In case you haven't read it yet, this might be helpful: Serial Type Families by Alejandro Lo Celso on TypeCulture.

There is no relationship other than their trade names. Expanded type families began in the 19th century. Early fonts such as Garamond did not have a bold face version until the 20th century.

Note that all the versions of Splendid have a similar x-height, and that between the script and antiqua versions, the thick and thin strokes are of the same weight (roughly). The idea is that if you need something done in a script font, you have one that was designed to match perfectly (e.g., you can get rid of the fora where people ask "What sans font goes well with [insert serif font]?"). The changes might be small or large. In my font, I have a 5 faces so far, and as you can see they all share certain qualities. The same x-, ascender, and descender heights, same colour (at least that's the goal), etc.

A (far better) example might be the Daam "Entity" by hrant, each of the faces DominationAvailable, Cristaal, and Brutaal, each go perfectly together yet have serve very different purposes. http://themicrofoundry.com/s_latin.html

«El futuro es una línea tan fina que apenas nos damos cuenta de pintarla nosotros mismos». (La Luz Oscura, por Javier Guerrero)

The driving concept behind a series can create a family. House Industries is great at this; Chalet, Neutraface, Luxury, and more of their families contain fonts that are related more by the ideas behind the faces than by design of the fonts.

On a related note, I wish I could be related to Greta.

It would be interesting to see how an abstract concept (possibly with soma shared formal attributes) can be adapted to a type family.

As much as I like the Luxury package; it is hard to see it as a type family, but rather variations of the same theme. Many foundries bundle fonts like this, usually grouped by purpose (Web, Script, Newspaper, Funeral, or whatever pack), but this is an arbitrary grouping driven by marketing, and the fonts were not designed from the first to be used one next to another.

What use of your (guifa) type family do you imagine? And Hrant how do you see your Entity project used? Have you seen any meaningful examples of your typeface in use?

> Have you seen any meaningful examples of your typeface in use?

What's a meaningful example?

?

!Lestatic and !Square Engine are thematic decorative type families. The common emphasis in !Lestatic is glyph "skeletons" and the metrics match within one width group. !Square Engine letters share a common naming convention, letterspacing and interchangeable components of "character", apart from the glyphs' shapes.

I'm presently planning on setting a version of a close friend's dissertation (she's getting her PhD on Saturday yay!), and then getting one of the book arts MFA students at my undergrad alma mater to bind it once printed. Unfortunately, until I'm done with my thesis I can't put in a lot of hardcore work on the font, and only get a character or two done every here and there as of late. But, the sample layouts I've done so far have the serif version for the body text, and of course italic as necessary. For captions and footnotes, the sans version is being used. This isn't too unlike a lot of journal style guidelines, where all figures and their captions must use sans-serif fonts. I doubt I'll use the fraktur variant for the book though.

One practical use for a font like hrant's (He normally posts rabidly on here, but on another thread someone said he's on vacation or something like that) would be for setting Faulkner's The Sound and the Fury. If you've read it, you'll rememmber that flashbacks/forwards were designated only by italicising the text. This was not the original way that Faulkner intended, rather, he wanted the different time switches to be represented by type setting that entire passage in a different colour. But, with many pages needing then three+ colours, for novel, it was considered too pricey at the time. (The only novel I can think of off the top of my head with a multi-colour printing in that style is House of Leaves, where the most commonly available edition prints the word "house" in blue, all certain passages in red. The "full version" also has parts in purple and (!) braille.) However, one could design a series of faces in the vein of Daam and have a different way of signally the transitions betwixt times, basing them off of the different feelings the narrators had on them.

«El futuro es una línea tan fina que apenas nos damos cuenta de pintarla nosotros mismos». (La Luz Oscura, por Javier Guerrero)

> He normally posts rabidly on here, but on another thread someone said he’s on vacation or something like that

Ahh, I thought perhaps they had him in for rabies treatment for posting so rabidly.

Underware's Auto and its three italics could be an example of a "character" or "funkiness" axis. They even advocate their being used within one design.

Although it's a bit more conventional in being related by weight, Porchez's Anisette and Anisette Petite could be a slightly unconventional example, what with one set being all caps of different widths and the other being single width caps and lowercase. But maybe they should be seen as simply a fragmented type family.

Oh, also take a look at H&FJ's The Proteus Project and Jeremy Tankard's The Shire Types. Both families are designed so their members are easily mixable.

Peter, when you write of Storm's trivial Splendida family

>I thought of something like Storm’s Splendid, but there I cannot trace what is it that links these four separate fonts into one family.<

aren't you in essence saying that you have discovered you might be asking the wrong question?

There is no relationship there; it is a frivolous, poorly-conceived gimmick == that I hope for Storm's sake sells well. In fact, let's for once be brutally honest. There is a relationship: it's called type fatigue. He's done too much.

