: “Today typographers use both sans and Roman type, set books both symmetrically and asymmetrically, use both flush left, ragged right and flush left, flush right. Today everything is stylistically allowable, allowable from the point of view of up-to-dateness.” [Fig. 10]

Fig. 10: “Integral typography” from Designing Programmes by Karl Gerstner (1964). Note the “bad” rags of the final two paragraphs, one with a “weird shape” and the other with lines too even and too short.

Gerstner was not alone among the leading Swiss designers of the 1950s through the 1970s in ignoring the subject of hyphenation. In his summary of “the ‘new’ typography” in 1961, Josef Müller-Brockmann insisted that “spacing between words should be uniform” but made no comment about flush left, rag right setting or hyphenation. Emil Ruder in Typography (1965) said only that flush left, rag setting is marked by “Even spacing, hence uniform grey.” However, he must have approved of hyphens because his text has them. [Fig. 11] Similarly, it can be presumed that Armin Hofmann was in favor of hyphens because Armin Hofmann: His Work, Quest and Philosophy (1989) uses them liberally (with up to four in a row) to achieve a rag as tight as Gill’s. It should be noted that Gerstner’s rags are looser than those of Hofmann or Ruder, suggesting that he tried to avoid hyphenation as much as possible without abandoning it entirely. [Fig. 12]

Fig. 11: “Introduction” to Typography by Emil Ruder (1967).

Fig. 12: p. 23 from Armin Hofmann: His Work, Quest and Philosophy (1989). There are two instances of four hyphens in succession on this page.

The ascendancy of the International Typographic Style since the late 1960s has led to a situation in which rag settings are the norm in design education (at least in the U.S.) even though the mass of printed material (books, magazines, newspapers especially) continues to be set justified. Design educators—even in a postmodernist world—ignore or disparage justified text and, by extension, hyphens. Justified text is not “modern” and hyphens mar the “clean” look of a text block. Flush left, rag right setting is not only a default, but it is preferred.

In the past, printers’ manuals, taking justified text settings as a given, devoted pages to laying out lengthy rules for word division or word breaks (terms used instead of hyphenation). The typographers of the 20th century seem to have ignored the subject not because they thought it unimportant, but because it did not concern them directly. In the pre-digital era when typographers had to prepare specs for compositors and type houses to execute, it was up to those parties to know the proper rules about hyphenation and to apply them correctly. The typographer only had to indicate his preferences in the initial specs and then check over the galleys to see if the setting was acceptable. Making changes was costly.

Despite their preference for rag settings, the Swiss typographers, both modernists and those of a more traditional bent, do not seem to have been greatly concerned about hyphenation. It appears they took it for granted as inescapable, especially for the German language. Perhaps this is why there is no mention of hyphenation in the post-World War II essays on typography and book design by Jan Tschichold (gathered as The Form of the Book: Essays on the Morality of Good Design (1991)) or by Jost Hochuli in Detail in Typography: Letters, Letter-Spacing, Words, Word-Spacing, Lines, Line-Spacing, Columns (1987). [Fig. 13]

Fig. 13: Typographisches Bewußtsein by Wolfgang Weingart (1971). Note the six consecutive hyphens in paragraph 1 (upper left).

The contemporary preference in design education for rag settings is reflected in design and content of the typography manuals of the digital era. Despite its sub-subtitle, Type: Form and Function: A Handbook on the Fundamentals of Typography by Jason Tselentis (2011) ignores hyphenation in its section on text typography. Neither Stop Stealing Sheep & Find Out How Type Works (1993) by Erik Spiekermann and E.M. Ginger nor A Type Primer by John Kane (2003) talks about hyphenation, though both use it in their rag settings. (Stop Stealing Sheep even has one page with seven hyphens on it, though not consecutively.) [Figs. 14 and 15] Type & Typography by Phil Baines and Andrew Haslam (2002) not only uses hyphens in its rag setting, but the authors address the topic directly. “Some people have an irrational fear about designing with hyphenated text whatever the alignment and regardless of what they may be accustomed to reading,” they declare. This is in stark contrast to Thinking with Type: A Critical Guide for Designers, Writers, Editors, & Students (2004) by Ellen Lupton. In discussing good and bad rags she urges her readers to “strive vigilantly to create the illusion of a random, natural edge without yielding to the sin of hyphenation.” She practices what she preaches. Her book is virtually sin-free with only five instances of hyphens that I could find, three of them for compound words. But her rags often leave much to be desired. [Fig. 16]

Fig. 14: p. 135 from Stop Stealing Sheep & Find Out How Typography Works by Erik Spiekermann and E.M. Ginger (1992). Despite the use of hyphenation, several lines seem shorter than they need be (e.g., the first line of the third paragraph).

Fig. 15: p. xi from A Type Primer by John Kane (2003). The rag of the left column is sloppy compared to the rags of the other two columns. The fault is not due to the use of hyphenation but to the choice of which words to break.

Fig. 16: p. 25 from Thinking with Type: A Critical Guide for Designers, Writers, Editors, & Students by Ellen Lupton (2004) showing how “ugly”—in my view—a wild rag created without hyphenation can look. Note especially lines 1 and 14, and the first six lines of the final paragraph.There is hyphenation elsewhere in the book, which makes a page like this inexplicable.

