Council for German Orthography officially allows use of u+1E9E

Christian Thalmann

That definitely looks out of place among the uppercase. No reason not to stretch that almost-vertical into a full vertical. It's based on a long s after all.

Andreas Stötzner

I should mention that a few days ago, the entire publication Signa Nº 9 – Das große Eszett – has been made publicly available under CC-licence. I hope it furthers insight on the matter, even though it’ll be not everyone’s cup of tea to read German.

https://www.pdf-archive.com/2017/07/13/signa-9-ausgabe-2017-cc/signa-9-ausgabe-2017-cc.pdf

John Savard

Christian Thalmann said:

Are you referring to the S3-like design?

Yes, that was the design I was referring to.

Christian Thalmann said:

That definitely looks out of place among the uppercase. No reason not to stretch that almost-vertical into a full vertical. It's based on a long s after all.

The lowercase one is, yes, but since there is no upper-case long s, the S in the uppercase version should be based on an uppercase S, rather than trying to create an uppercase long s, I feel. Basically, I am trying to grasp at existing models to avoid creating anything new, as far as possible.

Christian Thalmann

John Savard said:

The lowercase one is, yes, but since there is no upper-case long s, the S in the uppercase version should be based on an uppercase S, rather than trying to create an uppercase long s, I feel.

And yet it is the long ſ that is baked into the very concept of ß. None of those awful train-crash chimera designs read as a ß instinctually. Sure, if the context is obvious enough, you can guess its meaning, but then you might just as well replace it by a picture of a potted plant. Any glyph that requires analytic deduction to decypher will necessarily feel contrived.

Basically, I am trying to grasp at existing models to avoid creating anything new, as far as possible.

You are creating strange new things. If you wanted to avoid creating new things from scratch, you'd take the ß and lead it into capitalhood along the straightest possible path, changing as little as possible about its identity, as has been done for many other Neolatin characters, and as has been naturally done by non-typographer German users in handwriting for many decades.

Hrant Հրանդ Փափազեան Papazian

> Any glyph that requires analytic deduction to decypher will necessarily feel contrived.

At first.
Then it might pay off.

> changing as little as possible about its identity

No, changing the right things.

John Savard

On further reflection, I see what is wrong about my suggestions.

The form of the capital eszet in any given typeface should correspond to the form of the lowercase eszet in that typeface.

It is true that in a few Roman typefaces, the lowercase eszet is a character that is obviously the ligature of a long S and a short S. In those typefaces, the capital eszet should be the sort of thing I originally preferred: an obvious ligature of two capital letters S. Not with a middle bar, as I first suggested, but by joining them with a line at the top, since this form has already been used (or at least proposed) historically.

My ulterior motive is that, someday, some other language using the Latin alphabet might feel the need for an ss ligature in both lower-case and upper-case form, and thus it would be nice for the uppercase eszet to be solidly rooted in Latin alphabet tradition so that the glyphs as well as the codepoint could be used by languages other than German if need be.

In other typefaces, where the eszet is in its usual present-day form, to maximize legibility and differentiation from the capital letter B, I would suggest a form that is the Dresden form on the left, but the Frankfurt form on the right. Again, I think that form is already in use in some typefaces. (Well, Boxley, Canape, and Lenga come close, but they seem to be the Frankfurt form with a rounded corner on the left rather than the Dresden form on the left.)

If the foot of the lower-case eszet on the left side descends below the baseline, what about the foot of the upper-case eszet on the left side?

I would suggest that it should not descend below the baseline, to further help distinguish between the upper- and lower- case forms of the letter, unless the typeface is one of those in which the capital letter J descends below the baseline. However, I am willing to make an exception in the case of those typefaces where the capital letter U looks like an enlarged form of the lower-case letter u by having a foot on the right.

John Savard

Nick Shinn said:

True, but isn’t the very idea of oldstyle German type an anachronism? Weren’t all historic German types of the XVI–XVIII centuries in the Fraktur style?

Yes, the idea of "oldstyle German type" is an anachronism.

However, I do not think it is anachronistic or otherwise improper to set the German language in Garamond or Bembo... or, for that matter, in Palatino or Aldus.

As I noted in the thread about changing the slope of the acute and grave accents for some languages, while some typefaces have had special versions made for different languages where the letterforms themselves are designed, and while one famous book on typography did illustrate how Bodoni worked better for some languages than others and so on...