What would be more interesting to explore is the question whether formal relation (between for example sans and serif) is desirable or not. This is far from established. There is a lot of hype based on the aesthetically unproven assumption that a formally related sans and serif are desirable. [Really? Contrast does not exist as a beneficial force in art since cave drawings?] Clearly, a lot of this is hype and simply not to be taken seriously in a serious study. It is as bad as the assumption (which I have never found to be the case) that there exists, anywhere, a designer who is equally good at designing serif and sans serif type. I am speaking of course only about the very highest level of endeavour.

Every effort to form an italic from roman models has proved disastrous in the end. The best illustration is the very first integrated type family, Romulus. The ideal italic is not formally related to a roman; it is above all a contrasting element. Why should this not be true also of the sans serif that might be used in contiguous composition? That leaves the bold - - - the late 19th century bolds were not slavish replicas of romans but simply darker types that happened to work acceptably and agreeably with romans. I see no reason why this should not work today. I would even go so far as to say that it is more important to have a well-designed, well-implemented bold than to have a bold that is closely, literally formally related to the roman.

There is an impulse in these kinds of discussion that I deeply distrust. That impulse is the one that seeks to find a basis for type in some art or graphic theory. For me, it must always start at the functional level of the reader and the act of reading.

As everyone who has ever made a sans from a roman knows, you can't stick to the formal recipe blindly. That's because, as you see at once,

_a Sans Serif has a vernacular of its own_

and that is different from the vernacular of a serif typeface

just as the vernacular of italic is different from the vernacular of roman

Trying to impose unity on a family of types is to deny and disrespect that each constituent of a type family may have a vernacular and vernacular associations which should be honoured.

At any rate, that's a possible argument.

Martin Majoor has some theory about family-relations.
Ask Bill

formal parameters

These may be considered "design axes", an idea introduced to explain Multiple Masters.
This kind of interpolation derives, conceptually, from the Knuth metafont, and is "codified" in the Univers numbering system.
Carl Crossgrove's Relic, with its sober/inebriated axis, is an unusually conceived formal dimension, but really there is so much scope for imagination -- you can morph between all kinds of formal qualities with software like Superpolator: x-height, extender length, etc.

"Slant" is a paradoxical design axis, because in modern fonts it can indeed be varied incremementally in a purely formal slider-bar way, but in old styles it exists as an abrupt binary switch between quite topologically disparate roman/italic forms, or perhaps a three-parter (Auto, etc.)
I have termed these Cultural Axes, as the perception of a meaningfgul relationship between the poles/instances depends on a cultural awareness or knowledge.
Both formal and cultural design axes can be the basis of extended families, although in large OpenType fonts such as Hypatia, these may be subsumed within a single font as Stylistic Alternates.

The Globe and Mail fonts I recently designed leverage a cultural axis. GM News and GM Sans share x-height, extender length, stroke widths, some cheirographically informed joints, and general horizontal proportions, but the two alternates are conceived as the poles of a Humanist/Modernist axis. GM News is "some serifs" and old-style, whereas GM Sans is plain and modern. In the newspaper, the two variants are used in a variety of ways, e.g.
- For text emphasis, highlighting a word in a headline;
- For immediate layout contrast, as when a GM Sans overhead sits above a GM News headline; the most interesting of these is the contrast between a GM News italic (6 degree slant, classic old-style letter shapes) and GM Sans (10 degrees, a slanted version of the Roman);
- At the document level: GM News is the headline face in the news sections, GM Sans in the feature sections.

http://www.shinntype.com/Specimens/G&M.pdf for the type specimen,
http://www.newseum.org/media/dfp/pdf7/CAN_TGAM.pdf for the usage (most of the sans fonts on the front page are GM News, but the overheads and sidebar text are GM Sans).
These stylistic variants provide an extended design space for modernist page layouts, enabling many ways to contrast text while maintaining color and stroke contrast -- this equivalence being important to maintain the effect of locating the type on the plane of the page as flat graphic elements, without suggesting pictorial depth, which is what happens with stroke contrast.

**

There are cultural axes in several of my type families, Panoptica being the most diverse to the extent of becoming an alphabet system, rather than a typeface.
Unusual design axes in some other of my work:
Cultural: Eunoia; trendyness -- 1) Art Deco (1930s), 2) Op Art (1960s), 3) Techno (1990s)
Formal: Monkey; Spacing -- 1) visually pleasing spacing, kerned, 2) tight-but-not touching equidistant spacing. 3) monowidth.
Formal: Handsome; writing tool -- "a choice of pens".
Formal: Fontesque Ornaments -- the family principle applied to ornament, with optical scaling, monoline/stressed members, and plain/solid members.
Formal: Brown Gothic -- text and display variants of a sans, optically scaling the "distress" axis.

>This kind of interpolation derives, conceptually, from the Knuth metafont, and is “codified” in the Univers numbering system.<

Nick, do you ever think about what you write?

The Univers numbering system, apparently conceptualized by Frutiger, was originated in the 1950s. Knuth got interested in type 20 years later. So even if this statement were true, it would be the reverse of true. And couldn't you learn more about metafont before you start posting authoritatively about it?

Mac McGrew has argued to me that the first, primitive interpolation was done by Benton with his adjustable parameters for the pantograph.

