On Classifying Type

By Jonathan Hoefler

This text was first published in 1997 in Emigre No. 42.

If there is a Holy Grail of typography it is surely the Omniscient Typeface Classification System, which will organize and index the complete typographical output of mankind. Countless individuals have set to the task of developing such a thing: printer-scholar Theodore Low deVinne, historian A. F. Johnson, critic Beatrice Warde, educator Alexander Lawson, writer Robert Bringhurst, and scores of professional organizations. I suspect that every typophile who has ever spent more than ten minutes trying to locate an obscure typeface in a poorly organized specimen book secretly dreams of devising the ultimate system by which typefaces are classified.

I must confess that I don’t share my colleagues’ fascination with this challenge. Whether something is a humanist sans serif or a neo-grotesque seems largely irrelevant in light of current typography, where the motivation to create new typefaces has been transcended by the drive to invent new kinds of typefaces. Whether this is an artistic endeavor or an imperative of the marketplace is debatable, but it has remained a consistent theme in typography for nearly two centuries. Editors of type specimen books who don’t know where to file Remedy or Hard Times should take heart: a freakish design of 1815 was so vexing to compositors that they ultimately resorted to giving it its own designation. What subsequent generations referred to alternately as the Egyptian, Doric, Ionic, or Antique, history now remembers as the first Sans Serif.

This is not to suggest that the development of a classification system is futile but rather, that it is an infinitely complex task that requires the sensitivity of numerous disciplines – typography being only one among them. Attempts at typographic scholarship have long been impeded by the difficulties of describing typefaces, let alone organizing them. Where biologists have the benefit of the Linnaean system for comparing Homo Sapiens to Homo Erectus, and librarians can be certain that even as-yet unimagined volumes on architectural technique will be found under Dewey 692 (i.e., building construction practices), typographers have only the vaguest standardized terms for describing how Caslon differs from Janson. A taxonomy for type, if it were comprehensive, adaptable, correctable, expandable, generally accessible yet infinitely refined, would be of immense use to anyone connected with letters. If it chronicled the cultural, aesthetic, technological, and literary factors that have influenced type design – instead of postulating a neat progression of styles, implying an uncomplicated evolution – it might approach a more faithful record of the rich and complex history of typography. But the timeworn attempt to find a single best way to organize typefaces remains a hopeless pursuit: like asking “What’s the best way to drive across the United States?” It anticipates a simple response to a complex issue.

The Central Lettering Record

A research team at the Central Saint Martins College of Art & Design in London has recently taken up the challenge to build a better system, as part of an ambitious project to extend the resources of the Central Lettering Record. The Record, as its name suggests, is an archive of lettering of all varieties: as distinguished from typography, which refers exclusively to the study of printing types, lettering subsumes a host of related disciplines, each of which brings its own visual vocabulary to bear upon the shape of letters. The Record includes studies of epigraphy (stonecutting), paleography and calligraphy (handwriting, in varying degrees of formality), architectural lettering, and modern commercial signage in a number of materials. There are letters in neon, plywood, vinyl, and vacuform plastic – media which have yet to engender their own fields of formal study.

The interest in lettering inspired by Edward Johnston early this century began to wane by the 1950s, when British design educators started to take an increasing interest in modernism. To counter the declining interest in non-typographic lettering and to provide an archive for their students at the Central School of Arts and Crafts (later partnered with Saint Martins to become the Central Saint Martins College), Nicolete Gray and Nicholas Biddulph began to amass a collection of alphabetic artifacts and photographs which now forms the core of the Central Lettering Record. After a period of some disuse, a renewed interest in the Record has prompted both its reorganization and its expansion to include typographic developments of the past decade. In connection with this ambitious project, a research group within the college is working to develop a cd-rom on the history of typography, which will present both new and old typefaces in the context of the Record’s well-developed collection of typographic artifacts.

