Typeface classification/Catherine Dixon

The classification of typefaces is one of those topics which proves enduringly attractive in typographic debate. Yet, beyond such abstract discussions, the issue of how we might order type has very real implications on the way we write about and teach typographic history. It was this very practical aspect of classification which prompted my involvement in the field as I sought to address the particular problem of how to make accessible to students some understanding of the great diversity of typeforms now available to them.

My purpose here is to share the results of my experiences in developing a new description framework for typeforms - and here I should say Latin typeforms - with a view to prompting feed-back and a future pooling of knowledge, so that these ideas can be refined still further. And, while my interest really lies in the ongoing development of a practical structure for describing typeforms for the twenty-first century, the origins for such structures are an essential part of the typographic development of the twentieth.

Typeform description in the twentieth century

In 1995 I began work on a cataloguing programme for an archive, the Central Lettering Record, based in London at Central Saint Martin’s College of Art and Design. I was updating the archive, collecting examples of type design from the previous ten to 15 years or so. This material then had to be catalogued alongside the older historical typeface material, for input into a new database.

The database required a series of descriptive fields for typeforms - new and old - and that is when I first encountered the British Standard (BS 2961:1967). Common practice determined that the generic description of typeforms should use the category terms borrowed from classification schemes. And while it may seem strange to have turned to a system then nearly 30 years old, both the British Standard and the Vox system upon which it is based were then very much still in evidence.

The British Standard tells us a huge amount about attitudes towards the description of typeforms in the twentieth century, against many of which attitudes my work is in direct reaction. Here, I will focus upon just three essential aspects of the British Standard: its reliance upon a so-called ‘top down’ approach to categorisation; its embodiment of values which suggest that some types are more deserving of detailed description than others; and what it tells us about recent attitudes towards classificatory change.

First, the reliance of the British Standard upon a ‘top down’ approach to categorisation. The British Standard system was a very basic interpretation of the earlier classificatory proposals of the French typographic historian Maximilien Vox. These proposals, put forward in 1954, were far more sophisticated than it is often realised. Offering at first a ten-part classification, Vox revised his original proposal within months to a more compact nine-part scheme. The ‘Classification Vox’ or ‘Classification de Lure’ as it was also known, can be seen as Vox’s attempt to reflect his understanding of a complex and entangled field in which types were ‘living’ and as such susceptible to stylistic-interbreeding. His response was to address the similarities between types and also the possibility of their individual differentiation. His system combines a ‘top down’ classificatory approach whereby type content is divided across a ready-built infrastructure with a ‘bottom-up’ approach whereby the set category terms could be more fluidly used in combination to cater for those typefaces exceptional to the established norms. His intention was not a system confined to nine categories, rather one which could offer nine multiplied by nine, or still further multiples, if required.

Vox’s proposals provided a focus for classificatory ideas like no others before, and so by the time Britain came to publish its own national standard, adoption of the basic Vox categorisation schema was almost inevitable. But what was adopted in Britain, as in so many other countries, was no more than the basic nine category structure. Most of Vox’s intuitive strategy for being able to respond to the needs of individual typeforms through the multiple use of his category terms was lost, if not in direct translation then certainly in implementation. And Vox’s basic nine category structure - without his intuitive application - was really nothing more than nine descriptive buckets in which types could be placed.

The second main point I want to make about the British Standard is the way it embodies a judgement that some types are more deserving of description than others. The categorisation of typeforms grew out of a changing climate in production, when, during the nineteenth century, printers experienced a broadening in the range of typefaces at their disposal. Types were being introduced which were intended for setting at large display sizes as well as for book work. Type styles were also being deliberately ‘revived’ from earlier historical periods. But there was little consistency between type foundries in the terms used to describe them. Sans serif types, for example, were variously marketed as ‘grotesque’, ‘sans surryphs’, ‘gothic’, ‘doric’ and even ‘egyptian’ even though the latter term is more familiarly associated with typeforms with slab serifs. As the numbers of these new and ‘new-old’ types increased, it became necessary to find a way of ordering type, to ease communication between printers and clients, and as an organisational aid within the printing industry.