The rest of this weird post -- 'without suggesting pictorial depth, which is what happens with stroke contrast' (somehow Bodoni's pages are 'pictorially deep' whilst Calson's are not?) and the idea that the term axis should be broadened to encompass characteristics like 'trendy'(you hope!) -- I find a little too light.

Yes Bill, my post does inadvertently give the impression that the Metafont came before Univers, but do you really think I didn't know the right sequence, or do you just like to put people down so that you can show off how smart you are?

'...without suggesting pictorial depth, which is what happens with stroke contrast’ (somehow Bodoni’s pages are ’pictorially deep’ whilst Calson’s are not?)

Caslon's types are not without stroke contrast. The distinction I'm making is between serifed faces with stroke contrast, and sans serif faces that are nominally monolinear. Modernist page layouts are typically the Swiss or International Style of layout, not something involving Didone faces.
The challenge in the Globe was to handle a multi-element hierarchy in a modernist layout; usually in newspapers, that complexity is organized around contrasting sans types with serif types, but serif types undermine the flat graphic quality of the modernist layout. The solution was to use all sans display type, with contrast between stylistic variants of the same sans face -- "some serifs" (humanist) and plain (geometric).
As you say, "it must always start at the functional level of the reader", and that provided the impetus and the context for those stylistic variants.

I find a little too light.

If not a "cultural" design axis, how then would you describe, with suitable gravitas, non-interpolatable relationships between typeface family members? After all, that is the theme of this thread -- Peter wants to know about "some other attributes". It's a little too easy to dismiss them and say that roman-italic and regular-bold contrasts are enough.
Chalet, for instance, with its New York-Paris-Tokyo styles.
Auto, with its "unexplored palette" of three different italics.
Eunoia, with its Deco/Op Art/Techno era-related versions of a condensed, high-contrast sans.
Lisboa, with its plain and neo-humanist variants.

How are these different than the "Serial" families mentioned in Alejandro Lo Celso's Type Culture essay, such as Thesis?

>Yes Bill, my post does inadvertently give the impression that the Metafont came before Univers, but do you really think I didn’t know that, or do you just like to put people down so that you can show off how smart you are?<

Nick, I don't think I am smart, but I do have a reasonable grasp of basic typographical facts, and I do think before I write.

Of course I thought you 'didn't know that'. It's the only conclusion one can come to from what you have written. If you don't want people to think you are an idiot, think before you write! You're taking up precious minutes in the day of every one of the hundreds who read your posts. Don't you think it's perhaps worth a few minutes of reflection to make sure you don't write the opposite of what you mean? That what you _do_ write is actually worth reading?

>how then would you describe, with suitable gravitas, non-interpolatable relationships between typeface family members?<

I wouldn't. None of your exemplars merit gravitas-containing bandwidth, which is always in short supply. All you are doing here is devaluing the meaning of gravitas itself. You are using it like an ad-man. Your motives are transparently self-serving: self advertisement. A blog is not the best place to do that kind of thing so blatantly, because you are bound to get slagged. Save it for your client presentations.

The only way you can get away with this kind of thing on a blog is to say, 'Folks, here's a script I'm going to try on a really naive client. How do you think it'll run? What can I do to improve it?'

You don't have to worry about the client seeing the post. The client who buys that kind of hype probably doesn't read type blogs.

In the meantime has anyone got anything serious to say about Peter's original question? It seems he is not going to have an easy time writing this article. It looks like the one historically important designer who tried to conceptualize a wide-ranging type family was van Krimpen, and further, that it was most notably a failure in several respects. Of the two italics, one couldn't be set, and the other couldn't be stood. The sans experiments weren't successful either. It strikes me as significant that this is a subject very little discussed by the older generation of type designers -- those who really know how to design type -- such as Stone, Carter, Unger, Zapf, Frutiger, etc. It's also significant that there is no attempt in Font Bureau's ambitious Poynter family to embrace this kind of light aesthetospeak. I am almost ready to dismiss the talk of relationships as pure salesmanship.

Oh, for a milligram of gravitas!

While we're setting things straight, Martin Majoor made some factual errors on Peter's blog some time ago that have never been corrected. Scala sans does not make so many firsts as Martin claims, and surely he knows about the celebrated Today Sans?

The first sans serif superfamily was Volker Küster's Today Sans Serif, 1987. Two optical axes, extra light to extra bold, lining and osf figures, small caps, all f-ligs, true italic, ad infinitum, ad gloriam. An astonishing achievement for its day, and, still, today.

Actually it is wrong to speak of axes here, and earlier in Nick's post. Axis should only be used where interpolability is possible, for obvious reasons (i.e, the word only exists in type to describe practical, interpolable extremes; style is probably usually the best word to describe something that is related but not interpolable). Some use was made of interpolation in Today, but for example the extra bold weights could not be designed to be interpolable, and the optical axes are not practically interpolable either.

Drawings exist, in some stage of completion, for condensed and expanded versions of Today, but Küster seems to have lost interest in typeface design. It is one of the great tragedies. Why? Why? Why?

