Blue Pencil no. 13—Standard Deviations (the exhibition)

Standard Deviations: Type and Families in Contemporary Design
Museum of Modern Art
Paola Antonelli, Senior Curator and Kate Carmody, Curatorial Assistant

The Museum of Modern Art exhibition showcasing its new digital font acquisitions contains a short glossary of type terms. It is not only inadequate but inept. Here are all of the definitions and my comments on some of them.

Bitmap typeface 
A typeface in which the letterforms are composed of pixels, or “bits,” unlike a vector typeface, in which each letterform is rendered as a single outline.

Cathode ray tube (CRT) 
Similar to an early television, a CRT monitor renders images in large pixels on a grid.

[This not entirely true as early CRT machines used scan lines to render images. Cathode ray tubes in typesetting are the same as those found in television sets. The CRT emits a beam of light that is etched on a surface.]

An individual letter, also called a glyph or letterform.

[This collapses the critical distinction between a letter (or letterform), a character and a glyph. A character can be a letter, but it can also be a figure (numeral), a punctuation mark or a symbol. Glyphs, in typography, are graphical units and as such they encompass and go beyond characters to include writing marks in non-Latin languages.]

The part of a letter that reaches down below the baseline of the font, in g, p, and q, for example.

[This definition is slack. It not only leaves out j and y but it could include the tail of Q which descends but is not considered to be a descender. Furthermore, ascender, descender’s more significant counterpart, is not included.

A specific size and style of type within a typeface, for example, Helvetica 10-point italic is a different font than Helvetica 9-point bold. Since the digital revolution, “font” has been used colloquially to mean type family or typeface.

[This is a very confusing and inaccurate definition. The word has shifted meaning over time. Theodore Low De Vinne, in Plain Printing Types (New York: The Century Co., 1899) defines a font of type as “a complete collection, with a proper apportionment to each character, of the mated types required for an ordinary text.” (p. 165) This is a reminder that in metal type you need more than one of each character to compose a text. A font is thus like a Scrabble set with a differing frequencies of characters depending on language (see p. 167). As De Vinne says, “The type-founder tries to supply each character in proportion to its frequency of use, so that the printer shall have enough of every and not too much of any character.” (p. 165) A scheme or bill of type varies not only from country to country but also between metal and wood. But it does not normally vary by point size.

De Vinne shows (p. 166) one scheme of type in which there are 226 characters. A font refers not only to the frequency of characters but also the complete set of characters. De Vinne outlines a “so-called complete font” for roman and italic type of 253 characters (p. 169). Along with letters, figures and punctuation there are fractions, money signs, reference marks (e.g. pilcrows or fists), braces, dashes, leaders, space and quadrats, and miscellaneous marks (e.g. @ or the degree mark). In the digital era font does indeed refer to the design of a typeface—but not to a family of typefaces. Instead a family is a collection of related fonts. And font still has nothing to do with point size since digital type is largely size-independent unlike metal type.]

See character.

[See comment above about the definition of Character.]

Joining stroke 
A line that connects two letters, as in cursive handwriting.

[In calligraphy usually called a join. Joins can exist in non-cursive hands as well as non-script typefaces. See quaint ct and st ligatures as well as the unusual characters in Matthew Carter’s Walker typeface.]

The part of the character that extends outward from the stem. For example, a leg is what distinguishes R from P.

See character.

[See comment above about the definition of Character.]

A single character that represents the connection of two letters.

[Ligatures can consist of more than two letters. This is especially true in textura. Gutenberg’s fount had a number of three-letter ligatures. True ligatures involve overlapping strokes that subliminate the identity of the original letters. This is why ct and st ligatures (and their ilk) are characterized as “quaint”. They are only ligatures by virtue of an extraneous stroke.]

Point size 
The size of a font, based on its x-height. There are 72 points per inch.

[Point size is not the “size of a font, based on its x-height” but, in metal type, of the metal body bearing the character. This height was larger than the distance from the bottom of a descender to the top of an ascender. In digital type the measurement is similar, except that now there is no physical object, just a bounding box. Typefaes with the same nominal point size can have wildly divergent visual sizes. This concept should have been illustrated in the glossary. (Furthermore, it is only with Postscript that 72 points equal exactly one inch. In the Didot system, 72 points equals 1.186 inches and in the Anglo-American system—the one that dominated in this country until the advent of the Macintosh computer—it equals .9936 inches.)]

A short line that extends from the top or bottom of a stroke in a letter. It is a symbolic leftover from handwriting.

[A short line that extends from the top or bottom of a stroke in a letter,” the first part of the definition of serif, is merely incomplete. But the second part—“It is a symbolic leftover from handwriting.”—is laughable. A serif is a tiny stroke (not necessarily a line) that terminates a principal stroke of a character. Serifs are not confined to letters and they may be found on horizontal and curved strokes as well as on vertical ones. They derive from formal lettering, not handwriting; and, although their functional value has been a matter of debate, they are certainly not symbolic holdovers.]

The part of a character that is a vertical line, as in F or h.

