The Death of Serif Fonts

[Ryan Maelhorn]

Font selection isn't media darwinism.

[sans sleep]

@Héctor
Self distraction.

@Hrant: It's incredibly readable.
Allow me to disagree. Particularly the words underlined in the image below require some extra effort.

[hrant]

I said NotCarson is incredibly readable... It was a joke. :-)

hhp

[sans sleep]

Pardon my ignorance.

[quadibloc]

@rs_donsata:
Once people tought radio would kill the newspapers, then television was meant to kill radio, then internet was meant to kill television.

It's true that newspapers aren't dead yet.

However, television has meant that people only listen to the radio for music. You don't hear drama programs on the radio, like Fibber McGee and Molly, or Little Orphan Annie, or Captain Midnight, or, for that matter, Gunsmoke - which was on the radio before it got moved to TV - any more.

And the Internet is filling many of the otherwise idle hours of people that they might have spent watching television. And the effects of the many channels of TV on local broadcasters before that are well-known as well.

Now, what might be the next big thing after the Internet... well, whoever brings it about stands to make billions of dollars, so if I were able to foresee it, I would be doing it, not telling you about it. But I agree that since the Internet is so flexible, and it allows people to create, not just consume, it would seem like any future media could be placed on the Internet rather than superceding it.

Although it's hard to monetize stuff on the Internet, so that might be an excuse for putting stuff elsewhere - but then it's not likely that whatever it is would be nearly as big as the Internet, since it would cost more.

[Ryan Maelhorn]

I actually think NotCarson, and it's second cut evenmore so, is incredibly exquisite and beautiful. I imagine this is most due to a type designer's slavery to systematic thinking more than any rebelliousness.

but all fancy lines die in 2375.

[hrant]

Nevermind...

hhp

[quadibloc]

@Nick Shinn:
The idea that serifs will become functionally obsolete is a nice little myth of modernism.
Whether serifs are useful to the process of reading-as-deciphering-text is besides the point.
Type forms are invested with far more meaning than that, and it is not just an aesthetic veneer.

You may be 100% right, but I tend to be inclined to take issue with some of this.

I fully agree with your first paragraph... except that I'd be inclined to add "in the foreseeable future" as a qualification. I don't think it's absolutely impossible for serifs to go out of fashion.

While it would seem incredible to me if typefaces like Helvetica, Univers, Futura, and Gill Sans pushed out serif faces, that the evolution of typographic preferences could go in the direction of Optima, Radiant, Lydian, Freehand... in short, towards some other cues which add redundancy to the letterform... thus displacing serifs as the attribute of choice seems not entirely beyond the bounds of possibility.

When it comes to your second paragraph, I'm not sure that my unease with it is because of disagreement. I feel uncomfortable making definite categorical statements about matters of taste. My aesthetic preferences may not be the same as those of many or most other people, now or in the future.

So, if serifs aid "reading-as-deciphering", at least that's one objective straw I can hang on to in an argument. I may feel that serifs will continue to be liked for aesthetic reasons, but what can really be said with any definiteness about the fickleness of public taste and fashions?

Of course, we all know the story about Hitler and Fraktur. It's not beyond imagination that the Western world might be saddled with regimes that consider serif typefaces to be "bourgeois" and proscribe them. If that keeps on for long enough, even after liberation, sans-serif faces will be what people are used to and hence find legible.

When it comes to your third paragraph: I'm sure you're right, but I find your meaning to lie somewhat beyond my grasp.

Aside from the "aesthetic veneer", what is the meaning of type?

One example might be that, at present, a sans-serif typeface sends one kind of message when used in, say, a letterhead... and a serif typeface sends another. A student newspaper at a college might be set in Univers, but never a major daily.

But that kind of meaning tends to be relative to whatever the current baseline of fashion might be. So it doesn't seem to be much of a bulwark against a changed world in which using Optima instead of Univers sends the same message that using Times Roman instead of Univers currently sends.

I don't think you were instead talking about how typefaces allude to Jenson and Aldus and company; that would invite derision if it were meant to describe the habits and reactions of the general reader without typographic sophistication.

So this is why I'm having problems with your posting. I may, like you, feel convinced because of my strong subjective preference for serif fonts in some contexts that the Latin-alphabet using portion of the human race won't turn its back on them any time soon. But that seems to me to be just a personal opinion, rather than a finding based on hard empirical data, so I would feel ashamed to present it as a reliable conclusion to other people.