Your motives are transparently self-serving: self advertisement. A blog is not the best place to do that kind of thing so blatantly, because you are bound to get slagged.

Bill, you can misrepresent me and threaten me with further abuse, but I'll continue to discuss the topic of this or any other thread with reference to my own work.

**

The term "cultural axis" is a serious answer to Peter's question about the Splendid family -- as well as sharing common physical qualities such as x-height, sharpness, and high contrast, they are joined by a "cultural axis": a place, a time. Another cultural axis is that they are being marketed as a package -- present day technological and economic considerations are involved in that (sorry, bit of a circular argument). Implicit too is the understanding that these kinds of faces are often used together, again, a cultural understanding of genre. But was the Splendid concept successful? Have art directors produced layouts with three or four of these fonts together on the same page?

The issue here: although non-interpolatable styles may be part of the same family, is it more a conceptual or taxonomic kinship than a functional one? To what extent do typographers create layouts using both the sans and serif versions of a typeface? How much contrast is enough? As an art director, the only time I've used sans/serif contrast within a family, sort of, was combining with Perpetua and Gill Sans.

In her talk on Morris Benton at TypeCon, Juliet Shen noted that his Clearface types -- the serif and the sans -- were closely related physically, and placed a few pages apart in the same section of the 1923 ATF catalog. So the relationship ATF presented was a bit vague and unrealized, in terms of physical kinship, marketability, and usability (they were not shown in the same faux layout).

...the word [axis] only exists in type to describe practical, interpolable extremes;

An axis is a connecting line, that's all.
Given the overlap between interpolatability and non-interpolatability in type (Bill mentions Today's weighting), it seems reasonable to extend the use of "axis" to non-interpolatable relationships.

The crux of this must be the overlap of italicization and slanting.
An axis can connect as little as two non-interpolatable polarities, or have an infinite number of potential instances.

>An axis is a connecting line, that’s all.
Given the overlap between interpolatability and non-interpolatability in type (Bill mentions Today’s weighting), it seems reasonable to extend the use of “axis” to non-interpolatable relationships.<

What's wrong with the previously-used word, style?

Nick, you say axis is a connecting line, that's all.

But that's not what it means in type. It has a special technical meaning, specifically given to it probably by Adobe, or if not Adobe, one of the early workers in interpolation, such as Karow or Harper. It would be interesting to discover exactly when the term was first used, how it was defined -- it probably was defined, etc.

Since axis has always been used thus in type -- why add to semantic confusion and change it now?

Also, is axis really flexible enough for what you want to talk about? Why confine yourself to a line? I really think there must be a better word. What is it that you want to describe? It seems to me that you want to describe the relationship between two typefaces that do not necessarily have a linear or even predictable relationship. So why use a word steeped (or mired) in linearity?

Why make a word that means line not be a word that means line? (For as long as there's a line, interpolability is implied, is it not?)

You're not talking linearity here, but something more like spousal, neighbour, orbital (?) (rather than axial!) -- I can't think but there must be something better.

Looking back on this, I can see I am already confused by the semantics and not making any sense. So going back to the OED, in which most of the many definitions seem to depend on linearity, one nevertheless sees

b. fig. The ‘pivot’ on which any matter turns.
1860 MOTLEY Netherl. (1868) I. v. 169 The axis of the revolt was the religious question.

which is a great, flexible definition though only supported by one quotation.

Let me ask a simpler question: in type, isn't axis always properly used of a practical multiple master?

>Küster seems to have lost interest in typeface design

I'm not sure this is completely accurate. Maybe he is just too busy teaching? He is a professor at the University of Essen–Duisberg in northwestern Germany, where he teaches Typeface Design, Typography, and Book Design.
http://www.uni-duisburg-essen.de/fet/fue/eng/fb04/fb_04_kdg_06.htm

Perhaps Karsten Lücke would know more. Karsten was one of his students, and is a practicing typeface designer himself.

What’s wrong with the previously-used word, style?

There's not much wrong with it, and it is used in OpenType -- Stylistic Sets.
But "style" has a lot of meanings, and is thus vague.
The advantage of using the term "design axis" is that its focus serves to distinguish interpolable relationships from non-interpolable, which is the whole point of this exercise.
As you have noted with Today, weight is a design axis that may often be non-interpolable. In this respect, Times Roman is similar, and many Didone Italics have median serifs that are hooked in the regular weight, but flat in the bold weight.

Working on the modern/oldstyle versions of the Globe sans types, I found that what I was doing, for the purpose of adequate differentiation, was trying to drive the styles as far apart as possible, to the poles of their relationship, as it were -- so this too suggested that the relationship was axial.
And within this major modern/oldstyle axis, I found there were sub-axes; for instance horizontal proportion -- varied in the oldstyle, more equal in the modern; terminal treatment -- angled and slightly bent in the oldstyle, straight and perpendicualr in the modern; joint style -- pronounced in the oldstyle, neat in the modern. In each of these "parts of speech" of the design grammar, I had the option of varying the amount of difference between the two polarities.