At last year’s meeting of the Association Typographique Internationale in The Hague, Ed McDonald, Eric Kindel, and Catherine Dixon of Central Saint Martins presented a preliminary version of the cd-rom. While the project has undergone considerable changes in the past six months – including a seismic overhaul of the classification system that had been introduced in the winter 1995 issue of Eye – both the architecture of the system and its user interface represent a unique and refreshing approach to the documentation of type history. Perhaps more impressive than the work as it stands today, is its potential for future scholarship, which suggests that when complete, the cd will serve as a comprehensive review of more than five centuries of typography. While the cd itself remains a work in progress, the process by which this vast collection has been edited and organized already serves as an interesting study of the challenges facing the typographic archivist.

Where History Begins

The history of typography as conventionally told begins with Johannes Gutenberg and his celebrated 42-line Bible, printed circa 1455. However, current scholarship reminds us that Gutenberg is preceded by at least eight centuries of printing in the Far East demanding that this discussion be framed as the history of typography in the West, or more precisely, the history of European typefaces manufactured in the manner of Gutenberg. Ignoring for a moment the blackletter script after which Gutenberg patterned his types (because our discussion is about type, not calligraphy), we might proceed safely to the topic of major innovations in type, as they appear in the timeline, for almost ten years. The traditional next stop after Gutenberg is the work of the Italian humanists in the late 1460s, which culminated in Nicolas Jenson’s Roman of 1470 (which we know through countless revivals, including Centaur and Adobe Jenson). The types of this period show the sudden influence of the humanist miniscule, a style of handwriting which took hold as part of the broader influence of secular humanism in the late 15th century. This form of letter might be thought of as the last evolutionary stage of the Carolingian miniscule, a style advanced by Charlemagne in the 8th century, representing an attempt to consolidate the many kinds of regional handwriting used throughout the Holy Roman Empire – resisting this excursion into a discussion of politics and script, let’s continue with Jenson’s type and its many merits.

Aside from being attractive, Jenson’s type is unique in its approach to lettering, which involved reconciling the two alphabets of the ancient world, the majuscule and the miniscule. Majuscules are the letters of antiquity, familiar to us through architectural inscriptions and the typefaces based on them, such as Adobe’s Trajan. Miniscules are a sort of shorthand letter, handwritten forms indirectly descended from the majuscules and akin to our modern lowercase. Jenson successfully reconciled these two different alphabets by rendering the shapes of the miniscule in the style of the majuscule, taking written forms and remaking them in an eloquent vocabulary borrowed from inscriptions featuring stems, hairlines, and serifs. In so doing, Jenson and his Venetian cohorts cemented a typographic relationship which persists as a familiar if invisible dichotomy, that of the uppercase and lowercase. A broader discussion of Jenson’s innovations would include a survey of how inscriptional lettering changed under the Roman Republic and how changes in media forced the evolution of these letters into the quadrata, rustica, and uncial scripts which ultimately informed the Carolingian miniscule – but this is really the province of epigraphers, paleographers, and semioticians.

And so it goes, proceeding in many directions at once. Even in its infancy, type history is the confluence of many histories, and cross-pollination between these different areas of development only multiply as history advances. By the 19th century, when the dominant letter styles appear in many media, the lines of influence become impossible to chart. In Nineteenth Century Ornamented Typefaces (1976), Nicolete Gray discusses lettering of a style she terms “English Vernacular,” describing not only printing types but painted letters on shop fasciae, incised letters on monuments, and cast-iron letters on gravemarkers–never concluding that any of these examples is singularly supreme. By the 20th century, when the design of typefaces is commonly informed by literary and artistic influences which are decidedly outside the realm of typography, it becomes impossible to classify typefaces by their historical position alone. Type history invariably demonstrates that its exemplars, if not all of its participants, occupy the crossroads of many historical themes, none of which are fully within the scope of the study of printing types. The problem of when history begins is thus complicated by the larger question of what history includes.

One Man’s Language

With the understanding that typographic history is more than the history of type, we may better appreciate the approach taken by the Central Lettering Record, which presents a holistic view of lettering by making frequent excursions into related fields. Although this approach presents a more integrated view of type history, it is not without its own difficulties.

