Sharp S can appear at start of word

Helmut Wollmersdorfer

This week I looked up a word in my collection of dictionaries and was surprised to find sharp S at the begin of words. It's not German, but transcription of Yiddish by Duden into German phonetics. Of course this not an established standard and maybe dates back to the times before World War II, where the JIVO Institute in Berlin was an important center for this language. 

James Puckett

Please post some sample words so we can add them to our proofing documents.

John Hudson

I have heard that there is also a dialect of German in which this may occur, but have not seen examples.

John Savard

I have managed to see this with my own eyes, in the Jiddisches Wörterbuch by Duden; on page 23 of the 1992 edition, I see:

ßofek m (ßféjkeß) Zweifel (se) ...;

as an illustration in section 4.2.4.3 .

Franz Gratzer

German is my mother tongue but I am not aware of any words starting with an "ẞ". I wouldn't even know what dialect could have such a feature. In fact the capital version of the "ß" is even only known in small circles and hasn't even been a thing for very long yet. (However, I can imagine that some academic has decided on specific cases where this letter "should" be used at the beginning of a word. It doesn't seem to matter in real life.)

Many of you probably see the following ideas as barbaric and crude:

I feel that we should actually get rid of this very obscure and not really useful letter all together. Writing "sz" or "SZ" would be totally sufficient for the rare cases where pronunciation might be relevant.

I would be very surprised if even one percent of the population was able to find where the captial "ẞ" letter is hidden on our keyboard layout. (It is necessary to press 3 buttons at once and it is not even mapped on the same key as the lower case letter despite this being the case for all other capital letters.)

I am bold enough to suggest that much more frequent combinations of letters should get their own single letterform like "sch" and "ch" instead. We actually could create more conformity by not using different letters interchangeably for the same sounds. In German "x" are rather infrequent and "c" is either spoken like a "k" on like an "s". Therefore, we actually could make our language clearer by using "x" for the sound of "sch" (using "ks" instead of "x" in the rare cases where we would use an "x"). The same could be done with the very frequent "ch" by writing only "c" instead and actually use "s" or "k" when we want to use the sound of those letters. There are other cases that could be cleared out too. And the German "Umlaute" "ä", ö" and "ö" are mostly annoying since they hinder us to just use the standard English keyboard. It would be much easier to not need a specific regional keyboard layout for typing efficiently. The only annoying thing with replacing "ä", "ö" and "ü" with "ae", "oe" and "ue" is that they are somewhat frequent and we would need two letters instead of one. On the other hand it is already familiar to German speaking people to use this notation when for some reason one can not use these special letters – like in email addresses.

But you don't need to tell me that this won't fly because of tradition and backwards compatibility. Actually, the same seems to be true for the "ẞ/ß" situation.

In addition we haven't been able to find a sensible solution for gender neutral language for decades now. Useless gender specifications are woven into German so deeply that it would need a complete re-invention of our grammar to do so properly without creating ridiculously complicated constructs or versions that are only written and can not be expressed phonetically.

Sorry. I know this isn't really the proper place to discuss this. I just needed to express it in order to explain where I am coming from and why I am not fond of introducing additional complications into our written language.

Matthijs Herzberg

@Franz Gratzer I’m not German, but your opinion seems prudent and reasonable. I’d be inclined to agree with it, if it wasn’t for the fact that I get such a kick out of designing ẞ/ß for my typefaces.

Franz Gratzer

Matthijs Herzberg said:

if it wasn’t for the fact that I get such a kick out of designing ẞ/ß for my typefaces.

Interesting. I struggle with finding appropriate forms for letters that I am not familiar with. I guess in bigger teams people who have a fitting background because they come from cultures who do actually use those letters are appointed to develop adequate letterforms to a specific font. But I can't understand how I should even begin to understand how for example a Cyrillic or Greek letter should be adapted to fit into the formal language of my font face. In many cases legibility depends heavily on tradition and how should I know what specifics I could alter in order to fulfil both the formal language of my type face and the heritage bound forms for that specific letter. Of course it would be possible to start researching each letterform but I fear this would end up in a bottomless hole.