And, as an example drawn from recent discussion here, although today Armenian is often printed in typefaces with serifs in imitation of Latin and Cyrillic typefaces, it got along for decades with the Bolorgir style of type which had stroke width variation but no serifs.

For that matter, I've seen Gujarati typeset in faces that imitate Latin and add serifs, but that's not common except as a style of display type.

I suppose, then, after being reminded of this by mentioning Gujarati (and there have been illustrations here of even Korean being typeset in mock-Latin styles) the strongest argument for the persistence of serif type is that even if it does fall strongly out of fashion as a primary text type, it will still exist as a stylistic option - and stylistic options don't tend to get completely discarded.

[oldnick]

Aside from the "aesthetic veneer", what is the meaning of type?

Type has no meaning per se; rather, type is a means to an end. The “æsthetic veneer” is the more or less unique attitude which an individual typeface’s design contributes to the process of communicating information, whether of a factual or fanciful nature.

Serifs will never “die” because memories have emotive power; from time to time, we revisit the past and mine those memories, for better or for worse.

[Nick Shinn]

Myths may or may not be true.

Modernism was inherently reductive with respect to form, and classic modernism removed serifs, considering them to be unnecessary ornaments. That is the myth.

However, perhaps Modernism should have embraced serifs, if they do indeed perform a useful function to the process of reading.

Certainly, traditionalists believe that serifs are indispensible for optimum readability. That too is a myth. The present scientific intelligence, as far as I am aware, offers no evidence one way or the other.

In reading science, the myth seems to be that, beyond a few easily-identifiable categories, all the many different type styles which exist have no bearing on the process of reading-as-deciphering (no doubt because they don’t effect reading speed). That is what I mean by aesthetic veneer.

However, I find it hard to accept that typeface subtleties exist only to cue the reader into what sort of document she is about to read, and then vanish from awareness once she becomes immersed.

In art theory, the “innocent eye” was long ago discredited.

[oldnick]

Nick,

Not to put too fine a point on it but, in a classical sense, myths are always true because they explain First Principles. It is an unfortunate latter-day tendency to conflate “myth” with “urban legends” and the like.

Few things, once noticed, really vanish from awareness: more often than not, they're just moved to the back burner.

[Nick Shinn]

To put a finer point on it: Modernism and Traditionalism may be considered myths, in both of which serifs play a role irrespective of any empirically demonstrable function.

In those mythical roles, serifs are true to the precepts of the myths, whether or not it’s true that that they improve readability.

That makes perfect sense to me, because from my perspective as a typographer, more practical than ideological, I have always found it very useful to combine both in layouts with complex hierarchies of text, the contrast between sans and serif type proving practically indispensible.

[hrant]

although today Armenian is often printed in typefaces with serifs in imitation of Latin and Cyrillic typefaces, it got along for decades with the Bolorgir style of type which had stroke width variation but no serifs.

It's certainly true that many contemporary Armenian fonts -like Sylfaen- have too many serifs (which isn't just an æsthetic problem - it causes readability issues) and that mostly comes directly from Latin (and I guess Cyrillic) type. But traditional Armenian letterforms aren't exactly sans either... They often have what are basically serifs that are in fact so long they often get confused by non-natives for horizontal bars! The key thing to realize is the strong role of slant in Armenian authenticity, and that throws an interesting wrench into the sans-versus-serif debate, as well as the true nature of an optimal "italic" for Armenian. The latter probably remains the single biggest question mark in my own head.

Type has no meaning per se

I would counter that everything has meaning, especially something visible (since the human reality is so strongly dependent on vision).

The present scientific intelligence, as far as I am aware, offers no evidence one way or the other.

Neither does reading entrails, which frankly seems qualitatively equal in scientific rigor to "present scientific intelligence" concerning reading; but that still leaves us with regular intelligence, which points squarely towards serifs helping reading.

hhp

[quadibloc]

@hrant:
They often have what are basically serifs that are in fact so long they often get confused by non-natives for horizontal bars!

Hmm. Maybe this explains why some Armenian capital letters - or initials - have horizontal strokes so thin that they're actually nonexistent.

So some strokes that are believed by outsiders to be elements of the character are actually just large serifs? Then, indeed, serifs on serifs would create a confusing fractal effect...

Although if the Armenians in Armenia are happy with thoroughly Latinized typefaces, and only those in the diaspora, under the pervasive influence of other scripts (but, on the other hand, the Arabic script, the main outside influence on Armenians in Lebanon, unlike the Latin script, would seem to be less likely to have a corrupting influence...) value the distinctive characteristics of Bolorgir... then one might despair of finding anywhere an authentic authenticity of Armenian visual sense.

