A Refutation of

The Elements of Typographic Style

& also something of a backdoor defense of creative freedom

It’s the typography reference book you’ve heard of. The one everyone recommends to everyone else. If you’re a student it’s probably at the top of your list of resources; if you’re a teacher, you probably put it there. If you’re not a designer but have ever asked the question “What font should I use?” the answer you’ll most likely get is “Look into Bringhurst.”

Published in 1992 and assiduously updated since, The Elements of Typographic Style by Robert Bringhurst enjoys a high place in the designer’s library. It is an earnest and serious discussion of typography, and garners praise as being a virtuous model of the very art it extols. Four editions in twenty years speaks to a robust acceptance by the design profession.

It is also, as best I can estimate, one of those books everyone owns but few have read and fewer still follow closely in practice. Which is an odd status for a book, to be very highly praised while also in some way neglected, and which prompts the question, is it a good book? Is it useful and accurate and rich and wise?

As they say here on the internet, the answer just might surprise you. In fact, the deeper you read, the more you have to wonder just what the book is really about, and what is so wrong about having too many quotation marks? I hope to persuade you here that The Elements of Typographic Style is a flawed and misleading book — flawed in its advice on design practice and its coverage of the subject, and misleading on just what the real purpose of the book might be. At the very least, I hope to convince designers and anyone turning to the book for advice to take a much more skeptical look at its philosophy of design. (And for future reference: use Caslon. It’ll be fine.)

First let’s look at the heart of the matter: working with type. More precisely, that fundamental question of selecting a typeface for a project. The central principle in ETS [1] is given as:

Choose a typeface or a group of typefaces that will honor and elucidate the character of the text.
This is the beginning, middle and end of the practice of typography: choose and use the type with sensitivity and intelligence. §1.2.4 [2]

Here and there we get nibbles at the subject, but not until Chapter 6 do we come to the book’s advice on type as a creative practice. Broadly, this advice is the standard and reliable “Consider the subject and pick a type accordingly” school, as described in §6.2.1’s bicycle racing book example:

Choose faces that suit the task as well as the subject.
Briefly this means pick an “inherently good type” that’s a “good type for comfortable long-distance reading” [i.e., for the task at hand] and that’s “sympathetic to the theme.”

In this section’s very simplified example, a sympathetic type to the subject of bicycle racing means one “lean, strong and swift; perhaps it will also be Italian. But it is unlikely to be carrying excess ornament or freight, and unlikely to be indulging in a masquerade.”

The fundamental approach to working with type is the old reliable Type 101 approach of considering the text and selecting a face that suits it. Fine. It’s generally a feasible concept in the hands of someone with more, rather than less, type sensitivity. The appeal of this method (which is certainly not confined to ETS) is that it’s easy to commodify in simplistic examples and to pass on as teaching. It’s a place to start, is usually how it’s put.

The problem with this approach is that it’s always been an invitation to reductive readings in which something complex becomes something unnuanced, which can be matched to letterforms. There’s an assumed one-to-one relationship between the text’s subject and type criteria. Cycling = Italian, with no consideration of the intended audience, the genre, the writing style, and so on — all factors that designers in reality are taught to take into consideration. If you look for one quality in a text on which to base your typography, how can you a) avoid clichés and obvious, banal type choices; b) evaluate which quality, to the exclusion of others, to draw on for design guidance; and c) represent real complexity within a text or project? What value is guidance if it’s simplified to such an extent that the guidance no longer helpfully depicts the way designers actually design?

Looking at §6.4.1, we can see this thinking applied to typefaces themselves:

Letterforms have character, spirit and personality. Typographers learn to discern these features through years of working first-hand with the forms, and through studying and comparing the work of other designers, past and present.