Here's something from 1938, "The following Alternate Characters are put up in separate fonts at quoted prices and when used with Series No. 140, shown on preceding page, gives a very pleasing effect."

>Who said that until the 1990’s there were no italic small-caps?

Someone who didn't know what they were talking about.

Nick, the illustration is really interesting, can you tell us more about it, the foundry, the seller, the book? Is this an old Sol Hess design? I hate to say it, but I see here a possible genesis for the alternate characters in what is surely one of the most original of all typefaces, Renner's Steile Futura. How could anything in Steile come from such an undistinguished source? Yet it seems inferentially to be the case. Oh, this is awful! Similarly, some of the most admirable of Zapf's and Frutiger's ideas can be found in Goudy, in pre-war German types, etc. etc.

Nevertheless, I wouldn't call them an axis or a design axis. What I resent about broadening the term is that it was invented by Adobe (if it really was invented by Adobe -- I would want to check this) to mean a very specific thing. I think we should respect that specific thing, and come up with a different term if style is inadequate. I can see why you are unhappy with style, or with some variant thereof, such as 'integrated style'; I can see the need for an extension of the terminology -- not, necessarily, as far as the type world goes, but as far as the type to management world goes. Isn't there some better way of talking about this?

That said, what is interpolable? The very first Adobe MM with an intermediate master was ITC Garamond MM, which had one character which was not interpolable across the 'design axis' (is that a correct usage according to my admittedly inept but well-meant plea?) which required the intermediate master.

A definition of a font axis which meant only an 'Adobe MM interpolable axis' would therefore still not be completely simple. It wouldn't be linear -- although non-linear interpolation may not have been considered practical at the time these terms were being dreamed up. But, as we know, the Adobe model is incredibly simplistic; we have gone far beyond that conceptually. So what are axes?

I would vote that an axis should be something that is either literally linear or very near linear. Some other term is necessary. Is 'design space' the term we use for the totality of the design? The sum of the axes?

Here's where style should not be extended, obviously as I had naively originally proposed. For example, we can speak of the design space, the axes, of Minion Roman. The design space, the axes, of Minion Italic are something quite different.

So, for elements within a design space that are not axes -- what are they? That's supposing that you are going to limit 'design space' to the classic Adobespeak.

Bill, I appreciate your insistence on accuracy but at the same time I think you made your comments in a way that was more personal than was really needed to get your points across. I can't say that in a fit of pique I haven't done the same...

>I searched for the feature I read about and I found it. Strangely enough, Peter Bilak interviewed Martin Majoor (feb. 2003) and you can read it here: http://www.typotheque.com/articles/interview_martin_majoor.html. The part I’m reffering to is: Small capitals come from an old tradition, but designing something like Scala Sans Italic Caps was not done before.<

This transparently self-serving article by Majoor is factually incorrect. But let's take just the points you apparently took from it. First of all you wrote that italic small caps did not exist before the 1990s. This is completely untrue. They have a long history.

Then you wrote that sans serif small caps were not available before Scala. This is equally untrue. The first sans serif superfamily, explicitly containing every possible feature, was Today, as Majoor perfectly well knows. The other thing to keep in mind of course is that in metal you don't need an explicit small caps weight because you can usually use the caps of a couple of sizes smaller, and because of the optical reproportioning, they will have the correct appearance.

Speaking of optical reproportioning, Today has two optical sizes, Scala has only one, so it is still behind the curve.

Now, further to test some of Majoor's claims about Scala, do yourself a favour. Set some paragraphs in Scala, and make sure you have plenty of instances of words set between the right and left single quotation marks. Things like 'this', etc. -- the kind of thing you often find in the kind of museum catalogues that Scala is often used for, in fact. Find out what is wrong with these sentences, and find out why it is wrong, which will require a little detective work. Hmmmm. A lot of thought went into making that mistake -- which I think may be unique in type history.

'It is my conviction that you cannot be a good type designer if you are not a book typographer. I am not talking here about display types but about text types. A type designer must know how type works in a piece of text, he must know what happens with the type on different sorts of paper, he must know how a typeface behaves with different printing techniques.'

You can say that again, Martin!

Eben, you are right, I was too personal. But Nick is always so careless, this has been going on for years. What if people actually took this kind of thing seriously, some future researcher?

Anyway, I think I have more than met my match. Can anything exceed the arrogance, venom, mendacity and pomp of the Majoor article that what's-his-name cited? Like basically all it's saying is, 'everything else everyone ever did is wrong; I was the first to do it, and when I wasn't, I was the first to do it right'.

OK, but what is the original idea?

>Why are typefaces related?
Maybe because the designer says so!<

I agree with this, because it implies that relationship does not have to be measurable or obvious.

When you say, what is the relationship between the humanist miniscule and the humanist italic, I take that what you're saying is, there isn't any. Indeed there isn't. After a lifetime of designing chancery italics to fit his romans, van Krimpen admitted it was a mistake.