Two books tackle the issue of hyphenation in detail: The Complete Manual of Typography: A Guide to Setting Perfect Type by James Felici (2003)—which has an entire chapter of more than 30 pages devoted to “Controlling Hyphenation and Justification”—and Type Rules: The Designer’s Guide to Professional Typography (4th edition, 2013) by Ilene Strizver. Both describe hyphenation as a “necessary evil” and try to minimize it. Strizver, like Baines and Haslam, insists on no more than two consecutive hyphens—a position that is draconian in comparison with historical practice. (By the way, it should be pointed out that the obsession that typographic writers have with the number of hyphens in a row is because of a dislike of “ladders” in justified text. Ladders do not occur in good rag settings, no matter how many consecutive hyphens there are.)

As is obvious by now, the issue of hyphenation is inextricably linked to that of alignment, whether justified or ragged. Its presence in a rag setting is dependent on the nature of the rag. Writers on typography differ wildly on what constitutes a “good” rag, though they tend to agree more on the aspects of a “bad” rag. The latter has “weird shapes” in Lupton’s parlance, in which lines of increasing/decreasing length are to be avoided. In the view of some writers, lines that are nearly even in length are to be avoided as well.

Gill aimed for a rag that was as even as possible, while Felici and Strizver advocate a tight rag—like that of Bill and Gerstner—in which the margin is irregular but not too irregular. Strizver describes the ideal rag as “a gentle wave that makes slight in-and-out adjustments as the eye travels down the text.” A tight rag of this nature requires hyphenation. However, Robert Bringhurst, the author of Elements of Typographic Style (4th edition, 2012), ridicules such a rag as looking like a “pie-crust”. He prefers a hard rag without hyphens. What he terms an “honest” rag, others call a “wild” rag because the lack of hyphenation results in greatly fluctuating line lengths.

This essay was prompted by my annoyance over several instances of wildly fluctuating line lengths in various design magazines and books I read over the past two years. I found myself often stopping and rereading sentences because short lines in a wild rag made me think that a word or phrase had been accidentally dropped. But these were not “honest” rags. When I attempted to recreate several of them, I could not do so simply by turning hyphenation off. Instead, these overly short lines were clearly deliberate design choices. That discovery forced me to abandon my original plan to write a rant against designers who turn off hyphenation and to instead investigate how printers and designers have viewed hyphenation over time. [Figs. 17–19]

Fig. 17: p. 21 from Design Vignelli: Graphics, Packaging, Architecture, Interiors, Furniture, Products (1982). Note the sloppy rag (and widows). There is an extremely short line in the second paragraph of the right-hand column.

Fig. 18: p. 44 from Eye magazine 93 (Winter 2017). There are three lines in the left-hand column that, in my view, are too short.

Fig. 19: p. 18 (detail) from Ruder typography / Ruder philosophy by Helmut Schmid (2017). Even though hyphens are employed, the rag in the first paragraph has several overly short lines that could be improved by further (or different) word breaks. It is not as neat as the other paragraphs.

One thing that I realized in looking at unusual or bad rags in a wide array of design publications—including books on typography!—is that while a lack of hyphenation is often the culprit, there are other factors at play. I may have wanted to avoid talking about typeface choice, point size and line length (measure) at the beginning of this essay, but in the end I find that I cannot escape them. Balancing these factors is as important as deciding whether to hyphenate or not. Hyphens are not a necessary evil. If used with care and understanding, they are an aid, rather than a hindrance, to good typography. [Figs. 20 and 21]

Fig. 20: p. 43 from Modern Typography: An Essay in Critical History by Robin Kinross (1992). An example of an excellent rag, achieved with judicious use of hyphenation.

Fig. 21: p. 26 from herbert bayer: painter, designer, architect (1967). The setting of text in this book is very peculiar. The alignment is both ragged and justified at the same time. Because Bayer chose to set the text entirely in lowercase, he needed a way to make it clear where a new sentence began. He thus added space, pushing the line to the full measure. But elsewhere he allowed the lines to end naturally. The effect looks as bad or worse than the 19th-century typography that is often excoriated.

Thanks to Grendl Löfkvist for providing feedback as this article was in progress and to K.C. Witherell for suggesting the title.


5 thoughts on “Tiger Rag: Attitudes toward Hyphenation and Rag Settings

  1. johndberry

    This is a good article, with a wealth of examples. I particularly like: “Hyphens are not a necessary evil. If used with care and understanding, they are an aid, rather than a hindrance, to good typography.” And Paul’s pointing out that a “hyphen stack” only matters in justified setting; in ragged-right setting, it’s nearly impossible for hyphens on succeeding lines to line up into a visual stack.

    I wrote a couple of articles about just this subject almost 20 years ago. The software may have changed, but the principles haven’t.

    https://creativepro.com/dot-font-the-justification-for-hyphenation/
    https://creativepro.com/dot-font-the-hyphenation-of-justification/

    These days, with flexible publishing for the screen, we need to incorporate these principles into adaptive, responsive typography. Not a small task, but one that needs to be done.

    John D. Berry

COMMENT