The key word here is “discern,” which conveys an assumption that a typeface’s qualities are preexisting, inherent, always there to be discovered by the attentive designer. Whatever qualities a typeface has are already embodied in the forms of the face and in the type’s accumulation of history. Thus Bembo will always display and stand for whatever aesthetic or cultural values you want to ascribe to the Italian Renaissance. Likewise Futura and the “purity” of geometrical modernism. Understanding typefaces in this pre-ordained and ultimately limited way allows one to accept a dogmatic statement such as “The typographer seeks to shed light . . . by setting every text in a face and form in which it actually belongs” (§6.3.1, emphasis mine).

In this light, we might ask what typeface and typographic vocabulary should a designer use for, say, Lydia Davis’s 2010 translation of Flaubert’s Madame Bovary, written in 1856 and perhaps the ultimate example of the realist novel? What’s a French realist typeface? Is a Garamond face any more or less realist than Fournier’s or Didot’s, despite the fact that Garamond preceded the rationalist model of Romain du Roi by more than 100 years? Or or J.-F. Porchez’s Le Monde, designed in 1997? How much should we take into account that Lydia Davis is neither primarily a novelist nor, in her own excellent fiction, a realist? Plus she’s American, and a woman, and alive right now, whereas Flaubert is none of these things.

The historical-cum-nationalist approach is even more unworkable in non-book contexts — what typeface “belongs” in the corporate identity of a multinational food and beverage company founded in 1923 in America but with brands today in 155 countries, which is headquartered in Switzerland and Illinois, is publicly traded on the NYSE, and is best known for its macaroni and cheese products? I think you get the point, though these are not absurd examples. Practicing designers don’t rely on what amounts to a form of typographic identity politics because as a design method it’s simply too literal and too creatively limited.

Experience teaches us that neither texts nor typefaces can be reduced to a single meaning because the context in which a typeface is used is always intruding on the effects that a typeface will produce. Designers do, of course, strive to find the typeface that feels right for a given text or project. But the forces pressing on that sense of feel flow only partly from the personality of the typeface. In addition to historical echoes, typefaces have quite immediate contemporary associations; projects exist in specific contexts (from the locality of a Pilates studio to the globality of the internet); non-designers such as authors, editors, and clients exert powerful influence — this is the reality of designing with type. The decision of what font to use is far more complex a process than ETS ever portrays or investigates, despite asserting that this very practice is the “beginning and end” of typography.

No typeface is an ever-fixed mark. This turns out to be incredibly freeing, as it opens up greater creative space for designers. In fact, working with type is a process, perhaps more than anything, of unlearning the historical bondage of a given typeface, of chipping off the ossified shell of past uses to find ways to make a type’s qualities resonant with and relevant to the current age. Otherwise typography is just a matter of historical plug-and-play. Where Bringhurst uses “discern,” the better word would actually be “control” to describe the designer’s art. It’s more common to speak of a designer’s control over type and the page than his or her obedience to historically determined features, mannerisms, or aesthetics. Designers are free to work with or against a type’s qualities at will, thus unfixing those very features that ETS would have us believe are historically set.

Perhaps you’re unconvinced. Perhaps the spirit of ETS’s approach to typography — the honoring and the tradition and all — still resonates with you. Fair enough. Let’s turn instead to a more overt flaw in the book: its editorial structure, and how the very structure of the book itself fails to uphold basic principles of logic and hierarchy that are so central to the practice of typography. This flaw is plain in the sequence of chapters as well as the internal development of certain chapters.

The editorial logic of ETS displays an inconsistency to such a degree that I wonder if there’s any logic at all. Consider Chapter 3, Harmony and Counterpoint, which is divided into the following sections:

3.1 Size
3.2 Numerals, Capitals & Small Caps
3.3 Ligatures
3.4 Tribal Alliances & Families
3.5 Contrast

Size and contrast are primary formal considerations that interrelate with measure, margins, and textual elements, as well as type choice. Collectively these are the first matters a book designer addresses. Yet these considerations are given equal weight, according to the chapter’s hierarchy, with the secondary and tertiary technical factors of numerals, small caps, and ligatures. It’s even less clear how the rambling §3.4 fits in, with its “constitutional monarchy of the alphabet” and extended defense of never experimenting with anything.