The "ẞ/ß" is a bad example because there isn't actually a tradition for the capital
"ẞ" yet and it is used so rarely that most people wouldn't even know that this letter exists. Its design is usually so close to the lower case that most people would rather wonder why someone used a lower case letter instead of "SZ" in an all captitals headline, which has been the common solution until recently.

Christian Thalmann

Matthijs Herzberg said:

I’d be inclined to agree with it, if it wasn’t for the fact that I get such a kick out of designing ẞ/ß for my typefaces.

I'm Swiss, so I was raised in the mindset that the ß is weird and unnecessary. Now that I design type, I absolutely love it and use it consistently when writing High German. It's just großartig. The ẞ is just the logical conclusion of that train of though.

Andreas Stötzner

Franz Gratzer said:

… legibility depends heavily on tradition …

very true.

… how should I know what specifics I could alter in order to fulfil both the formal language of my type face and the heritage bound forms for that specific letter.

how? by learning about those typefaces.

… Of course it would be possible to start researching each letterform but I fear this would end up in a bottomless hole.

well, no. To research and learn about e.g. Cyrillic (whatever) is what I call work. Without this kind of work you’ll not realise what is going on.

_ _ _ _

* A healthy and lazy New Year’s Day to you all *

Franz Gratzer

Andreas Stötzner said:

To research and learn about e.g. Cyrillic (whatever) is what I call work. Without this kind of work you’ll not realise what is going on.

Since an international font has hundreds (if not thousands) of glyphs I feel overwhelmed by this outlook. Despite it I still suspect that I can't overcome my cultural predispositions that easily. It might be somewhat manageable for us to deal with glyphs that are roughly based on Latin letter forms (even if I feel it is deceiving at times) but it is easy for me to see how hard this becomes when I look at a font with Latin letterforms that has been designed by someone from a very different cultural background. Only yesterday I browsed through a lot of fonts on Google and the letters and spacing of Latin glyphs of such authors looked awkward and wrongly balanced to me. I wouldn't blame them for it because it isn't trivial to grasp the relations if you aren't shaped by this tradition. I bet the same would be true if I for example attempted to design Japanese or Arabic glyphs. Even if I beforehand took the time to study several specimens. And I have no convincing means to overcome that.

But I guess I don't need to do everything. Of course not every font needs to contain all glyphs. But I am not sure yet how far to go is sensible.

I guess the question then becomes what is the minimum set of glyphs you consider practical in a typeface that is meant to be used in at least North America and Europe? If we consider smaller populations we should of course include native Americans too. Even with the limitation to North America and Europe I end up with something that seems to be to much to handle properly.

I suspect, in the end, it very much depends on what I am aiming for. And every font family can be extended in case someone has the need for it. And – to be honest – most likely I will never create anything that becomes that widely used that people would even want extensions of its glyphs set. Even if I ignore special letterforms from neighbouring countries here in Europe that share the Latin alphabet. But then again: A font might might not be used solely because it didn't have the needed glyphs yet. I myself often excluded font faces from my list of families I rather liked to design with just because they didn't support German "Umlaute".

Scott-Martin Kosofsky

For those of you who are not familiar with it, Yiddish is a West Germanic language spoken by Ashkenazi Jews and written in Hebrew script. It contains a significant number of Hebrew and Aramaic words, as well as words from Slavic languages, reflecting the culture of the Jews of Eastern Europe. Yiddish was my family’s secondary household language and I grew up speaking and reading it. In later life, I became involved with it professionally.

The Yiddish word sofek ספק ( the English noun “doubt”) is a direct borrowing from the Hebrew safeik. In Western Yiddish, it was more common to see “doubt” expressed as “Zweifel,” taken from German. Be that as it may, there is no phonological reason to express the initial consonant with an eszett. I don’t have a copy of the Duden Jiddisches Wörterbuch (just ordered one), so I can’t access their reasoning. Did they transliterate all other words beginning with the letter samekh with an eszett? To do so seems absurd to me; to single out this word for such treatment is, on its face, ridiculous.