By the turn of the new century two new publications offered typeform classifications: Southward’s Practical Printing from 1898 and De Vinne’s The Practice of Typography: Plain Printing Types, of 1900. Written for the benefit of printing practice, by printers, each classification took its lead from practice, employing terms then in common use.

The typefounders were quick to join the printers in taking up the classificatory challenge. In 1903 the French typographer Thibaudeau devised a system for the historical material of the Peignot foundry and in 1911–12, HL Bullen worked on a similarly analytical system for the library of the American Typefounders (ATF). Both systems also introduce a new consideration of presentation, making good use of diagrammatic illustration to reinforce their underlying rationales.

Then there followed the scholarly surveys. There was a need in the early twentieth century to consolidate the different strands of historical research, following the interest in reviving early types for contemporary use. This resulted in DB Updike’s Printing Types of 1922, Morison’s On Type Designs Past and Present from 1926 and AF Johnson’s Type Designs from 1934. Within these historical overviews, the morphological evolution of typefaces is used as an organisational structure, each key stage in the development of typeforms being summarised by a category and typically a chapter.

Implicit in all these early systems, and especially in the latter and most influential historical models, was an evaluation of typographic history and current practice now nearly 100 years old. Yet, it is their common category structure which has continued to inform the underlying schema - if not the exact nomenclature - of most later systems.

For example, early twentieth century evaluations of both historical and contemporary design practice favoured roman types over display, for reasons both aesthetic and commercial. This basic premise remained largely unchallenged, informing the basic Vox categorisation and that of the British Standard. If you consider the British Standard, ‘humanist’ types are formally distinguished from ‘garalde’, even though the formal differences are very subtle and such a distinction is only appropriate for very few types. But large numbers of slab serif types, clarendons or ionics (that is bracketed slab serifs) and egyptians (that is square-ended, unbracketed slab serifs) are simply grouped together.

A contemporary evaluation of typographic history and current practice would or should, I think, reveal that roman types are no longer a central concern. Display types have not only encroached upon the scholastic exclusivity that roman enjoyed, but also upon the commercial monopoly that they once held in book printing. Distinctions between text and display are now increasingly irrelevant, with the greater subtlety that has been introduced into sans serifs and slab serif designs, leading to a wider application of such types for text purposes. And as text is increasingly found in environments outside of print, so typeforms are being modified to the requirements of new technologies, the screen being the most obvious of these. More generally, type design has, with the advances made in digital programming, been made accessible to anyone with the necessary software. Reduced overheads combined with ease and speed of production and distribution have seen the market broaden considerably. In these competitive times when traditional roman text faces are so widely available - it seems that every manufacturer has its own version of Bodoni and Baskerville - the emphasis in the market-place has shifted from roman toward ever greater novelty in form.

Yet these shifts have not been matched by a similar shift in the approach taken to typeform description. And this leads to my third main point about the British Standard: what it tells us about recent attitudes to classification.

The British Standard was published in 1967 and has remained unchanged. Yet to try and implement this standard within current contexts is hopeless. Using the British Standard within the Central Lettering Record, I found it to be inappropriate for much contemporary typeface design - too much recent material could only be accommodated within the ‘graphic’ category. The diverse range of typeforms caught in this classificatory ‘catch-net’ were not actually being described at all.

Alternatives to the existing systems have been put forward, but have struggled to find the support required for wider implementation. The authors of a recent study of types, for example, praise the classification work of Gerrit Noordzij as offering a more intelligent approach, but in finding it hard to apply, continued to persevere with the German interpretation of Vox.

For some, it seems acceptable to retain classificatory structures that exclude this new dimension of practice. And here classificatory apathy, combines with the idea of classificatory censorship to explain the inadequacies of many of our existing systems. The direction in the pursuit of formal novelty has sometimes proved so controversial as to call into question whether such types should even be acknowledged, let alone afforded descriptive attention. So more recent developments are generally omitted from the broader view and considered in isolation, if at all.

