What Makes an Oldstyle a French, Italian, Dutch or English?

brampitoyo's picture


So we've been classifying and identifying oldstyle typefaces using the above terms forever. But let's clear things up -- beyond the usual x-height proportion and greater/lower face contrast and weight -- shall we?

I find this ParaType article to be very helpful as a starting point. My theory is, there's gotta be a simpler way to organize oldstyle faces.

Of course, I could be totally wrong and this is just another futile attempt to reinvent the wheel.

Is there a simpler way to organize oldstyle faces? What makes an oldstyle a French, Italian, Dutch or English.

Don McCahill's picture

Well, Venetian is easy, the cross bar on the e is at an angle. Otherwise it is a Garalde. I've never tried to split into English/French/Dutch.

William Berkson's picture

Well, these things are complicated because there are always exceptions.
The first old style, Jenson does have the angled crossbar on the e, but it also has a lot of calligraphic details.

The typefaces done by Griffo for Aldus were less calligraphic and had the straight crossbar. Garmand and other french had Griffo's type, and if I've got it right basically lightened it a bit, with smaller serifs--an elegance.

The Dutch, influenced probably by black letter, increased the x-height of these designs, increased the weight of the serifs and the general blackness of the letters. They also, as G. Noordzij pointed out, were willing to change the stress on different letters, reflecting writing by pen which changed the angle of the pen.

Caslon was basically imitating Dutch faces, adding his own personality and take on the letters. Baskerville, also English, started a different look by taking Caslon's face, making it wider, with smoother curves, upright stress, and more regularity. Generally a 'neoclassical' feel.

So Jenson=Venetian, the school of Aldus to Garamond is another, and the Dutch are another. Caslon is an English variation on the Dutch. Baskerville is something new. He is taken up on the continent and his ideas taken to their extreme by Didot and Bodoni. There are definitely lines of influence, but each designer had his own take on letters also.

ebensorkin's picture

You might look at the text of Jan Middendorp's 'Dutch Type'. The initial text talks about this a little. The book is expensive but maybe you can find it in a bookstore or a library. Even though it's expensive, I think it was well worth it. Here's a link:


The Dutch Oldstyle is usually described as having shorter ascenders & descenders plus a larger x height resulting in less metal use per glyph and more economical use of paper as well. The Italian & french are descriobed by Karen Cheng as 'lighter & more refined' and the English as more sturdy & less refined.

Keep in mind that none of these generalizations is based on clean or easy separations - and the real data underling these genralizations is complicated & messy. For one thing - lots of the type described as english french etc was actually made in the low countries. Or Germany. But the urge to tell a story & generalize in the face of real complexity is too strong to resist... Unfortunately it is often less than acurate too. But take what I say with some salt. I have only been looking at type seriously for a year or so.

Palatine's picture

I might as well add to this:

How about "German" type? I know Pradell by Eudalde (and now Typerepublic) is a serif face meant to have a distinctly Spanish feel, or at least to recall a distinctly Spanish wsy of drawing legible serif type.

Esta by DSTYpe is meant to be distinctly Latin-American in flavour, for example. While, say, Quaestor Serif by the same is meant to evoke the style of Classic Roman letterforms.

So what might Germany's answer be to faces like Pradell, Andrade, Esta, and the like?

brampitoyo's picture

Ah, yes!

I completely forgot about the Catalan and Latin font scene. Perhaps due to my ignorance, I haven't seen much typefaces of these flavors other than Eduardo Manso's Relato and the many creations of DSType and Textaxis. Letras Latinas is one good resource, but they're so ecletic (which, in a sense, is good.)

But honestly, I've had a lot of trouble pointing out their oldstyle/transitional faces' unique characteristics.

So the categorization of French, Italian, Dutch and English (and Catalan) would make much more sense if seen in light of their historical progression, rather than their typeface constructions, then?

dan_reynolds's picture

There is a ton of German type. I hope that the Germans who read and post here will forgive me for leaving things out, especially all of the great blackletter type, which would be just too much to go into here!