Adam Jagosz

@Helmut Wollmersdorfer I think what you encountered was a phonetic transcription of the word. Older dictionaries (and maybe still some newer ones intended for the layman) often use the reader's native alphabet in a grotesque manner to present approximated pronunciation at the reader's minimal effort.

Scott-Martin Kosofsky

@John Savard, can you send a picture of the entry? I’m thinking it might be an IPA symbol (International Phonetic Alphabet, not India Pale Ale) rather than an eszett, perhaps one of the sibilant fricative symbols.

Nick Shinn

What is the relevance of this grammatical oddity to type designers?
I doubt kerning would ever be necessary between ẞ and preceding punctuation or a following minuscule/small capital glyph.

Florian Pircher

This would finally make initial swash alternatives of ß and ẞ viable ;​)

Christian Thalmann

Be that as it may, there is no phonological reason to express the initial consonant with an eszet

The reason being that an initial s- is pronounced as [z] by default in German, so a different spelling is needed to remind casual readers of the difference.

I remember trying to use the Greek crash course from the Kauderwelsch series while on vacations in Greece and running afoul of the improvised phonetic transcription that used ß for [s] and s for [z]. As a Swiss guy, I wouldn't have dreamt of starting a word with [z] to begin with, but then often ended up reading the s as [s] rather than [z].

Nick Shinn

@Florian Pircher

This would finally make initial swash alternatives of ß and ẞ viable ;​)

Yes, of course—I had forgotten…

Franz Gratzer

I can't recall any case where an "S" at the beginning of a word is pronounced like a "Z". At least how German speaking people would pronounce a "Z". Can you give an example? The "S" in "Sonne" (aka "sun") is pronounced exactly like in "some". What don't I get?

Christian Thalmann

Franz Gratzer said:

I can't recall any case where an "S" at the beginning of a word is pronounced like a "Z". At least how German speaking people would pronounce a "Z".

The notation [z] implies International Phonetic Alphabet (IPA), not German spelling. The IPA symbol [z] stands for the voiced alveolar fricative.

Can you give an example? The "S" in "Sonne" (aka "sun") is pronounced exactly like in "some".

That's false. The S in Sonne is voiced («stimmhaft», like the Z in English zone), whereas the S in some is unvoiced («stimmlos»). You can observe the distinction in English loan words in German, e.g. «Sex» [sɛks] vs «sechs» [zɛks].

In fact, the voicing of «s» in syllable onsets strikes me as a uniquely German thing; I can't think of another language that does that. (Then again, I don't know that many languages.)

Franz Gratzer

Christian Thalmann said:

The IPA symbol [z] stands for the voiced alveolar fricative.

Oh, allright. Sorry. This explains my error.

The S in Sonne is voiced («stimmhaft», like the Z in English zone), whereas the S in some is unvoiced («stimmlos»).

I fear I am really bad with phonetics. I am having a hard time hearing any difference at all. (There is probably a reason I failed in the language lab when I attempted to study Mandarin. I wasn't capable of hearing the difference between proper and wrong intonations.) I probably should get more careful when making assumptions based on my limited sensitivity for phonetics.

Peter Baker

Christian Thalmann said:

In fact, the voicing of «s» in syllable onsets strikes me as a uniquely German thing; I can't think of another language that does that. (Then again, I don't know that many languages.)

It was a feature of southeastern English up to the time of Shakespeare, who makes fun of it in King Lear IV.6:

Good gentleman, go your gait, and let poor volke pass. An chud ha’ bin zwaggered out of my life, ’twould not ha’ bin zo long as ’tis by a vortnight.

RichardW

I think you mean southwestern, as in Mummerset.  The shift is also to be found in Dutch, though the voiced fricatives are spelt with 'v' and 'z'.