Failure to incorporate such types within historical surveys has created an artificial end-point in typeform history, with much contemporary practice entirely dislocated from that of the past. Alexander Lawson’s Anatomy of a Typeface from 1990 refrains from discussing more recent types. But it is the unrevised publication in 1998 of Geoffrey Dowding’s Printing Types, originally from 1961, that best epitomises this stagnation in historical accounts of type design.

But where to start? It is perhaps with this question and the prospect of the sheer enormity and difficulty of the situation that we find the overriding reason for so little classificatory advancement in recent years. Certainly the situation seemed one in which there were more questions than answers. In defence of his use of the open-ended term ‘decorative’ (effectively in place of ‘graphic’) Lewis Blackwell acknowledges that: ‘It sounds like a loose term, and it is - but how do we describe the rapidly increasing number of founts that do not draw on one particular historical tradition or form of production, but are distinguished by being sports that draw on the varied visual culture of their time?’

A new response

The increased individuality of form within type design during the 1990s has prompted a number of more recent classificatory proposals. Within these systems, such is the emphasis upon the individual typeface, however, that formal description can tend towards the abstract.

Keen to avoid such abstraction and communicate type description information in ways accessible for users both familiar and unfamiliar with type I formulated a proposal for an alternative response. Within my own description remit, I wanted to provide a description framework able to overcome the fracture in contemporary discourse and make sense of what was happening now, a framework able to consistently describe the formal character of all typefaces across an approximately 550 year production period. (In very broad terms, the formal character of a typeface can be understood as being how the punchcutter or designer intended it to be seen in use.)

Work began — badly. My own start was determined largely by early project time constraints. And so perpetuating the ‘make do and mend’ philosophy which had come to dominate the classificatory field, I returned to the British Standard, and adopted the simplest strategy of trying to modify the existing schema. The contents of the overloaded ‘graphic’/catch-net category were accepted and dispersed across a series of newly added categories. These ideas were summarised and published along with a working proposal for a revised set of categories in Eye 19.

But trying to modify only one category led, in the end, to failure. The problems of the old system were not confined to one part, and the instability of the whole only became more obvious in the process of trying to add to it. Of no lasting value in itself, this early proposal did come to represent a turning point in the design process. Two key areas for re-evaluation were highlighted: First, categories as a mechanism for describing typeforms; and second, visual presentation in the communication of typeform description.

The work on extending the existing schema showed that categorisation, certainly a similarly ‘top down’ approach, could not provide a long term solution to the existing description difficulties posed by the ever-broadening range of sources used by type designers. The extra categories initially proposed scarcely scratched the surface of the problem. Yet to increase the number added would lead to an unwieldy system: too many categories each with too few typefaces. Acceleration in the field of typeface design/production meant that to keep a system updated would anyway be impossible. Even an ambitious and admirable system such as that in Rookledge's international typefinder, designed by Christopher Perfect & Eichii Kono, has, without the means of regular revision, become quickly outdated.

Nor could categorisation offer a practical or sustainable response to the hybridisation of form characteristic of much recent practice and of twentieth century type design practice generally. Although in part fuelled by what Robert Bringhurst describes as a ‘surfeit of historical awareness and self-mockery’ symptomatic ‘of the phase we call postmodernism’, the lifting and mixing of formal elements from other letters or sources within type design is neither transitory or new. In her study of nineteenth century display types, Nicolete Gray reveals the origins of this ‘new freedom’ in the introduction of the continental Latin-Runic typefaces, commenting that, ‘although so far no very drastic changes have been made, categories are becoming blurred and classification complicated; a new era has begun’.