[The stem of a letter is also called a spine.]

The lines from which a letter is formed; a lowercase h, for example, is composed of two strokes: a stem and a leg.

[Why is this plural? Depending on the nature of the lowercase h, it can be composed of one, two or even three strokes. A cursive h such as that found in roundhand is a single continuous stroke, complete with loop. A constructivist h composed entirely of straight lines (so that it resembles a chair in profile) has three strokes. See The Stroke: Theory of writing by Gerrit Noordzij for an extended and influential discussion of the concept of stroke in writing and calligraphy and its relevance to type design.]

An exaggerated serif that embellishes a letterform.

[A swash is not an exaggerated serif. It is an extended stroke that embellishes or decorates a letterform.]

Titling face 
A typeface designed to be used in large sizes for titles or headlines.

[The traditional definition of a titling face is a typeface consisting solely of capitals, lining figures and punctuation. Its letters are larger than usual because the absence of lowercase letters (specifically descenders) allows the full area of the typeface (literally the face of the metal type body) to be used. Titling typefaces tend to have stunted Qs and non-descending Js. The name comes from the common use of such large letters on the title pages of books. The concept of a titling face is of little relevance in digital type where size is no longer controlled by the typefounder. The only instance where it still has a bearing is to describe typefaces that do not have full character sets, but instead are limited—such as my Kolo, Donatello and Bermuda—to capitals, lining figures and punctuation. Such typefaces are no longer limited to titles or headlines.]

A set of letters in different sizes and styles, united in form and look, that are designed to be used together. Also called a type family or face.

[Originally, typeface referred literally to the design of the character on the face of a piece of type metal. From there the term has come to mean the design of a group of related characters (not only letters) “united in form and look” but not comprising “different sizes and styles”. A typeface is not the same as a type family. The latter is a set of related typefaces, most often various weights and widths of a roman and its companion italic. Increasingly, the definition of family has been stretched to include serif, sans serif and mixed serif variants. Getting this term wrong undermines the whole notion that Standard Deviations is about types and families.]

Vector typeface 
A typeface in which each letterform is rendered as a single outline, unlike a bitmap typeface, in which each letterform is composed of a collection of pixels.

The height of the lowercase x in a typeface, upon which the heights of all other characters are based.

[This is overly literal and it puts the cart before the horse. The x-height (the z-height in older American books on type) describes the height of the body of a lowercase letter and is only meaningful as a guide to the proportion of the body to the ascender height first, the capital height second and the descender depth third. The height of the x (or the z) is merely a convenience and not what the type designer is really concerned about.]

I sent the Museum of Modern Art glossary and my comments to James Mosley, former librarian at St. Bride’s Printing Library in London and type historian par excellence ( to get his feedback. Here are some of his views on the terms under consideration here. Note that they do not always agree with mine, indicating how difficult some of these terms are to define, both across specialties such as calligraphy and type design but as well as across languages (including British and American English).

“Font” is a curiosity. I instinctively insist that in historical terms it is equivalent to a “casting” of type at one time from one set of matrices, so that another quantity, cast on another occasion, even from the same mats (as we typefounders know them), would be strictly a different “font’, to be kept in different cases—and quite wisely, since it might have come from a different mould, and might have a different alignment and set and even a slightly different body. But I keep coming across historical instances, quite old ones, where “font” or “fount” is clearly a synonym for “design”, as in our rather loose present-day usage. I am not nearly consistent in my own use of this kind of term.

I do find “letterform” a quite useful global term for designs that are common to “type and calligraphy”, attempting to make “letterform” an equivalent to the universal German term “Schrift”—but I am not sure if many people understand what I am trying to do. [I tend to use “letterform” in this manner as well.]

“Ligature” is another fossilized survival from writing and also from metal type, and it can be made from three letters (ffi) as well as two. When the original letters cannot be made out (which is generally true of the roman form of &) I do not think it should be called a ligature—perhaps having become a “glyph”, for want of a better term.

I flatly decline to get involved in defining a “serif”. [Too bad as this is a very contentious term and it could use some standardization.]

Discussing the “point” is pretty pointless, but after that odd period in the 1960s (or when ever it was) when zealous typographic theorists were wanting to insist on a metric scale, we seem to have gone back to them, because they are easy to grasp as a system that gives a pretty universal visual picture. It doesn’t matter if we are using a seventy-second of an inch (which is a sop to the anti-metric US population) or the older point of 0.0138 inch or 0.351 mm (or is it 0.3515 mm?). If I say ten points, or fourteen, you will know pretty well what I mean. The world seems to be accepting that Fournier didn’t invent the point, but stole it (from Truchet). [On this latter point see Mosley’s blog mentioned above.]

The entry for “x-height”—something “upon which the heights of all other characters are based”—is essentially wrong. Reading something as confused as that makes me want to throw the whole thing away, a long way away. Pity, since a good vocabulary of this kind, well-informed and handled by someone with a grasp of clear, simple language, like Carter père (and indeed Carter fils too) [Harry and Matthew Carter respectively] can improve one’s own technical language.