Joseph Moxon said it best in 1683:
A font of type is, as noted elsewhere, a unit of purchase or acquisition. The primary characteristic of a font is that it is finite. In metal type for handsetting, a font is something from a particular typefoundry which includes just one typeface at just one variation (e.g., roman, but not also italic) at one weight (e.g., bold, but not also light or extended) at one body size (e.g. 12 point, but not any other size). Sometimes a font would include uppercase, lowercase, figures (numbers) and points (punctuation), but at other times typefounders would sell these separately. A font includes only a particular set of sorts.
The distinction between a font and a typeface is more, not less, important in the digital world. While digital type removes almost all of the limitations of finiteness present in metal type, one remain: the finite set of sorts. As it becomes more common to use type multi-lingually in the 21st century, the fact that every font is a finite set of sorts becomes an important in sitautions which earlier printers could often simply ignore.
A "font" of matrices (or matrix font) is the set of matrices from which a typefounder casts a font of type. It is subject, typically, to all of the restrictions of finiteness noted above. (Save that when a typefounder sold uppercase, lowercase, figures, and points separately, typically the matrix font owned by the typefounder would, in fact, include them all.)
For the Monotype display casters (the Type-&-Rule Caster, the Giant Caster, the Supercaster, and the Thompson Type-Caster) for sorts casters now of historical interest such as the Nuernberger-Rettig or the Compositype, for traditional foundry pivotal casters, and for "foundry automatic" casters such as the Barth or Kustermann, a matrix font includes one matrix per sort (because the machine casts one type at a time).
However, for the Linotype and Ludlow (and also the now very rare Linograph), a matrix font typically includes many matrices per sort (because the machine casts an entire line of type at a time, and each character in the line requires a matrix).
When casting from Linotype or Ludlow matrices, a typefounder will often distinguish between an ordinary font of matrices intended for casting composition and a "typefounder's font" of matrices which contains only one matrix per sort.
In the digital world a series is most commonly simply a single size scaled linearly to other sizes. More sophisticated scaling, as was common with metal type, takes into account the fact that the proportions of the letterforms should differ at various sizes.
However, the idea that the members of a series are mathematical scalings of each other (however crude or sophisticated) is itself a modern idea. It cannot predate the introduction of mechanical pantographic methods in the making of type (roughly the 1890s for metal type and the 1840s for wood type). Prior to this, and occasionally thereafter, each member of a series was created individually, harmonizing with the other members, but not generated from them.
This can be seen in Bodoni's Greek types from the 18th and early 19th centuries, for example, where different members of the series are simply not the same. Indeed, a close look at 18th century specimen books reveals it for all types now commonly (retroactively) grouped as series. In a more modern context, the 19th century ornamental face Rustic No. 2 ( Figgins), revived in the 20th century as Rustic ( Phoenix/ Los Angeles/ Barco/ Skyline) has different ornaments on different sizes. Ornaments and other design elements appropriate at one size do not necessarily work at another size.
Note also that it was not until the 20th century (with some precursors in the late 19th) that it was common for the various members in a series to line (share a compatible vertical alignment) with each other, or with other types. While in our simplistic way we expect this today, the introduction of "standard line" by American Type Founders (and equivalent standard lining by other typefoundries and matrix makers) was the subject of much controversy (indeed, outrage) at the time. While type made to a standard line is incontestably easier to use the question of whether or not it can be as good as type designed to line individually for each member of a series is unanswerable.
The terms "font" and "series" are essentially typefounder's terms; terms used by people who make type. They describe matters of manufacturing. The term "family" is essentially a type designer's or type historian's term; used by people who classify type. It describes a matter of opinion.
Since the creation later in the 20th century of comprehensive and visually very unified families such as Univers, we tend to think of a typeface family as being so unified, but this is not necessarily so.
Indeed, although an individual manufacturer would certainly only include their own members of a family when speaking of a family, by extension it is not unreasonable for users of type to speak of a type family as including series from multiple, competitive manufacturers. (Users of metal and hot metal type, of course, recognize that while this is true in the abstract, there are significant practical issues involved - particularly the compatibility of lining.)
(But note that when a series has members which differ not simply in size and proportion but in features, as Bodoni's Greeks or Figgins' Rustic No. 2, the digital-era preconceptions about "typefaces" must be modified to accomodate the richer realities of the 18th and 19th centuries.)
Consider, though, the case of a typeface made and re-made over many years, by many manufacturers, with many variations. "Caslon" is a valid designation for a typeface, but it encompasses many different families of type from many manufacturers over the course of several centuries. So here typeface is a superset of type family.
But consider a small, important counterexample: Warde's Arrighi was cut some time after Rogers' Centaur, with the intent that it serve as a "companion italic" for Centaur. They are companions, indeed, but they are not the same: Arrighi and Centaur are different typefaces. Yet both might reasonably be termed members of the "Centaur family." So here typeface is a subset of type family.
It is necessary with a term such as "font" to be precise, because fonts are made things, and making requires precision if it is to work at all. With a term such as "typeface," such precision is not necessary (and perhaps not desirable) because typeface is a term for the discussion of type - and when talking about something you can get away with ambiguity in a way that you cannot when making it.
All portions of this document not noted otherwise are Copyright © 2011 by David M. MacMillan and Rollande Krandall.
Circuitous Root is a Registered Trademark of David M. MacMillan and Rollande Krandall.
This work is licensed under the Creative Commons "Attribution - ShareAlike" license. See http://creativecommons.org/licenses/by-sa/3.0/ for its terms.
Presented originally by Circuitous Root®
Select Resolution: 0 [other resolutions temporarily disabled due to lack of disk space]