In Scala, the relationships between sans and serif and weights are obvious, but where it falls apart completely is in the relationship between roman and italic, which is poor. Although it is one of the most attractive typefaces of the 1990s, one must admit that the italic is slanted too much and is neither particularly readable nor particularly harmonious with the roman. Indeed it almost seems arbitrary. I have always thought the chancery-ish italic of a contemporary typeface, Minion, was inept.

Yet, as we know, highly related italics (i.e. slanted romans) don't work. So the talk of relationship is rather bogus.

So we come back to:

>Why are typefaces related?
Maybe because the designer says so!<

And all the rest is just hot air.

But was this what you meant?

yes

but I guess that doesn't help Peter with his article.

OK, but what is the original idea?

>Why are typefaces related?
Maybe because the designer says so!<

I agree with this, because it implies that relationship does not have to be measurable or obvious.

When you say, what is the relationship between the humanist miniscule and the humanist italic, I take that what you're saying is, there isn't any. Indeed there isn't. After a lifetime of designing chancery italics to fit his romans, van Krimpen admitted it was a mistake.

In Scala, the relationships between sans and serif and weights are obvious, but where it falls apart completely is in the relationship between roman and italic, which is poor. Although it is one of the most attractive typefaces of the 1990s, one must admit that the italic is slanted too much and is neither particularly readable nor particularly harmonious with the roman. Indeed it almost seems arbitrary. I have always thought the chancery-ish italic of a contemporary typeface, Minion, was inept.

Yet, as we know, highly related italics (i.e. slanted romans) don't work. So the talk of relationship is rather bogus.

So we come back to:

>Why are typefaces related?
Maybe because the designer says so!<

And all the rest is just hot air.

But was this what you meant?

Yet, as we know

I am inclined to favor a real italic too but there are better & worse sloped romans and I would choose a great sloped roman over a poor italic any day.

So what are the better and worse sloped romans?

What is the ideal italic and for what purpose?

Let's not consider the case of the beautiful italic in which long passages, usually of poetry, are set, but rather the case of the italic used for the very occasional word of emphasis.

I take as a test an Iris Murdoch novel. She rather overuses her italics, but they occur at moments of great stress in high pitch reasonably contemporary situations. Baskerville is not the kind of thing you want here; neither is Garamond or Scotch. They're all too florid. Instead of seeming intense, the emphasized word seems silly -- in Murdoch. (of all the most conventional choices, pre-Porchez Sabon is the best)

The kind of inconspicuous not too cursive, not too humanist, not too anything italic that Carter designed for Charter seems to me a good model here.

I also must say that if italic is only being used very occasionally, I can't really object to sloped romans in sans serifs, but I have never been comfortable with the sloped serif italics that I know. Stone Cycles is another case where it just works for me. If it's a case of setting a book, then I'm likely going to go for a Stone or Matthew Carter face, regardless of other considerations, because I know the fonts will actually work..

So what are the better and worse sloped romans?

I am not prepared to go toe to toe on that one. I am just interested in the point that absolutes rarely exist in type. Instead fonts/Type are a deeply contextual thing - as your very good point about Iris Murdoch points out quite well.

One reason I am developing new respect for the sloped roman is some of the examples Hrant showed me when I visited him. They were great. I was solidly against them before that. Imagine what you would have said about Sans Type at the beginning of the 20th century vs now. A great deal has been done and you can't esimate the poential ogf the sans the same way any more. I think sloped romans may 'grow' in the same way.

Hrant of course likes them better than Italics. I haven't gone that far yet myself.

A side Q: What Iris book do you like the best?

"...but there I cannot trace what is it that links these four separate fonts into one family"

Peter,
Other than the bold and regular didone pair, neither do I see the link. I am quite intrigued by the concept you stated though. Typically, families are weight, slant, and width variants with common features. The italic is less family like out of the need to be distinguished in text. That is all simple stuff and done many times. Even making a serif and sans family fulfills some corporate need for commonality of brand. Again, there are plenty of these in recent years. The intriguing thing to me is your hint of purpose in mood or tone set by the family variants. I have not seen this exactly yet. The closest thing might be Thomas's recent Hypatia but this is stylistic sets within a single font weight rather than different fonts in a family.
I have done a pair of display fonts with different tone called "Boulder id" and "Boulder ego" with alter ego to follow. There is an attempt at shift in tone but I doubt that this is what you are talking about.

ChrisL

Here is sample of “Boulder id” and “Boulder ego”
The ego version is more tame and reads better at smaller sizes than id.

ChrisL

I remember reading recently that most of Frutiger's typefaces are loosely related structurally, obvious examples excluded. There is the obvious example of Egyptienne F and Univers, but also his other text faces have very similar skeletal structures.

And Nick noted Joanna and Gill Sans as being related above.

Had any or all of these typefaces been designed and created in the 1990s/2000s they would have been marketed as “families”, although there are few things that formally relate them to each other in ways that families are related today.

Had any or all of these typefaces been

This seems like little bit of a reach to state this so clearly.. but it's an interesting idea.

Nick, your tourist gothic example is super creepy but also very interesting. Thanks!.

>I think sloped romans may ’grow’ in the same way.