The technical terms specific to any area of study are likely to evidence their own biases, and the collision of many disparate vocabularies inevitably introduces ambiguities. A term as seemingly neutral as “Roman,” for example, even divorced from its geographical, national, historical, and political meanings, is fraught with ambiguity: to a typographer it means “upright types,” but in orthography it denotes the Latin alphabet, and to calligraphers it describes a style of lettering peculiar to the 15th century. Sometimes terms specific to typography have multiple or even contrary meanings, leading to such delightful oxymorons as Serif Gothic and Times Roman Italic. The terms “Antique,” “Gothic,” and “Old English” are among typography’s most widely used and, entertainingly, also the most convoluted: “Antique” can mean a slab serif (Antique No. 3), a sans serif (Antique Olive), a humanist book face (Zapf Antiqua), or simply anything that looks old (Caslon Antique). Similarly, “Gothic” can refer to a sans serif (Franklin Gothic), or a blackletter (Totally Gothic), or occasionally both (Gothic Gothic). “Old English” is a hornet’s nest, best avoided altogether.

Terminology also shows its roots. The very words used to describe the inscription on the Trajan Column might betray the concerns of a stonemason, an architect, a geologist, a lapidary, or a type designer, since the specialized vocabulary used by any profession is inevitable colored by particular areas of interest. The language typographers use to discuss non-typographic artifacts is no exception, as seen in the terms adopted by the Central Lettering Record: for example, “Augustan,” “Rustic,” “Tuscan,” and “Wedge” are among the stylistic groupings used to organize its collection of inscriptional photographs. Although these terms were developed to highlight a specific area of interest, even they are prone to confusing many lines of inquiry: “Augustan” is a historical marker, indicating those letters made in the style popularized under Caesar Augustus (27 b.c.–a.d. 14); “Rustic,” divorced from its adjectival meaning, is a paleographic term which describes a particular style of handwriting; “Tuscan” is a term coined by 19th century typographers to describe several genera of ornamented printing types; “Wedge,” though the least colorful of the lot, is the only term to describe letters on a purely visual level.

Apples, Oranges

The conflation of historical, technical, and formal terms is a common problem with typeface classification systems, seen especially in the scheme adopted by the British Standards Institute in 1958, and revised in 1967. For instance, while its “Slab Serif” and “Lineale” categories are purely visual, “Garalde” is a historical denomination (a synthesis of Garamond and Aldus), and its “Script” category takes into account the intention of the design, for it includes only “typefaces which imitate writing.” Other systems muddle things further by expanding their criteria to include usage, such as the DIN (Deutsche Industrie Norman) classification which subdivides “Classical Styles” into four sub-sections, three of them chronological (early, late and modern) and the fourth reserved for typefaces designed for newspapers. The situation gets worse with systems that include geographical designations, like those which maintain a distinction between “Dutch Old Styles” and “English Old Styles,” without ever elaborating on these groups’ taxonomic differences. Much of this is sheer jingoism and can be largely attributed to Daniel Berkeley Updike and his seminal Printing Types: Their History, Forms, and Use, a standard desk reference on type history since its publication in 1922. Despite its usefulness, Updike’s scholarship is marred by extraordinary prejudices at every turn. Without exception, every achievement in three centuries of Dutch typefounding is somehow ascribed to the influence of the French (with whom Updike is inexplicably enamored), while the Dutch are portrayed in shockingly racist stereotypes: Plantin, Van Dijck, and the Elzevirs are commended for their shrewdness, their diligence, their thrift. Updike spits about the mediocrity of “a rugged little Dutch type,” a typeface which is praised earlier in the edition where it is properly identified as the work of Frenchman Robert Granjon. This is par for the course in Updike.