I don't think we should encourage a school of thought that says that any language written in the Latin alphabet should only be typeset in typefaces historically used with that language - or newly designed and derived from the tradition of those typefaces. It's fine and dandy if a Czech type designer designs a new typeface by drawing on the history of Czech typography. But to deplore its use to set text in Spanish - or the use of a typeface designed for Spanish to typeset Czech - well, the basic result would be to impoverish the typographical choices of speakers of a host of languages.

National cultures should be preserved and cherished, but they should not be allowed to limit people's choices or otherwise dictate to the people to which they belong.

It was Francis Bacon who said "Money is a good servant, but a bad master", and that quote has been paraphrased to apply to other things, such as habit and technology; the same principle applies to one's national culture.

John Savard

From

https://typography.guru/journal/germanys-new-character/

I learned that the need for the capital eszet had become critical since the orthographic reform of 1996 (although perhaps I did so by misunderstanding what I read). But it also noted that German speakers in Switzerland eschew the eszet.

This created a thirst for more information on my part which led me to Wikipedia. From there I learned that the 1996 spelling reform affected all the main German-speaking countries, including Switzerland; it was not something that happened in Germany only.

Of course, this raises the question of just how the Swiss can manage without it, if an uppercase form is now critically needed due to the reform - but the answer is clear enough; Swiss schoolchildren had to memorize the things missing from their country's spelling system to identify words, just as English and American schoolchildren must memorize the things they need to cope with the spelling of their language. Germans could not be expected to instantly pick up a large body of knowledge for which they previously had no need.

Also, I learned from Wikipedia that the long-s with round-s ligature was not the sole source of the eszet; there was also a long-s with z ligature, and the eszet is currently used where either of these were used. Obviously, then, if I was inclined to advocate that a capital eszet should be unmistakably built from two capital letters S, I am the sort of person who, based on that, would now (be daft enough to) suggest that this distinction be restored. (Since the eszet is also sometimes called the sharp-S, one could split the two names between the two characters!)

Wikipedia noted only that there were rules for which one of these to use; no doubt they are recorded somewhere in German, at least. (A different Wikipedia article explains that this distinction comes from the Heyse spelling, and is absent from the Adelard spelling, so now I know where to look.)

John Savard

John Savard said:

In other typefaces, where the eszet is in its usual present-day form,

I see now that the proper term for this form is the Sulzbacher form.

Oh, yes: here is what I would be tempted to inflict on the German speaking world:


but with the option of using this instead



and the rule for deciding which eszet to use is also a simple one: when the new orthography calls for an eszet, it is the sz one; when an eszet was formerly used, but the new orthography took it away (due to switching from Adelung to Heyes) use an ss one!

Andreas Stötzner

@JohnSavard: with all due respect, John, your lenghty elaboration explains less than it pretends to; the images in your last posting are completely unsuitable to make anything more lucid or to guide anyone in the right direction.

Have a break, read this brand new article, if you wish:
https://qz.com/1033265/germanys-century-long-debate-over-a-missing-letter-in-its-alphabet/

Nick Shinn

I continue to be impressed with the title of this thread, so important sounding!—authoritative, erudite, technical and obscure. And with an f_f_i ligature to boot. One more time:
Council for German Orthography officially allows use of u+1E9E

Hrant Հրանդ Փափազեան Papazian

> the images in your last posting are completely unsuitable to make anything more lucid or to guide anyone in the right direction.

Too conservative.

John Savard

Andreas Stötzner said:

or to guide anyone in the right direction.

It does make sense that only the native speakers of German could really see what the right direction is.

But I know I have learned a lot from both postings in this thread and from searching for more information on the topic.

Yes, the notion of having two versions of the eszett in German, even though that was true for the lower-case eszett in Fraktur typefaces for a time, most likely is ludicrous. And, even if the 1996 spelling reform was not entirely popular, bringing back the lost eszetts but as a different form of the eszett is the sort of compromise that leaves both sides unsatisfied.

Thus, while I accept the charge that I am supplying suggestions that may point in the wrong direction, I suspect that at least for some I am making things more lucid, even though it is only by relaying information I have seen elsewhere: the central fact being that the capital eszett is needed because it affects the pronounciation of the vowel before it. (And the eszetts dropped by the 1986 reform are the ones that didn't do that, and they are the ones that, in Heyes orthography, were represented in Fraktur by the ss-ligature instead of the sz-ligature.)