The lack of editorial logic has two consequences, aside from diminishing the book’s general coherence and thus authority. First, it runs counter to one of the basic principles of book typography: establish a clear and rational hierarchy of text elements. In publishing houses, the editor and copy editor establish a manuscript’s logic, working in service to and conference with the author. Traditionally there’s great discipline and thought put into this process so that a book’s logic is clear. The designer’s task is to cast this editorial structure into a corresponding visual system (this is the very point that §1.2.2 makes). The point being, editorial logic precedes typography. It has to. And yet in ETS, a book that deals with typography, we have precious little editorial logic in the book itself. A book with an erratic editorial structure is a poor example indeed of the values of proper typographical structure.

The second consequence is that the lack of logic disorients the reader, which mirrors how inconsistent typography can mislead the reader. Many times I had to stop and try to figure out why I was jumping from the openings of a book (§4.1) to endings (§4.3) and back to frontmatter (§4.5). How does it makes sense that we read about text figures (§3.2) and analphabetic symbols (Chapter 5) before addressing typefaces wholistically in Chapters 6, 7, and 11? Why are the type-focused chapters (6 and 11) interrupted by page geometry (Chapter 8) and minute font software mechanics (Chapter 10)? None of this would have been difficult to set in an order that made sense to the reader.

Book design is perhaps the most logical, top-down, and systematic of all the genres of design this side of designing in code, and yet ETS can’t sort itself out rationally. A book must make sense structurally as well as linguistically and visually. Books, especially ones on relatively technical topics, which read haphazardly and don’t reflect true practice, are not usually recommended on all those syllabi.

Perhaps we can look to the goals of the book to suss out why it’s organized this way. Just what kind of book is ETS? The foreword is somewhat indeterminate about the matter. Is it a rule book? A proper style guide somewhere on the Elements of Style – Chicago Manual spectrum? In the foreword, Bringhurst describes the book variously as a simple list of working principles, a short manual for typographic etiquette, a rulebook, and “neither a manual of editorial style nor a textbook on design.” It’s a book that claims to be many things while not ever overtly admitting its true purpose (a polemic, as we’ll see).

There are in fact many sections that emulate a manual of editorial style. Where the book’s principles of typographic etiquette intersect with editorial style is where ETS as a useful manual falls apart because ETS promotes nonstandard, antiquated, and otherwise ill-advised techniques that are out of place in text design today, or more specifically out of place in mainstream American book and magazine publishing. Not to mention the wild stylistically lawless frontier of the very medium you’re reading right now.

Chapter 5, Analphabetic Symbols, is the book’s primary discussion of editorial style as it relates to typography. Even within the somewhat specialized confines of book design, many of ETS’s style points lead the designer into an aesthetic that modern readers will find intrusive if not off-puttingly pretentious, assuming editors would allow some of the practices ETS suggests, which is doubtful. Following this chapter’s advice can result in a variety of odd and mannered effects. See for example:

§5.2.1 Use spaced en dashes – rather than em dashes or hyphens – to set off phrases.
§5.2.3 Use the em dash to introduce speakers in a narrative dialogue.
§5.4.1 Minimize the use of quotation marks, especially with Renaissance faces.
In trade publishing these are all matters decided by authors, editors, and copy editors, not designers.
§5.1.3 Use the best available ampersand.
§5.1.4 Consider the lowly hyphen.
[All this mixing fonts for specific glyphs gets awfully tedious and error-prone, even with Find & Replace.]
§5.4.4 Eliminate other unnecessary punctuation.
[!!! Do so only if you don’t mind putting it all back in when the proofreader corrects the proofs.]

