Type Classification and Bringhurst

John Boardley's picture

Re-read this today,

Rigorously scientific descriptions and classifications of typefaces are certainly possible, and important research has been under way in this field for several years.--The Elements of Typographic Style, page 121, version 3.1

Anyone know which research he was referring to?

Miss Tiffany's picture

Bringhurst is referring to such work as the Vox system. Catherine Dixon gave a very good talk about it at St Bride in 2002. Typeface classification/Catherine Dixon

Michel Boyer's picture

The Panose system describes fonts with numbers. Whether it makes it more scientific is arguable.

[edit] There is a nice section on font classification in Haralambous' book Fonts and Encodings.


paul d hunt's picture

typeface classification schemes
(just for others' reference)

John Hudson's picture

The Panose system describes fonts with numbers. Whether it makes it more scientific is arguable.

The purpose of Panose is to map similarities of weight, width and other proportions, along with some stylistic details, between different fonts. Insofar as it does this based upon precise measurement of these features, calculation of proportional ratios, etc. it is objective and methodical (i.e. what people tend to call 'scientific'). Of course, the goal of Panose was font matching and substitution: a way of coming up with the 'best fit' font if the original font in a document were missing. So it works -- assuming font developers accurately recorded the Panose data, which most did not -- within that conceptual and practical framework. It isn't a classification system per se, and the numbers won't necessarily tell you what you want to know about a typeface, but it should tell you what other types are similar in certain ways.

Unfortunately, like most type description or classification schemes, it breaks down completely beyond the Latin script.

Nick Shinn's picture

There are some expectations about type classification which assume that a scientific method would produce accurate identification of type specimens.

But the distinction must be made between classification and identification, and between typographic images and fonts.

Trying to identify a typeface from a specimen of a few letters, using a classification system, won't get you very far without expert knowledge. But the same could be said for trying to identify an animal from its tracks--there's a lot of missing information.

William Berkson's picture

In the mid-West where I grew up, we had a useful classification category that I think applies to any claim that a system of classifying typefaces is "scientific."

That technical term is "BS" :)

Such classifications are useful or not, and for different purposes. That's it.

Si_Daniels's picture

The Panose spec, eventually landed with Monotype Imaging, via HP via Elseware (A Seattle company founded by ex-Adobe people)...


It's primary author Ben Bauermeister is back at Adobe. We'd hoped to have him on a panel at TypeCon but sadly he had other commitments.

John Boardley's picture

Thanks guys. I must admit to being a little surprised at,

Rigorously scientific descriptions and classifications of typefaces are certainly possible...

I guess it depends on how one defines 'rigorous'.

Thanks for all those links too.

An important distiction, and a rather nice analogy. Thank you.


dezcom's picture

It seems the the moment you have a classification system is the moment the arguments begin as to what fits where. There are always faces that don't fit the scheme exactly and the time and energy it takes to agree to disagree on which pigeon hole to select far outweighs the value of the system. This hardly seems scientific to me but I am not a scientist. As long as we don't take the persuit too seriously, there is some basic value in general categorization just for the sake of description. If you have the argumentative mindset, you might like dueling intellectual arguments over categories but you really have to enjoy that sort of thing.


eliason's picture

Developments on the web have changed some approaches to classification that might be worth pondering in relation to type. Here's an interesting article: Ontology is Overrated
Shirky's observation about the Balkans and Asia made me think of Vox's categories of Garaldes and sans-serifs. Seems like Shirky identifies some of the same types of pitfalls that Dixon did. But following his lead in envisioning a folksonomy of type looks like a very different kind of solution than Dixon seems to be proposing. (The link to the pdf with more details in the Dixon article seems to be broken, so I'm a bit mystified about what her solution actually is)

Rob O. Font's picture

Back at the top..."Rigorously scientific descriptions and classifications of typefaces are certainly possible..." says RB, but if one asks for a rigorous scientific description and classification of a typeface...what would that be? The periodic table of typographic elements, Elements of a Type's Style has not yet been written, and even it was, without it's companion volumes, Elements of a Type's Use, and Elements of a Type's Output, what do you know?

Panose, e.g. asketh, and I slightly paraphrase: What number (from 0 to 9) would best describe the X-height, (i.e. the relative height of lowercase characters, presumably to the uppercase;) and the treatment of uppercase glyphs with diacritical marks. That's one number, like it's a toothbrush and a stun gun. John Hudson points out that Panose breaks down "completely" beyond the Latin script. It breaks down partially before it finishes the Latin type circa 1994...and in classification theory, partial breakdown and complete breakdown can be hard to distinguish, I think.

Dez says, "It seems the moment you have a classification system [...] the arguments begin as to what fits where."
My experience is quite different. Conflicts occur naturally in the formation of classification systems that, by the nature of what they are trying to classify, cannot be complete or fully described, variety in the characters themselves making rigorous scientific description impossible. Nevertheless, once the system is completely argued out — what fits where, is over, the arguing stops, and the ignoring begins. ;)


dezcom's picture

"...the ignoring begins."

I certainly agree with that :-)


enne_son's picture

Type varies across numerous distinguishable dimensions. The most important of these are contrast and construction. These alone make classification possible and useful. If you are insensitive or not attuned to types of contrast or to patterns of construction you loose your grip on form and your sense of where to go. The effort to categorize alone can be enriching, because it forces attention to distinguishable dimensions. So this has a pedagogical benefit. Without it the 'kinship domain' of type is almost unsurveyable, and it becomes difficult to get yourself oriented or locate oneself within it.

The danger is that a scheme that codes for too-few dimensions or gives pre-eminence to subordinate dimensions blinds one to the relevance of important others.

Classification or filling-in a niche is not a destination, just a servicable and sometimes fertile strategy or heuristic.

Nick Shinn's picture

John Hudson's pastiche of Borges:

Syndicate content Syndicate content