John Savard

Scott-Martin Kosofsky said:

@John Savard, can you send a picture of the entry? I’m thinking it might be an IPA symbol (International Phonetic Alphabet, not India Pale Ale) rather than an eszett, perhaps one of the sibilant fricative symbols.

Oh, yes: it is definitely not an IPA symbol, as can be seen by the context:

Franz Gratzer said:

I would be very surprised if even one percent of the population was able to find where the captial "ẞ" letter is hidden on our keyboard layout. (It is necessary to press 3 buttons at once and it is not even mapped on the same key as the lower case letter despite this being the case for all other capital letters.)

Surely the reason for that is obvious. Capital eszet was introduced to German after 1981, and, indeed, after mid-1985, when the IBM 7531 Industrial Computer was introduced, the first device to include the Model M style of keyboard later popularized by the IBM PS/2 from 1987.

So Microsoft provides a keyboard layout that includes the capital eszet on keyboards people already have, rather than one that would require purchasing a new keyboard (or at least one or two new keycaps).

Even the Euro currency symbol suffers from the same embarassment, for the same reason.

Looking at the German keyboard, the space taken up by the three umlauts does seem to constrain the choices for making an alternate layout which gives eszet the same prominence as one of the umlauts, a key of its own with both the upper-case and lower-case versions in the unshifted and shifted positions respectively.

However, if one changes the ß? key to a /? key, with # as an alt, and ° as the shifted alt, then one can move ' to replace ° as the shifted character on the caret/degree key (with the multiplication sign as the alt) and now both shifts of the #' key become free for the eszet and its capital form. So a new German keyboard could be defined, people could buy keyboards in that style if they wish, and then select the alternate layout for them.

Franz Gratzer said:

In addition we haven't been able to find a sensible solution for gender neutral language for decades now. Useless gender specifications are woven into German so deeply that it would need a complete re-invention of our grammar to do so properly without creating ridiculously complicated constructs or versions that are only written and can not be expressed phonetically.

Of course one could eliminate gender from German, since

a) German has a neuter gender in addition to masculine and feminine, unlike French, and

b) The English language exists, proving a Germanic language can be modified to lack gender.

Of course, such radical change to the language might not qualify as a "sensible" solution.

Florian Pircher

Both Android and iOS offer ẞ on the German keyboard now.

Peter Baker

RichardW said:

I think you mean southwestern, as in Mummerset.  The shift is also to be found in Dutch, though the voiced fricatives are spelt with 'v' and 'z'.

Perhaps general southern: with "southeast" I was thinking of Dan Michel's 14th c. Kentish Ayenbite of Inwit, where the shift is represented in the spelling. I suppose the distribution of the feature would have been different by Shakespeare's time.

Interesting about Dutch.

Franz Gratzer

John Savard said:

So Microsoft provides a keyboard layout that includes the capital eszet on keyboards people already have, rather than one that would require purchasing a new keyboard (or at least one or two new keycaps). [...] Even the Euro currency symbol suffers from the same embarassment, for the same reason.

I don't even use Microsoft. But I of course understand why even on GNU/Linux we don't have a better solution yet.

However, if one changes the ß? key to a /? key, with # as an alt, and ° as the shifted alt, then one can move ' to replace ° as the shifted character on the caret/degree key (with the multiplication sign as the alt) and now both shifts of the #' key become free for the eszet and its capital form. So a new German keyboard could be defined, people could buy keyboards in that style if they wish, and then select the alternate layout for them.

That sonds like an excellent suggestion. A shame that we can't just make it the new standard without a tediously lengthy process. (Or we could get rid of the German oddities all together and instead use the English keyboard.)

Of course one could eliminate gender from German [...] such radical change to the language might not qualify as a "sensible" solution.

I don't think we didn't attempt this just because we fear we might lose the ability to understand older texts. I guess the degree of suffering wasn't big enough yet to make sensible use of the neuter gender where appropriate. If we want to go fully gender neutral we might even stop using gender specific grammar in natural persons too. This way we wouldn't need to find new pronouns for non-binary genders.