A short, short, short list:
1. Garamond type came to Frankfurt with Jacob Sabon around 1580.
2. At first glance, Walbaum might look quite similar to Bodoni or Didot types, but it is really quite different, and (maybe became) quite German.
3. Lots of fantastic 19th Century sans serif type.
4. Walter Tracy brought a beat down on German serif typefaces of the first half of the 20th Century, writing (if I remember correctly) that Koch Antiqua was the only good one.
5. But there was a lot of great serif type… some of which, unlike Koch Antiqua, were real text faces (although Koch Antiqua in metal was used for books, too). Great types were cast from designs by Schneidler and Tiemann, for instance. I'm sure that there were many more.
7. Optima.
8. Despite my misgivings about some of it, the East German communist regime had a state-sponsored type foundry, which brought out some interesting work.

brampitoyo's picture

That's so true Dan. Faces like Palatino, Schneidler and von Hesse's Carmina looked distinctively German. Prillwitz, Koch Antiqua and Walbaum are good examples, but I'm focusing mainly on the Oldstyle to Transtitional genre.

What makes them distinctively German, though? That is what I was trying to figure out all along.

All I could think of is their Glyphic quality and sharp, tapered serif, but not much else.

dan_reynolds's picture

>What makes them distinctively German, though?

Their heavy use in Germany is one thing. Another specific "German" feature about a lot of late-19th century through 1940ish typefaces is (too) short descenders (someone out there please correct me if these too-vague details are inaccurate): The German type foundries had some sort of standardized sort measurements… these were based on the proportions of Fraktur. All fonts within a point size had to have the same baseline "height". Fraktur typically has a tall x-height, and small ascenders and descenders. Roman, or Antiqua typefaces don't tend to look good with the same big x-height… so that could be dropped down, and the design could have relatively longer ascenders. But the descender length couldn't change because the baseline was uniform. This allows for the easier mixing of types, I guess.

I think that German typefaces before Helvetica (begun in Switzerland but expanded by D. Stempel AG in Frankfurt) are almost all very calligraphic in origin, with exception of the sans serifs and some of the Modern (Didone) designs.

Palatine's picture

That's fascinating. Of course, I forgot about good old Palatino, still one of my favourites. Palatino Nova, of course, is causing some excitement as well.

brampitoyo's picture

That's some mighty specific and useful information you got there, Dan! I find that short descender to x-height/ascender proportion is, indeed, a highly appropriate way for classifying Teutonic typefaces.

dan_reynolds's picture

Bram, type founding was a highly industrial process for at least a good hundred years or so. The Industrial Revolution brought standardization to all sorts of things, even places where it did not necessarily need to be.

This is interesting in light of the fact that no German government ever really declared Fraktur to be its official type style of choice (a vote in the pre-WWI Reichstag was more or less deadlocked between Fraktur and Antiqua), although the NS-Regime did ban it towards their end. I guess that standards are another matter altogether.

William Berkson's picture

The short ascenders are not only German. The 'Dutch' faces had shorter ascenders and descenders, also probably under the influence of black letter.

The shortened descenders was the fashion also in the US from the late 19th through mid century at least. Some thought very short descenders were important to economical setting. Merganthaler Linotype faces were also cast on a 'standard' baseline, which enforced short descenders. They had to make special provision for 'art' faces with full descenders.

dan_reynolds's picture

Many of Goudy's faces have quite short descenders, don't they? Didn't Monotype matrices come in short and long alternates for some designs?

Speaking of Economy, Dr. Karl Klingspor wrote a book around 1950 chronicling the Klingspor foundry… in one sense it seems like a swan-song for a company whose days were numbered. Anyway, he shows several pages of examples illustrating how Fraktur is more economical than Antiqua (more characters per line, more lines per page within the same allotted space). In these days after WWII, I suspect that it was in the offing that books were not going to be set in Fraktur anymore. But the book is still an interesting read.

brampitoyo's picture

The industrial revolution and the mechanization of the press are very good points on the standardization of types. And yes, I remembered from reading somewhere that Monotype and Linotype has a rough classification system comprised of text faces and art faces. The former complied with their matrix standard while the second can really go out and about.