Most typeface classification systems attempt a purely historical approach, categorizing typefaces according to the time of their genesis. The 18th century types of Van Dijck and Caslon share many characteristics and are commonly collected as “Old Styles.” Bodoni and Didot’s types of a century later are distinctly different and are designated “Modern.” Between these extremes are the types of John Baskerville and others, conveniently labeled “Transitional.” Aside from the obvious problem that compartmentalized systems demand rigidity, such classifications distort the historical record by equating chronology with typographic style. This becomes especially awkward in a discussion of 19th century types, when entire categories of typeface develop coevally. Immediately following the period during which the Modern style flourished, a number of styles developed simultaneously, including sans serifs and slab serifs. The category that is listed directly after “Modern” in a classification system might understandably be mistaken as the next evolutionary step, even if this is not really the case.

Assigning types to discrete time periods also intimates that historical styles are visual explorations which have long since been completed. As a result, such systems are utterly unable to cope with typefaces of the late 20th century which are only incidentally related to historical styles. This invariably leads to confusion: typefaces of vastly different construction are corralled under a single heading (“Slab Serif” includes 19th Century Egyptians as well as 1974 Lubalin Graph), modern designs not explicitly designed in any historical style are grouped with the oldies (Gerard Unger’s Swift typeface of 1985 is considered a “Garalde” ), and designs which don’t fit any of the molds are summarily expelled. Not all the exiles are freaks, as can be gathered from any type specimen book which includes special designations like “Twentieth-Century Romans,” a questionable grouping usually included for the express purpose of accommodating Times Roman.

Much of the scramble to organize typefaces has come from type manufacturers, since commercial typefounding brings a special urgency to the question of how typefaces are listed. No discussion of classification would therefore be complete without at least a little revue of some of the more colorful developments in the private sector. Bitstream’s approach is rather ecumenical, giving equal weight to sans serifs and “typefaces adapted from typewriters and daisy-wheel printers,” among other comprehensive categories (although things take a turn for the worse in category 15, where “typefaces outside the typographic norm” are ghettoized as “Exotic” .) Letraset forecloses the discussion with their “Graphic” category, which warehouses some two-thirds of the faces in its library falls. The FontShop’s FontFont range has recently been reorganized along conceptual lines, bringing to typography an updated version of the nine muses: “Amorphous,” “Destructive,” “Geometric,” “Handwritten,” “Historic,” “Intelligent,” “Ironic,” “Pi & Illustrations,” and “Typographic.” (Note that a typeface cannot be both Historic and Intelligent, and that only some typefaces get to be Typographic.) Perhaps it was inevitable that the beloved Dewey decimal system would be interpreted typographically, as David A. Mundie has done with A Field Guide to Type Classification, published on the Internet. Beginning with standard morphological differences (serifs vs. sans), Mundie provides subdivisions for progressively subtler typographic distinctions, right on past several decimal points to an almost fractal level of refinement. The result is that Univers Black, which Adrian Frutiger outfitted with the workaday number 75, is designated: “, R flag roman [498].”


In The Elements of Typographic Style, Robert Bringhurst outlines an interesting approach to type classification, which lays the foundation for a more sensitive conservancy of type history. Recognizing that the period of a typeface’s development will reflect not only its style but its manufacturing process, Bringhurst introduces two complementary scales for classifying type. The first is a series of artistic movements, borrowed from the fine arts: Renaissance, Baroque, Neoclassical, and so on. These labels ignore the vagaries of x-heights and serif designs, and instead attempt a reference to the artistic circumstances relevant to each period in typographic history. Intersecting this system is a second scale, which indicates the original form of a typeface: foundry type for hand composition, hot metal type for machine composition, phototype, or digital fonts. By classifying typefaces under both methods, a system emerges which allows for the comparison of visually related designs without overtly suggesting relationships between them. Therefore Fournier’s types of the 1740s might be stylistically grouped with their 1926 interpretation for machine casting Monotype Fournier, as well as Fleischman’s No. 65 (1738), not to mention more recent designs that are only arguably their descendants, such as Matthew Carter’s Charter (1987). By locating history along perpendicular axes denoting artistic and technological concerns, Bringhurst dispenses with the usual laundry-list approach and hints at something far more compelling to emerge.