Browse through almost any mainstream book or magazine published in your lifetime and you won’t see these practices applied. Furthermore, all of the examples above are points that are determined by authors and editors and house style; they are simply not decisions designers make alone. ETS frequently asserts that the designers must “honor” the text and then goes on to advise practices that would often override the author’s intentions and substantially intrude on the text as well as the reading experience. The designer’s task, after all, is to enable the reader’s free and smooth experience of words, a task that’s intrinsically dependent on shared conventions of punctuation, notation, hierarchy, and legibility. When and where these conventions are broken starts with the author, not the designer.

Authors are almost totally absent figures in ETS, by the way. Editors, copy editors, and proofreaders are all almost totally absent in ETS’s gloss on the bookmaking process, as Bringhurst repeatedly suggests that designers make critical editorial decisions. For example §4.2.2, “Use as many levels of headings as you need: no more and no fewer” — when in reality a designer should, must, and will use exactly as many headings as the author creates and the copy editor specs. For example §4.3.1, “Footnotes that extend to a second page (as some long footnotes are bound to do) are an abject failure of design” — which is an awfully grandiose stance to take on the lowly footnote and anyway irrelevant if the author has written long footnotes because who is going to let the designer cut footnotes? Which designer among us would have (could have) had that argument with David Foster Wallace?

As a discussion about how books are made, ETS is gapingly incomplete. The lack of a true picture of how a manuscript takes form as a work of typography is one of the most obvious and serious flaws of ETS as a whole, even if you granted that its conception of typography is essentially limited to book design. The many crucial omissions in the editorial-to-design process are misleading, especially to generalist designers who work with text but don’t have the benefit of the disciplined manuscript pipeline of book publishing.

Many designers cite the quality of the writing as one of the more enjoyable elements of the book. Written in a grandly metaphorical style, ETS is forested with chestnuts such as:

[A page full of letters] can lapse into a typographical slum, or grow into a model of architectural grace, skilled engineering and simple economy. Broad suburban lawns and wide typographical front yards can also be uninspiringly empty or welcoming and graceful. They can display real treasure, including the treasure of empty space, or they can be filled with souvenirs of wishful thinking. Neoclassical birdbaths, effigies of liveried slaves or stable boys, and faded pink flamingoes all have counterparts in the typographical world. (§4.1.3)
In a badly designed book, the letters mill and stand like starving horses in their field. In a book designed by rote, they sit like stale bread and mutton on the page. (§1.1.2)

Presumably even starving horses would pass on stale bread and mutton? (I have cherry-picked these bits, but I think it’s fair to say they’re representative of the writing style.) An embarrassing shade of purple, the writing is the literary equivalent of setting prose in Spencerian script. To write about typography, shouldn’t the language be direct? In the same way that typographic logic should reflect editorial logic, so too the prose style of a discussion of typography ought to emulate those better angels of typography: clarity and precision. These are after all the values that Strunk and White uphold in The Elements of Style. And what’s more, typography doesn’t need this kind of dressing up. Typography is already an inherently noble pursuit. Typography, by any measure, is housed within the heart of human endeavor in practically all cultures, and has been for millennia.

Various online comments about ETS have left me with the question of whether it’s the prose itself that designers take as the token of the book’s validity. While there’s often disagreement with minor points in the book, a general acceptance of its overall stance seems to prevail, which would indicate an acceptance of the authority of the book’s voice. However, there’s not much basis for according authority to a book if you disagree with its actual assertions.

Even still, perhaps you will object that the real value of ETS is that it teaches us how to think about typography. The book is a model, it’s often said, of high standards in typography.

Read closely and you’ll see that the opposite is true. ETS is not a teaching document. If you happen to share its taste and values, it’s a validating document to be sure, but it’s a book of assumptions, not arguments. It certainly doesn’t present a variety of design approaches in order to explore their merits. How can we learn to evaluate typography if ETS gives no model for this kind of comparative critique? Critical elaboration would certainly help to make ETS a teaching document (as the Chicago Manual is, by the way, with its copious explanations of its choices) rather than the disguised polemic that it more fundamentally is.