The Legibility Group's newspaper face was very easily identifiable, though. Noticeably short descender and ascender, normal to condensed width, noticeable open counterspace, Transitional-Modern roman construction, and so on.

The Dutch typefaces are just like that, but with an Oldstyle construction in mind. Indeed, 20th century romans like Times New Roman was based on Plantin.

Teutonic faces, on the other hand, seem to have a moderate x-height (caused by short descender but normal ascender), an Oldstyle-Transitional construction and that glyphic quality (see the hook of 'a' on Palatino)

ben_archer's picture

Hi Bram

Well yes we know what you mean... the problem originates with the Vox terminology because it equates oldstyle (your term) with 'Garalde'. 'Garalde' is a compound name that conflates two different categories - the oldstyles of GARamond and the late venetian styles of ALDus.

William is right about the essential progression of the differing national styles historically, but by the time he mentions Baskerville (transitional) and Dan mentions Walbaum (modern) we are no longer in oldstyle territory, and outside of the scope of your question.

Interestingly enough, Jaspert (1958) lists only 30 faces in the Old Face section of the Encyclopedia of Typefaces – so in a sense, this is already the simplest way to organise the oldstyle faces. A key attribute would be the irregularity* of genuine oldstyle faces when contrasted with the technical refinements that produced later types.

The situation only gets really complicated when trying to include all those 20th century revivals and text faces within an oldstyle classification – which is where Robert Bringhurst's ideas about a historically-based classification come in useful.

Related articles about the relative uselessness of classification in general are Jonathan Hoefler's 'On Classifying Type' which is mentioned at http://typophile.com/node/17412 and also a great discussion on this from last year at http://typophile.com/node/9757

Historical and structural distinctions between French, Italian, English and Dutch oldstyle typefaces are reduced to English/Dutch vs. Continental Taste by Catherine Dixon in her 'Typeform Dialogues' project, which is an attempt to work up a new classification system and iron out the vagaries of the old systems. One of her significant findings was that type designers consciously 'aim for the gaps' in classification systems – if you think about Unger Fraktur (1793) being a hybrid between forms of traditional blackletter and the proportions of the (then) modern roman of Didot, then you can see that this idea has a lot of precedent. I guess it was like the Dead History of its day.

Although 'Typeform Dialogues' is currently on hold, Dixon's system is written up pretty comprehensively in Baines & Haslam's 'Type and Typography'.

* This is a real problem if you think about where/how to classify Beowolf because while it's based on an oldstyle face, the technology that creates Beowolf, and its resulting irregularity of forms are really not 'genuine' oldstyle.

brampitoyo's picture


That's why I put the word "oldstyle" instead of "Garalde", "Aldine" or anything. Speaking of which, Aldus' Venetian is earlier than Garamond's style, too.

My aim is to figure out whether classifying oldstyle faces (and just oldstyle faces) by region is relevant and logical. To my knowledge most resources only made distinction between Venetian, French (as in Garamond's), Dutch and English, but only vaguely mention the unique features of each.

Put simply, I want to find some identifying characteristics that, although cannot be applied universally, can be used to categorize Fournier as being French, Caslon as being English, and Plantin as being Dutch.

Bringhurst's historically based clasification can be one of the characteristics, because it doesn't discount newer faces like Palatino, Perpetua, and such.

I suspect that another good characteristic is the x-height to ascender to descender proportion. This method, in my opinion, is more accurate than simply saying "This typeface has a large x-height and long ascender" because it takes the three proportions into account. As you can see, I attempted to differentiate Dutch oldstyle from German oldstyle with this method. I'm not sure if it hits the jackpot, but it's a good place to start.

Irregularity can be another good point, but I would see it in a different light: the italic. As we all know, italics was originally not designed to work as roman companions. Dutch and English oldstyle faces have italics that are 'regularized' (read: designed with the roman in mind), while Venetian and French roman and italic were two separate entities.

But this idea, born solely out of curiosity, could just be another useless attempt at catching the wind. So I really don't know.