Since typeface classification systems are rarely constructed with concern for the needs or interests of their users, let alone anticipating their level of sophistication, they rarely raise the question of who might benefit from their use. Mundie’s Field Guide to Type Classification, for all its obscurity, is one of few systems that directly confronts the question of readership. Presenting his system as a field guide rather than a comprehensive taxonomy, Mundie makes a clear distinction between handbooks for beginners and reference works for the proficient. His botanical parallel is a good one: “[A] field guide is free to use categories such as ‘pink flowers,’ while the taxonomic botanist is apt to concentrate on whether the pistils are adnate to the perianth.” The student learning to recognize typefaces undoubtedly has different concerns from those of the art director who is likely to be more interested in the stylistic and cultural references of a particular font, as well as its location in a catalog. The historian’s interests are different still, as are those of the bibliographer, the cultural anthropologist, and the increasing number of nonprofessionals who have discovered an interest in type through their font menus.

A New Reference

Questions about organization and audience amount to curatorship, an aspect of type classification that has never really been articulated. “Lettering is often interesting from many points of view,” wrote Leonora Pearse about the Central Lettering Record, “so to make a classification according to one aspect of style may in fact prove more restricting than helpful.” Mindful of the fact that classification systems inevitably leave an editorial imprint on the material they organize, the research group at Central Saint Martins has allowed their system to evolve with the understanding that when complete, it will serve as only one of many means for navigating the history of typography.

Although rooted in the archival resources of the Record, the cd has evolved in an atmosphere which strongly reflects the human resources of the college. Under the stewardship of Jesse Collins, Herbert Spencer, Anthony Froshaug and most recently, Phil Baines, the College has engendered an interest in typography, perhaps overshadowing the other lettering arts. Despite the cd’s focus on printing types (especially those of the past fifteen years), it shares with the Record an approach to type history integrating the new and the old. “Our first thought,” says editor/designer Eric Kindel, “was to provide the students with an overview that incorporates the familiar with the obscure, such as historical background material.”

The cd consists of four basic sections. The first is an electronic specimen book which contains more than 150 fonts (recorded as anti-aliased bitmaps) whose characters can each be fully displayed at enlarged scale and alongside the same characters in other typefaces (as many as eight at a time). In the second section, most of these faces are presented in the context of original use, i.e., in books and ephemera gathered from the Record. Included as high-resolution scans, these examples come from a variety of sources, including not only type specimen books and type foundries’ promotional material, but posters and ephemera. A split-screen interface allows both sections to be viewed at once, thus a reader might examine Monotype Bembo alongside its historical progenitor, a type used by Aldus in the Renaissance novel Hypnerotomachia Poliphilii, which is shown as a striking enlargement. A third section contains historical synopses of all the typefaces included on the cd (provided as searchable text), and the fourth and most comprehensive section contains an elaborate timeline of type designers. The cd also offers various other materials, such as QuickTime animations (no cd is complete without them), which generally support the larger reference work.

The job of distilling half a millennium of type history down to 150 typefaces is an unwelcome task, but one that has been skillfully handled by Kindel and Dixon. Among the typefaces chosen are not only those which have enjoyed commercial success, but those which represent significant developments in technology. The technological front represented by Monotype Imprint (the first custom-designed typeface for machine composition) continues with Apollo, the first face explicitly designed for photo composition, and on to Adobe Minion, the first Multiple Master. Other aspects of the relationship between design and technology are introduced through the inclusion of machine-readable fonts such as ocr-a and ocr-b, and machine-inspired fonts such as Citizen and Beowolf. Still other typefaces are included as examples of different perspectives on a historical model, such as the three Bodoni revivals, the early Bauer Bodoni, the more recent ITC Bodoni, and the “bastardization,” as Kindel puts it, Monotype Bodoni. Typefaces which are crucial to the discussion but remain unavailable digitally are illustrated in the first image section, such as the generous number of Garamond revivals which are shown as foundry specimens.