John Hudson

I think you mean southwestern, as in Mummerset.  The shift is also to be found in Dutch, though the voiced fricatives are spelt with 'v' and 'z'.

Drink ze zider.

Helmut Wollmersdorfer

I detected it in [1] Jiddisches Wörterbuch, Duden, 1992.

For original (means 'book') in YIVO orthography:

ספֿר

ßejfer -> ßojfer

Looked in my two other German-Jiddisch dictionaries.

[2] Rosten, Jiddisch - Eine kleine Enzyklopädie, Deutscher Taschenbuch Verlag (dtv), 2013, also uses ß for samekh.

ßejfer (EN-US: Seyfer, Sefer)

[3] Sigmund A. Wolf, Jiddisches Wörterbuch, Helmut Buske Verlag, Hamburg 1986, uses ss.

ssefer 

He explaines the transcription in the preface as samekh -> ss, ß and refers to Bernstein, Ignaz, Jiddisch in ..., Warschau 1908 and Landau, Alfred 1911.

[4] Solon Beinfeld, Harry Bocher, Comprehensive Yiddish-English Dictionary, Indiana University Press, 2013, uses the YIVO orthography and YIVO transcription 

[SEYFER]

[5] Peter T. Daniels (Editor), The World's Writing Systems, gives 

samekh: IPA [s], YIVO s

Searching older books:

[7] Sieben-Sprachen-Wörterbuch, 1918, https://archive.org/details/siebensprachenw00prusuoft/

Without diacritics ("unpointed" quadratic Hebrew), no transcription:



[8] Ave-Lallemant, Friedrich - Das Deutsche Gaunertum - 3. Band (1914), p. 246 https://archive.org/details/AveLallemantFriedrichDasDeutscheGaunertum3.Band1914543S.ScanFraktur/ has ss



and uses Ashkenasi Cursive (could not find a digital font for this writing style). Maybe I will reconstruct it if I find better specimen.

Many early books (15th to 17th century) printed in Germany used Rashi https://en.wikipedia.org/wiki/Vaybertaytsh, a semi-cursive Hebrew. There is a font Noto Rashi Hebrew on fonts.google.com which claims to support Yiddish but IMHO lacks some codepoints like the precomposed U+FB2B HEBREW LETTER SHIN WITH SIN DOT (Other_Letter) = sin, and U+FB4A HEBREW LETTER TAV WITH DAGESH (Other_Letter) = tof.

Helmut Wollmersdorfer

Scott-Martin Kosofsky said:

@John Savard, can you send a picture of the entry? I’m thinking it might be an IPA symbol (International Phonetic Alphabet, not India Pale Ale) rather than an eszett, perhaps one of the sibilant fricative symbols.

In a printed book I can't tell if it is an IPA symbol. In regular style it looks exactly the same (no descender) as the sharp ß used in the German words in the same book. In Italic it has a descender as the the sharp ß usually has a descender in cursives. The IPA sign for bilabial fricative is a beta (b like beta) and always has a descender.

Adam Twardoch

Franz, are you Austrian or German? 

In Austria, the voiced S tends to be less voiced than in most Germany*. I think it’s also the case in Switzerland. But in central and northern Germany, S is very clearly either »z« or (in front of some consonants) »sh«. And this is exactly the reason why the double-S was introduced, from which ß evolved. 

* Salzburg in most German-speaking regions sounds like »zalts-burg«, but in Austria it often sounds like »ssalts-burg«.

John Savard

Franz Gratzer said:

That sonds like an excellent suggestion. A shame that we can't just make it the new standard without a tediously lengthy process. (Or we could get rid of the German oddities all together and instead use the English keyboard.)

Incidentally, I want to thank you for your post, as it inspired me first to update a page on my site about keyboard arrangements which discussed a news item from 2016, where France was seeking to modify its standard for the AZERTY keyboard to support upper-case accented letters... and then to do a Google search, from which I found that in 2019, they actually completed the process, and issued the new standard, NF Z71-300.