I have yet to read all of your recommended articles -- which are all very relevant to the topic -- aside from Mr. Bringhurst's.

But I have a hunch that I should start with Mr. Hoefler's posting first.

Keep the comments coming!

ben_archer's picture

Happy reading, Bram

You're not the only American I know who's commented on an unwillingness to embrace the Vox terminology... I think Forrest Norvell did the same last year.

Somewhere – maybe in Bringhurst, or more likely in Lawson's 'Anatomy of a Typeface' – I recall seeing a comparison chart of slope degree and/or irregularity of italics within a range of faces; has anyone else seen this lately?

As for Dutch and English oldstyle faces have italics that are ‘regularized’ I have to disagree.

One of the attributes of the Caslon Old Face that always made me reluctant to use it was (and still is) the unharmonised slopes of the italic. Jaspert states In the italic, as in many old face italics, the capitals are irregular in their inclination; the A, V and W appear to be falling over.

When you examine the italic lowercase close up, it's apparent that the slope of the main strokes of b,d,f,h,j,k,l,p and q are in no way consistent with each other, or necessarily with the rest of the mean-height lowercase characters.

The model for Caslon's work, especially in the italic, was type by Cristoffel van Dijck, Amsterdams leading typefounder of the preceding century. This irregularity of slope is of course related closely to the change of the stress on different letters, reflecting writing by pen which changed the angle of the pen that William is talking about at post #3. These are not rationalised italics but letters based on the inconsistencies of handwritten forms.

Not all handwriting needs to look unbalanced. My reasons for discriminating against Caslon are the same as my reason for disliking Rotis – there is an intrinsic lack of balance in the face, and certain characters look like they're about to fall over. Once you notice this, it's very hard not to see the same flaw ever after on each successive viewing.

To argue for a regional or national variation in rationalising italic slope angles relatively within the italic (or to the construction of the associated roman face) – you would have to go to a later historical period, and here the French founders of the early 18th century have the advantage, but by this point we are dealing with transitional typefaces, rather than oldstyle.

William Berkson's picture

>These are not rationalised italics but letters based on the inconsistencies of handwritten forms.

The irregularities in the slopes of old style italics generally do have a logic to them, and it is not really a handwriting logic. For example generally the ascenders are more vertical than the stem of the n. I suspect this is because a more extreme slope would create kerning problems. The lesser slope of k ascender compared to the l, for example, is because of problems in drawing a balanced looking old style k (looped top) with an extreme lean in the ascender.

In the case of Caslon in particular, the AVW in the original Caslon have their thick diagonals nearly the same slope as the H. This makes the thin diagonals slope more extremely. You can argue with it as a bad decision, but it does has some visual logic to it. I believe that Moxon, the early English type designer, recommended it. It works to an extent with the caps as initial caps with lower case letters. In an all-caps italic setting it clearly doesn't work very well. But then I don't think they were planning to use italic all caps back then. The frequent presence of the swash JQTY as standard is a sign of the same approach to italic caps as leading letters.

ben_archer's picture

Ah yes, I'd forgotten about the angles relating to the kerning... however a comparison of the different angles in mean-height lowercase n, u and v, and also the main strokes of p and q, would still support what I'm saying about the slope inconsistencies in general.

I think my point was to claim that particularly the Dutch and English oldstyle italics were not harmonised (or regularised, as Bram says).

And yes, in the case of Caslon I can argue with it as a bad decision and think other faces are superior, but then we have a lot more to choose from these days. Later (better?) reworkings of Caslon, like Imprint show a much more orderly conception of what a regularised accompanying italic should look like.

brampitoyo's picture

I think I failed to define the word 'regularized'. By that, I meant that the Italics were designed to go with the Romans, regardless of their slope angles and such.

But indeed, orderly slopes was more prevalent the more recent you go in years. In that respect, English face like Caslon had a more 'regularized' italic in comparison to Janson.

The fact that the digital rendition of Garamond from Monotype has such an individual italic is an entirely different thing.

Syndicate content Syndicate content