In the final analysis, this is the most problematic point. The high-flown language of ETS both underpins and disguises the argument that really feels like the book’s central purpose. The roving subjectivity of its aesthetic guidance actually makes a lot more sense if we take them rather as a set markers in an overall design philosophy. What ETS delivers is a defense of historically conservative design practices asserted along not only aesthetic but moral lines.

Notions of honor, integrity (§6.5.2), and dignity assert a moral cast to typography. The presumption is that one can design dishonestly or dishonorably by making choices that are right or wrong in terms beyond the simply aesthetic. It’s tempting to say Right and Wrong to make the point (it takes typography to make the distinction);

[Logotypes] tend to push [typography] toward the realm of candy and drugs, which tend to provoke dependent responses, and away from the realm of food, which tends to promote autonomous being. Good typography is like bread: ready to be admired, appraised and dissected before it is consumed. (§3.2.3)

Certainly a design can be deceitful by misrepresenting its subject in some graphic way (i.e., the beautifully alluring cigarette pack) — that’s not the kind of dishonesty we’re talking about here. In ETS there’s a truth quality to typographic choices themselves:

The underlying truth of the blank page must be infringed, but it must never altogether disappear – and whatever displaces it may well aim to be as lively and peaceful as that original blank page. (§4.1.1)

We end up absorbing this bias without realizing it, perhaps, lulled by neoclassical metaphors.

It doesn’t take too close a reading to see which typographic styles are connected to the good and the right in ETS. From the typefaces that are mentioned repeatedly — cast an eye over the Index’s mentions of various typefaces — to stylistic suggestions to page makeup, typography of the Renaissance bears the standard over all other periods.

Like Renaissance painting and music, Renaissance letterforms are full of sensuous and unhurried light and space. They have served as typographic benchmarks for five hundred years. (§7.2.1)

ETS’s position on typography after all isn’t so different from saying the best movies were made in the 40s in Hollywood and so we, today, should be making black and white movies to uphold the tradition. Imagine a filmmaking manual that argued for this. Should we emulate Renaissance typography any more than we emulate its forms of dress or cuisine, its painting or music or speech?

I certainly recognize that Robert Bringhurst may assert whatever taste he prefers. There’s a time and a place for it in certain projects, but as broad guiding principles, ETS’s techniques are dubious, which calls into question the book’s function and value as a style guide. For a manual of style that is poorly structured around its topic, inaccurate in its portrayal of design practice, and gives thin, simplified advice on a complex subject, this historicist bias is all the more problematic and amounts to a conservative polemic. Nothing wrong with polemics, but reader beware. And skeptical.

The position I’m framing is not old versus new — please don’t mistake my argument. It’s not traditional versus modern, Bembo versus Futura. Rather, at issue here is restriction versus potential: protect a specific set of choices versus open the field to the exploration of everything. My point is that it’s counterproductive, countercreative, to morally charge the art of working with type. We wouldn’t moralize cadmium red or C sharp major when teaching brush technique or the scales. To do so in typography stunts the full breadth of expression that designers need to draw upon. There is no method, T. S. Eliot said (of becoming a critic), but to be very intelligent. Take it from a Modernist. Take it from a poet. Use all the fonts. Use all the tools. Denying them is counterproductive to the exploration of new idioms and new aesthetics in the service of new eras and new histories.


1. While writing this piece, I got in the habit of using “ETS” as shorthand for the full The Elements of Typographic Style. Please forgive the abbreviation here — I’ve grown accustomed to it and it’s far more economical. No disrespect is intended.

Also, the terms “typographer” and “designer” are more or less interchangeable in ETS, but I use only “designer” to refer to anyone who works with typefaces in creative practice, regardless of media.

2. All citations refer to the Fourth edition. Throughout, these numerals refer to sections and subsections of the book.