Christian Thalmann

I don't see the merit of introducing a ligature for /s_s — after all, the sequence behaves exactly as expected. A vowel followed by a consonant cluster is short by default. Otherwise, you might as well call for an /n_n ligature.

As I understand it, part of Andreas' reaction to your graphics was also the fact that even in the lower sample, which is acceptable usage, the /ẞ is extremely ungainly, and if anything we shouldn't increase the number of bad examples in proscriptive materials.

Hrant Հրանդ Փափազեան Papazian

An "nn" ligature might be just the thing, if done right and deployed judiciously. The true power of ligation still eludes us (largely due to the disease of Modernism).

Andreas Stötzner

John Savard said:

… that only the native speakers of German could really see what the right direction is.

That is not the point. Neither was it the direction of my remark.

John Savard

Andreas Stötzner said:

That is not the point. Neither was it the direction of my remark.

I certainly will affirm that your post was addressed to the substance of my posts, and was not in any way an out-of-hand reaction to someone not a speaker of German commenting on this matter, and I regret if what I posted appeared to imply otherwise.

I am merely admitting that my qualifications in the matter are limited.

However, based on what I have been reading about the issue of the upper-case eszett, even my seemingly outré suggestion that two eszett characters, one an ss-ligature and one an sz-ligature, each with its own upper- and lower- case versions (with the sz-ligature ones using the existing Unicode codepoints) is, in fact, defensible.

The two major schools of thought on German orthography - Adelung and Heyes - each have something to commend them.

In the Heyes orthography, the eszett is used only in those positions where it makes a difference to the sound of the preceding vowel. Where that does not happen, "ss" is used instead.

But the Adelung orthography addresses another concern: a consequence of the Heyes orthography is that a number of German words will have "sss" - the letter s, three times in a row - in them, and this interferes with readability.

And, so, if one had a second eszett to put in the Adelung-only positions, occurrences of the letter s three times in a row could be avoided - and, yet, because the second eszett is distinguishable from the regular eszett applied according to the Heyes orthographical rules, it remains unambiguous how vowels are to be pronounced.

Also, since there are historical texts typeset using Fractur typefaces in which just these two forms of the eszett were provided, asking for a second lower-case eszett in Unicode, at least, is legitimate on the ground of being able to accurately quote those texts and utilize electronic versions of their typefaces.

Christian Thalmann said:

even in the lower sample, which is acceptable usage, the /ẞ is extremely ungainly, and if anything we shouldn't increase the number of bad examples in proscriptive materials.

I will certainly admit that the quality of my artwork leaves much to be desired.

Andreas Stötzner

> historical texts typeset using Fractur typefaces in which just these two forms of the eszett were provided …

Would you care to prove that with a sample?

> asking for a second lower-case eszett in Unicode…
… will get you a dead sure NO from ISO and UTC because this is not the sort of things which the level of text hard-coding is meant to be for.

John Savard

Andreas Stötzner said:

> historical texts typeset using Fractur typefaces in which just these two forms of the eszett were provided …

Would you care to prove that with a sample?

I got this from the Wikipedia article on the eszett which claimed that this was how German was spelled and printed between 1879 and 1901 in Austria when Fraktur was used.

I have found one illustration, in a journal on German orthography, that shows that the different ligatures at least exist in Fraktur, but I will see if I can find a better example.

André G. Isaak

I only see a single form of eszett in the above example. Can you perhaps repost that image with one occurrence of each form highlighted?

André

John Savard

Here is another example:



None the less, with respect to this example, as to the other one, you are completely correct.

While all three forms of ss (or in one case sz) are present, the only one that is a ligature is the sz-ligature, as present in naturgemäß and Ausmaß. The other two forms are not ligatures; the double long s in Ergebniſſen, or the long s followed by the short s, which is what I have been suggesting as the second eszett, as in daſs at the beginning of the selection.

John Savard

Actually, the double long s, if not ſs, which is what I actually wanted to show as a ligature in Fraktur, is or can be a ligature, as can be seen from this illustration of a printer's type case for lower-case Fraktur:



on the right-hand side in the top row, just to the left of the eszett.

Thus, my examples attest to the orthographic existence of both ſs and ſz (ß), but not their typographic existence as separate single characters.

So one could use existing Unicode characters to avoid the occurrence of "sss" by writing words like Paſsstraße - returning the long s as a character to German typography in Roman (Antiqua) typefaces and not just Fraktur. A ligature, though, would still be needed as a way to make it possible to distinctively capitalize the long s (although a capital long s as an individual character exists in the alphabets of at least two languages).