Most important point. Taste and perception are never stationary; everyone changes over time. What seemed impossibly crude 20 yrs ago might seem merely functional today. When I first talked to Lino in 1995 about reviving Pilgrim, Bruno didn't want the rather nice modified slanted roman, but a completely new, cursive italic, as was the hot fashion of the minute. By 2005, when I was beginning to get serious about it again, the modified slanted roman looked really cool to Bruno and anything else would be ridiculous.

One man, evolution of taste over a short period. But there are lots of subtle things we can do to make slanted romans better. I have got to the point of calling cursive italics politically correct italics.

The Good Apprentice? But I am having trouble rereading it.

Viva Tourist! Re the variant forms, has anyone ever seen anything like these before? Was it really Hess who conceptualized them first? However, even if Renner did get the idea of the variants from Hess, there is no further comparison to be made. The two weights of Steile and its italic are executed on a supersubtle plane never, I would argue, achieved before or since. The weight relationships between the medium and bold alone are a miracle of non-interpolable brilliance.

But Nick is always so careless, this has been going on for years. What if people actually took this kind of thing seriously, some future researcher?

Bill, this is the internet, it's collaborative. The idea is we act as each other's editors, and the thread is bigger than any one author. If someone expresses themselves poorly or gets a fact wrong, hopefully someone else will ask for a clarification, or correct it -- as you have -- but without using it as an opportunity to gloat and preen.

As for "always being so careless", it appears you are saying that in an attempt to justify your rudeness.
Or it may be that you think it is careless of someone to hold different opinions to your own!
For instance, you say "Yet, as we know, highly related italics (i.e. slanted romans) don’t work," as if it were a fact, thereby condemning those of differing opinion to carelessness. I prefer Eben's position that "absolutes rarely exist in type".

Speaking of slanted romans, compare the original Optima italic to the one in Optima Nova. It has been said that Hermann Zapf had always wanted a calligraphic italic for Optima; nonetheless, in my opinion the slanted roman works better.

So now it looks like there is a choice of italics for Optima, somewhat like Fairfield.

Yesterday I removed some of my posts, because it lead to "off-topic" discussions about the quality of some typefaces.
I have tried to turn Peters' question around and asked myself: What determines that types are NOT "family"? If typefaces, which don't have the same name, can they still be related? Is there another way of looking at "family-relations"?

If you look at several different typefaces from the same typedesigner, you can see some similarity. One can recognise types drawn by Luc[as] de Groot (compare Corpid and Thesis) . The designers' role is important, I think.

Another thing about family-relations has to do with 'derivation'. (I used an online translator for this word) Many designers talk about influences, inspiration or "based upon...". Every typeface is based upon its predecessors or certain elements from an ancient face are re-used in modern types. It's allmost like there is some family-tree which passes on "type-genes" onto the next generation.

Example: Different versions of Garamond are based upon the same "genetic code" which lies in the orginal Garamond (15th century). Still they look pretty different (they have mutated?)

Other example: In FF Legato Evert Bloemsma used characteristics from traditional faces (like Bembo, Garamond and Lexicon) and reused the diagonal contrast in a sans-serif to improve readability.
On the micro level of typography a typedesigner can be challenged by the idea to (re)connect, at least optically and to some extend, the single units of our script into ‘whole’ words**. Most typedesigners are aware of this but at the same time there has always been a strong tendency to build rhythms of thick vertical stems. Not only since Bodoni was drawn in the eighteenth century but even already in the Middle Ages when books were handwritten ( below). This might be a consequent expression of the concept of our alphabet and culture (splitting things up into units and building uniform patterns and sequences) but too much stress here isn’t nice (to read).
(Originally from evertbloemsma.nl, but the website is offline)

Is it possible to detemine the "genetic markers" in typedesign?

P.S. English is not my native language and therefore some phrases might look weird.

Arjen, I don't think it's that complicated.

On the one hand, it's required that family members should share one or more interpolable dimensions (vertical metrics, stroke contrast, and stem width are most obvious). Horizontal proportion can't be measured quite so accurately, but "eyeballing" the similarity works fine.
On the other hand, if the requirement of a shared interpolable dimension is met, then if the foundry says it's a family, it is.

The oeuvre of certain type designers, such as Gerard Unger or Eric Gill, contains many physical congruencies that satisfy the first stipulation for widespread family grouping, but ultimately unless the fonts are published as a family, they aren't.

The "genetic" metaphor is off, because unlike the biological world, in type different generations of the same typeface are not considered to be part of the same family.

A social metaphor for family may be be more correct; sharing the same household.

>So now it looks like there is a choice of italics for Optima, somewhat like Fairfield.

Nick, just for the record, Ruzicka didn't design the 'caption' italic variants of Fairfield. These generic simple italics were designed by various people at Lino (amongst them Kaczun) originally at the behest of National Geographic, if I have the story right. They're not very elegant or characterful, they're not in the least authentic, artistic, aesthetic, or anything like that. But they are awfully functional -- and isn't that what type is supposed to be? PS anyone whose only acquaintance with Fairfield is the digital should go out and find some metal settings -- it really is a good type.