Unlike digital publications which are fundamentally books or educational films, the appeal of the Central Lettering Record cd is that it is presented not as a story, but as a database. Unlike conventional type histories in which typefaces are arranged chronologically, gathered stylistically, or merely listed alphabetically, the materials on the cd are not offered in linear order. The hypertext structure of the cd requires that the reader leap from font to font in pursuit of points of comparison, and the narrative about type history that emerges does so not on the screen but in the user’s mind. “Our hope was to put in as much as possible,” says Kindel, “and to allow the connections to form for themselves. There are certain ideas we’re hoping to draw out – namely the context in which the types have evolved, the technology and influences underlying them, what predecessors’ types represent, and what they themselves were based on – but I’d argue that we ultimately take no opinion about the merits of the material included.” This approach is less laissez-faire than it sounds, for included with the cd may be a series of scholarly essays that will refer to the material on the disc in different ways. These routes through a common terrain might be thought of as typographic travelogues, and their parity with the user’s journey serves to remind that the hierarchies posited by any type history, and every type classification system, are highly subjective.

A New System

With the imperative to present contemporary and historical material in the same source, chief content researcher Catherine Dixon set about developing a classification system as a supplementary means of accessing the material on the Central Lettering Record cd-rom. Since its conception, the system has undergone a number of changes, many of which are a response to the research team’s own explorations of their cd, explorations which continue to suggest new ways of organizing material. Many of the original categories which reflected typefaces’ formal qualities have been discarded in favor of more abstract terms which take into account their intentions. The “Industrial Vernacular” class, represented by Erik van Blokland’s Trixie typeface, has since been absorbed into a new category, “Emulative,” which is less wedded to a particular fashion. “Categorizing Trixie as an ‘Emulative’ typeface suggests that its employs strategies that may have been tried before,” says Dixon. “There are countless Letraset faces which are emulative – all those LCD types, for instance – but if you think about it, Gutenberg’s type is really emulative, as well.” The “Sampled” category has been similarly redefined, no longer referring only to typefaces made from other typefaces, but now encompassing designs that “take an aggressive stance towards history,” as Dixon puts it. Whether it will be possible to uncouple “Sampled” from a larger cultural ethos has yet to be seen; like “Industrial Vernacular,” it connotes strong existing cultural and temporal associations. And the part of the project that involves determining conceptual underpinnings remains slippery, as Dixon is quick to point out: “It’s a major trap, talking to designers about what they’ve done. The hype is largely after the fact.”

The latest iteration of the system calls for a plane, formed by the intersection of a timeline on the one axis, and a list of identified typographic models on the other axis. Like Bringhurst’s approach, typefaces are evaluated according to multiple criteria, but the cd’s system takes a unique approach to how they are plotted on the grid. Historical sources might be represented by a single point on the grid – a foundry type made by William Caslon, for instance, might be at the intersection of a.d. 1732 and “Garalde” – but later designs that refer to historical material in different ways might be represented by a range encompassing a number of points, each of them a typographic reference. The corner of 1732 and Garalde might find itself a part of many neighborhoods, appearing in the profiles for not only explicit revivals like Adobe Caslon and Caslon No. 540, but more ironic interpretations like Mark Andresen’s Not Caslon – and, even more unexpectedly, the Kennerley typeface which Frederic Goudy claimed as his homage to Caslon. The system becomes especially interesting when these sorts of unforeseen patterns emerge.

As the system continues to evolve, it threatens to become increasingly difficult to represent visually. Recent review suggests the need for a way of displaying non-visual references, which the team now hopes to incorporate as a listed appendix – time permitting, as the project’s funding expires this summer. What happens then remains uncertain, but the team hopes to develop subsequent releases of the CD, some of which might be necessitated after the profession has had a chance to respond to the classification system. “In conversation with type designers,” notes Dixon,” it has become clear that they aim for the cracks between categories within type design, combining elements of this with elements of that and so on.” In spite of its flexibility and openness, perhaps this classification system is raising the bar, encouraging today’s samplers and emulators to find new ways of hacking the system tomorrow. I addition to thriving on challenge, type designers are, to a person, smart alecks: the chance to have an entire category named after them in version 2.0 may prove irresistible.

Jonathan Hoefler is the director of The Hoefler Type Foundry, a design studio specializing in the development of typefaces in the Proto-Hoeflerist style.