John Savard

Here is another illustration of the sort of thing I propose:


For the upper case, SS and SZ, as ligatures or even as logograms, like the Dutch IJ. For the lower case, the Antiqua ſs and ſz or ſʒ ligatures, basically the ancestor of the Sulzbach form and the Fraktur form... so that not only is there a form for the uppercase eszett, but as well the issue created by applying the Heyes s-rule to Antiqua faces instead of Fraktur faces of having the letter s three times in a row is solved.

Hrant Հրանդ Փափազեան Papazian

Interesting. If you could fuse the two letters into an elegant whole, it might be an option.

Christian Thalmann

It's a contrived non-solution to a contrived non-problem. A ſs ligature only makes sense in a world where the ſ/s distinction is still present. It's been abolished for good reasons.

There is nothing wrong with the spelling «Passstraße», no more than with or «Schnellleim» or «Krummmesser» or «Sperrriegel» or «Spannnagel».

In fact, a spelling that reduces «ss» to a single glyph misleads my reader's intuition to pronounce the preceding vowel as long, since there seems to be only a single consonant following.

Johannes Neumeier

John Savard said:

Here is another illustration of the sort of thing I propose:


For the upper case, SS and SZ, as ligatures or even as logograms, like the Dutch IJ. For the lower case, the Antiqua ſs and ſz or ſʒ ligatures, basically the ancestor of the Sulzbach form and the Fraktur form... so that not only is there a form for the uppercase eszett, but as well the issue created by applying the Heyes s-rule to Antiqua faces instead of Fraktur faces of having the letter s three times in a row is solved.

Your examples bring to mind another fraktur quaint, namely stacked vowels for umlauts, but I dislike that form equally much:


Also, and this might sound vague, but the sharp left top cornered ẞ variant to me feels russian... these boxy shapes at the top, like in П and Л, just stick out like a sore thumb in a text set in german, to my eye anyway.

Hrant Հրանդ Փափազեան Papazian

Interesting observation on Russian. You have a point, although to me the more authentic form of Er for one has the top-left corner round.
https://typography.guru/forums/topic/480-designing-the-cyrillic-er/

John Savard

Johannes Neumeier said:

Your examples bring to mind another fraktur quaint, namely stacked vowels for umlauts, but I dislike that form equally much:

I do not disagree with this.

I think that particular shape for the capital eszett is flawed; it might belong in University Roman, but it would not be fitting to most other typefaces.

I had wished to avoid trying to have an uppercase long s, because I view that character as belonging only to the lower-case in the Latin alphabet, but it has a clear practical advantage that the failure of that shape illustrates.

So, in this diagram


on the left you see what my own personal preference for a shape for the uppercase eszett is, to remain consistent with the genius of the Latin uppercase alphabet as something to be carved into marble with a chisel.

And on the right is the compromise I am now willing to accept as better suited to the needs and habits of German speakers.

Both shapes could be easily modified to turn them into ligatures,


but Greek manages perfectly well with Ξ, Russian manages perfectly well with Ы, and Dutch manages perfectly well with IJ, so I do not believe there is any real need for that; the dot under the Z in the first case, and the nesting of the Ezh in the second case under the upside-down J that is the upper-case long s, are quite sufficient to make it clear that this is a single glyph and not two.

John Savard

As for some other languages: I have left a comment (which will need to be moderated, as I have just joined to comment) on Typography.Guru to note that it is Poluustav of which Hrant was thinking when he suggests a Russian Er should have a rounded corner, and that it's a misconception to think of Cyrillic as a descendant of Latin, as it was directly derived from Greek (although Latin influenced it too, as the pairs Aa and Ee show) - and, also, it's worth mentioning that the 1996 spelling reform in Germany is nothing compared to what is currently being proposed in France.

And, of course, not to claim to erudition which I do not possess, I had to poke around on the Internet to find that "poluustav" was the name of the thing I sought to mention.

Personally - while the Greek uncial in, for example, the Codex Hercleianus, is what Cyrillic writing was derived from, and in that script, rho had a rounded top left corner, Greek writing has changed since that time, and I don't see why Russian writing shouldn't be allowed to change as well.

Christian Thalmann

Just noticed I had an ungainly form of /Germandbls stuck in Eau de Garamond. Fixed it with the Zürich form.