The difficulties I was facing became all too clear when I saw the visual presentation of my early proposals in Eye. I quickly recognised that the regimented series of uniform, immutable and insular boxes belied the complexities of the narratives being drawn out in practice. A more accurate visual presentation of these narratives would, I felt, encourage the subtlety of description required.

The visual mapping of classificatory ideas is nothing new. The early systems of Thibaudeau and Bullen both utilise visual methods, though perhaps the most outstanding example of visual mapping in the field is the contribution made by Beatrice Warde in 1935. Building upon the morphological principles of her former ATF colleague Bullen, Warde developed a hierarchical model for the classification of text faces which shows categories within a visual structure explaining their formal descent. Visual presentation here is not simply a secondary illustrative device, as in the case of Thibaudeau. Understanding of Warde’s system is dependent upon its visual layout, the relative positioning of the categories to one another being an essential part of her overall thesis. With this work Warde established an important precedent for visual classificatory argument. Jan Tschichold and Rudolf Hostettler certainly made good use of layout in presenting their schemes of the 1940s and I would also recommend looking at the simple though effective family tree ideas explored by Lawrence Wallis, writing as John Wulfrun for Print in Britain during the late 1950s and early 1960s.

For my own part, the exercises I undertook in playing with the relative positioning of classificatory categories led to their more formalised diagrammatic representation, determined by hierarchical and historical relationships. Then as it became clear that many types required a finer degree of descriptive focus, the emphasis shifted towards a visual analysis of the criteria defining these categories. The result was the identification of a series of three description components - sources, formal attributes and patterns - which together form a new description framework [see Visual overview pdf].

Sources describe the generic influences informing a typeform. They were seen as a series of larger groups into which the existing categories could be ordered: roman, handwritten, nineteenth-century vernacular, and so on.

Formal attributes are the basic individual units of description that refer to a typeface’s design and construction. Eight main kinds of these attributes were identified - construction, shape, modelling, terminals, proportion, weight, key characters and decoration - each with a further sub-menu of its own.

The new description framework operates on the assumption that the formal character of every typeface can be explained in terms of its specific configuration of these two components: sources and formal attributes. However, while providing the micro level of description so often required, too great an emphasis on differentiation tends to obscure the macro view of a given type’s shared formal relationship with others. For this reason the third element in the new framework, patterns, list the most common recurrent configurations of sources and formal attributes.

The whole description framework is presented in diagrammatic form, providing a map which shows the range and contextual relevance of each of the three main description components. The diagram also shows how, while patterns are key to understanding the early centuries of typeface design, from a certain point onwards, consistent relationships between formal attributes and sources dissolve amid the diversity of practice. The diagram also serves to more accurately represent the history of type design practice as a series of parallel developments rather than, as is typical, a cohesive story in gradual evolution.

Everything about the new description framework - the organisation, utilisation and presentation of its information - enforces the intention that it should constantly reveal how it is working. Description is not centred on an inflexible structure within which typefaces are made to fit, either conceptually or visually. Typefaces can now be brought to the framework, which provides a system of reference against which they can be examined individually and a description built to the requirements of each one.

The future

As any other classificatory endeavour before it, this new description framework is a product of its purposes and of its time. And as Walter Tracy so rightly points out, such a system should not be an end in itself. Tidy-mindedness for its own sake leads nowhere. But this new description framework certainly provides my students with a structure for learning more about typeforms. However, there is still plenty of room for improvement and development. Nomenclature is one of the most controversial areas in the field of type design, and needs to be addressed. And then the formal scope of typeface design has not been determined once and for all. It is only a matter of time before new classificatory challenges will appear, and our understanding of typeforms will need readjustment again.

The new framework, while fully operational, should not, though, be considered as a complete system. I am well aware that it will need to be added to. So in trying to overcome the closure characteristic of the pigeon-hole approach to classification, a flexibility has been built into the new framework, to address this issue of change. Additional description components can simply be added as needed, and without undermining the existing content. And that is where maybe this system does depart from others before it.


Catherine Dixon is a designer and writer working, researching and teaching typography in London