I can't say I have a real problem with Optima slanted. For me the focus on these problems is trite. What Optima needs is not a cursive italic. What it needs is size-optimized weights. Hype, hype, hype! Size-opt doesn't sell, therefore, it's ultimately of little interest to manufacturers

>I have tried to turn Peters’ question around and asked myself: What determines that types are NOT “family" ...

Brilliant thinking, because if you can't say what something is not, you can't say what it is. It makes mincemeat of the original question. If you consider all the metal sizes and weights of Times New Roman, you would be hard pressed to say they were related ... obviously they are, but not very closely and in some cases scarcely at all. The relationship is more a matter of intention than of organic fact. Another way of saying what you have already said. So in Times metal, there's a lot of apparent conceptual conflict, but functionally, it all seems to have worked out.

Maybe I have not understood the original question very well. I thought he was looking for a completely different paradigm for family than the traditional weight/slant concept. Much of the variation in types like Times is do to different people doing the work and on different equipment. I still think Peter is looking for some way to describe family relationship beyond the typical.

ChrisL

> I still think Peter is looking for some way to describe family relationship beyond the typical.

Does that presuppose that there is a persusasive way to describe 'typical' type family relationships?

I disagree that the differences in metal Times are due primarily to diff people doing work/diff equipment. There was still an overriding art director, Morison. I think function dictated the differences -- function and exigency. But that is merely an opinion. We can never really know.

Nick, just for the record, Ruzicka didn’t design the ’caption’ italic variants of Fairfield.

I never said he did.

Size-opt doesn’t sell, therefore, it’s ultimately of little interest to manufacturers

Actually, "size-opt" is of great interest to manufacturers, especially if their font development model involves serifed types for newspapers and magazines. Think H&FJ, FB, Peter Bilak's Greta, or Josh Darden's Freight. And then there's Robert Slimbach's neo-renaissance types for Adobe -- Garamond, Garamond, Minion, Jenson, and Arno.

In my experience as a manufacturer, "size-opt" is a good product, (a) because it sells, and (b) because it doesn't take a huge amount of work to extend optical weights (Bold Italic is far more trouble). I've made optical variants of a variety of types -- usually one size sells better than another, but it's hard to predict which it will be.

The idea is we act as each other’s editors, and the thread is bigger than any one author.

Exactly.

Bill, your points about size specific type are interesting. Is that a theory or based on actual comments conversations? I honestly would love to know! Certainly not every type is worthy of size specific treatment. Just for the record I think size specific is uber-hot feature in a face.

What you were saying about Bruno is fascinating.

Maybe I have not understood the original question very well. I thought he was looking for a completely different paradigm for family than the traditional weight/slant concept.

Exactly. I too went back to re-read the question and I agree with Chris he is asking not about the typical sense of family but instead about some way of talking about as he says character or humour. The sense I get is that he would prefer that there be some more discrete more analytical way of describing things than the 'feeling' 'voice' or 'humour' of a font. To my knowledge there isn't such an intellectual structure existing yet. The substitute for that I have seen is direct comparison of glyphs across weights to describe how they go bold for instance; or between a glyph - such as a lc a between two fonts - made by the same person. So while I understand the urge - I think the whiffly nebulous terms like 'voice' may be the only recourse when speaking of the whole font. At least for now. I would love to be wrong about this.

Peter?

Arjen,

On the micro level of typography a typedesigner can be challenged by the idea to (re)connect, at least optically and to some extend, the single units of our script into ‘whole’ words

Your points about Evert are confusing to me. On the one hand you are talking about the contrasts but on the other you are talking about the whole words. It sounds like you might be making a synthetic point there - but I am not getting what it is. Also, how does this relate to the sense of 'family'?

I should also mention that Evert's ideas about whole words have intense appeal to me. If you have more of writting from Evert on this topic and are willing to share ( offline - or on another thread, because it's not on topic for this thread ) I would be grateful.

Eben,
This is a very fast and general summary:
Traditional faces, like Bembo have diagonal contrast. In sense this connects the individual letters and creates "better" word-images (NL:woordbeelden) which improve readabiltiy and/or legibillity.
This has (for as far as I know) not been seriously used in sanserifs. Evert mentions some typefaces which have such feature, but they are "overdone".

The point I wanted to make, but Nick corrected me allready and I believe he's right: Different sources of inspiration could been seen as "fathers" and "mothers" of a new typeface. The carry the genes of their predesesscors (?). Because the influence of the designer is so big, this aspect of being a family is not suitable for type.

I'm not sure if I said it correctly...

When someone is making a font with a certain ’feeling’ ’voice’ or ’humour’ one would naturally think of a purpose for this font. A 'sad' typeface should be use for 'sad' communication: someone died...

Thanks Arjen!

A ’sad’ typeface should be use for ’sad’ communication: someone died...

I am not sure the voice of a font is best thought about in that way. Instead I think voice is a factor in the overall effect. Put another way; voice is important is only one aspect and may even be deliberately worked against to good effect.

It is definitely time for me to go back & re-read Evert!