So I suppose one will have to venture to contact Armenians in Lebanon to find out their typeface style preferences... although they likely have far more pressing concerns.

[hrant]

horizontal strokes so thin that they're actually nonexistent.

I have a theory that this might come from blindly emulating eroded inscriptions.
http://armenotype.com/wp-content/gallery/highlights/dsc_0377.jpg

In terms of the difference between long one-sided serifs versus crucial elements, it can get pretty subtle. Here's an illustration from my recent Armenotype article*:
http://armenotype.com/wp-content/uploads/2012/04/illustration2.gif
That's from my first-ever outline font, which I made by staring (without magnification) at the letterforms in a document set in Compugraphic's Barz (possibly the most ubiquitous Bolorgir, at least back then). Look at the bottom-left corners of the descenders: it was probably hopeless (I was too inexperienced to appreciated the total disappearance of tiny details at text sizes on laser printers on lousy paper...) but I tried to implement a difference between serifs versus actual horizontal bars - the «ի», «խ» and «կ» have serifs, while the «լ» and «ղ» have bars.

* http://armenotype.com/2012/04/archipelago/

if the Armenians in Armenia are happy with thoroughly Latinized typefaces

Not all of them are happy. And some of them are "happy" because they see it as a necessary evil, sort of how -I and others believe- Ataturk converted Turkish from Latin to Arabic - to play along with the victorious West after WWI. As for the rest, I hope they'll attend my talk in Yerevan in the second week of June... :-)

--

The Armenian communities in Lebanon have severely weakened over the years, and many of those that remain have trouble staying Armenian (something our US communities were previously infamous for). It's ironic that they were strongest at the height of the Lebanese Civil War. I still have a lot of family there (including a cousin who used to be the editor of Aztag, once the premiere diasporan newspaper) but sadly I can't consider the Armenian diaspora in Lebanon to be trend-setting in any way these days.

BTW, Lebanon -still- has two official languages: Arabic and French. Coupled to the fact that English is the lingua franca, any Lebanese person (especially an Armenian one, since we're more Western than the average Lebanese) is at least as influenced by Latin forms as Arabic ones. There are even certain strata of Lebanese people who want to deny they're Arab, to the point of making some Arabic words sound French. :-/ A classic example is the word for rat, "jardon" (affectedly pronounced with a "gh" for the "r", and a nasal "n") which in fact is nothing but the Arabic word "jirthaan"! Lebanon is funny. Anyway, I digress...

hhp

[dberlowgone]

HP: 'Neither does reading entrails, which frankly seems qualitatively equal in scientific rigor to "present scientific intelligence" concerning reading;'

[He's kidding again.]

"...but that still leaves us with regular intelligence, which points squarely towards serifs helping reading."

...if device resolution/printing allows them to do their job... if not, according to regular intelligence, they almost all must fall off.

[hrant]

If the resolution is low enough they do get too thick/tall, but I wonder if strategically letting them touch (laterally) might actually promote boumas. Serif-wise ligation.

And to in a way rephrase something I wrote above: contrary to what tradition might say, serifs don't need to be an all-or-nothing proposition. We should really be looking at where they need to go beyond "well, that's how my pappy did it". Does this in effect make them a "structural element" instead of mere bling? Unsettlingly, but yes it must.

hhp

[dberlowgone]

HP: "I wonder if strategically letting them touch (laterally) might actually promote boumas"

That's an interesting question, of course, but if the market is any indication, and outside of high resolution-specific specifications, (e.g. the Kindle), serifs have been relegated to the back seat. And I'm pretty sure from the designs of nearly all serif text faces that I'm familiar with, that serifs touching is not beneficial.

"...serifs don't need to be an all-or-nothing proposition..."

But when it comes down to executing a readable design in low resolution opentype-free environments, some-serif fonts show serifs on only I, 1 and maybe a few others. Why? "Sometimes-some-serif" fonts might be more promising, when contextual layout is more widely used. In the meantime, I'd very much like to see your trials of a bouma-optimized sometimes-some-serif font, which I think you think you can do now. The place I think feets could be more beneficial are on beginnings and ends of words, and on stems above and below the x-ht, incl. caps.

[hrant]

But/and what about "traditional" semi-serif designs, where the leftward ones (typically on heads) and rightward ones (feet) can still maintain one-pixel gaps (without stems becoming too distant)?

hhp