# Why is »ß« substituted with »ss« rather than »sz«?

The letter ß is called Eszett, literally meaning s z.

However, when the letter is not available (or when a word is in all caps), ß is almost always substituted by the digraph ss rather than sz (e.g. STRASSE rather than STRASZE).

While this fact is well documented (e.g. on Wikipedia), and it’s also documented that substitution by sz used to be more common, I have never seen an explanation for why the change in rules happened, or indeed for why ss is even considered to be a valid substitution for Eszett at all.

## 6 Answers

To attack the premise of the question: What are the arguments for substituting ß with sz?

A lot of things happened to German spelling (and pronunciaton) since the appearance of the letter eszett. In particular, what once made sz the preferred choice of letters to represent what we now write as ß¹ is long gone. So, while the eszett bears the letters s and z in name, this does not exactly refer to the letters s and z as they are used by modern German spelling, but to those letters as they were used centuries ago.

With other words: Apart from the linguistic and typographical history, the relation of eszett to s and z is by name only – the same way that the letter w is uu by name only in English. Thus the name somewhat is the only argument for substituting it by sz. In the same way, one might ask, why the letter ç is not substituted by cz, since that is its origin.

In modern German, the combination sz is only used in a few loanwords (e.g., Disziplin, Szene or lasziv), where it represents different sounds than the letter ß does (see also Toscho’s answer). On the other hand, the ss always represents the same sounds as ß, and only differs by indicating a different length of the preceding vowel: ß is preceded by long vowels, ss by short ones. (In the old spelling (before 1996), the difference is a little bit more complicated, but the argument is essentially the same.) Therefore, ss is the substitution that is closest to German spelling conventions.

¹ and what eventually gave rise to the sz ligature (more precisely: ſz), which rose to letter status and on which eventually the modern ß was based

• “Szene” is actually an interesting example since its original pronunciation (the one I learned in school) uses the same sound as “ß”. For whatever reason, it has very recently changed to the /s-z/ pronunciation. – Konrad Rudolph Jun 24 '14 at 11:45
• @KonradRudolph If you go back to the Greeks it's /sk/. – ALLCAPSWARLOCK Nov 30 '15 at 13:09
• “aspects of German spelling due to which it made sense to use sz where we now use ß are long gone.” I tried to show in my answer that this is not, in fact, the case. The letter s is used for initial /z/ and final /s/ in German, because they are distributed complementarily (if initial /ts/ is exempt). Alas, in medial position, i.e. between vowels, both /s/ and /z/ occur (also /ṣ/ and /ẓ/, Eisenberg says) – and /ts/ etc. Single s is reserved for /z/ here, doubling implies short vowel. All other voicing pairs (soft/hard) use separate letters, e.g. d/t, so it’s not an issue elsewhere. – Crissov Dec 2 '15 at 9:12
• @Crissov: (Hoping that I understand you correctly:) I meant: “aspects of German spelling due to which it made sense to use sz and not some other digraph where we now use ß are long gone.” I do not dispute that there was a need to have some distinct representation form where we now use ß. – Wrzlprmft Dec 2 '15 at 20:38
• There are not that many digraphs possible graphotactically, though. ſſ or ss should work the same as tt, nn, pp etc., so they’re inappropriate after long vowels or diphthongs. ſz made just as much sense as tz (which could have been but not cz instead). cſ works similar to ck and ch, and ſc could have been possible as well, but cs and sc are missing the desirable ascender – well, so does sz which could switch to a descender . zh is associated with /ʒ/ nowadays, but (like zch/ʒch) would have been less confusing than sh (considering sch/ſch). – Crissov Jan 25 '17 at 22:15

I’ve found some interesting explainations on Wikipedia.

• Origin of the various substitutes.

In the late 18th and early 19th century, when more and more German texts were printed in Roman type, typesetters looked for a Roman counterpart for the blackletter ſz ligature, which did not exist in Roman fonts. Printers experimented with various techniques, mostly replacing blackletter ß in Roman type with either sz, ss, ſs, or some combination of these.

• Why SZ could (or have been) be used instead of SS.

The traditional orthography encouraged the use of SZ in place of ß in words with all letters capitalized where a usual SS would produce an ambiguous result. One possible ambiguity was between IN MASZEN (in limited amounts; Maß, “measure”) and IN MASSEN (in massive amounts; Masse, “mass”). Such cases were rare enough that this rule was officially abandoned in the reformed orthography. The German military still occasionally uses the capitalized SZ, even without any possible ambiguity

• Since I studied some Hungarian pronounciation, I remembered that they use to read SZ as the german Eszett sound. Here’s another quote from the link I mentioned before:

If no ß is available, ss or sz is used instead (sz especially in Hungarian-influenced eastern Austria).

I’m not German but this is what I actually found. In my opinion, as someone said, the ss way has just evolved out of ignorance. Probably at some point for some reason (and this was you question) people started to use SS instead of SZ and they became used to it. Maybe because is nicer to see or faster to write?

I would add just that, when something for some reason becomes a convention, it isn’t anymore important if it’s wrong or not. Languages are constantly changing and the aim is not to be correct (i.e. to follow traditional rules) at speaking but to communicate in the most efficient way and understand each other. Otherwise I should start to speak Latin and want Italian people to understeand me.

• well... I thought it was interesting, why this down vote? – Cindie Nov 30 '15 at 8:00
• I dislike the sentence evolved out of ignorance but otherwise +1. – Jan Nov 30 '15 at 12:22
• The part that addresses the actual question starts with "in my opinion" and contains only that. – Carsten S Nov 30 '15 at 13:58
• In particular, the asker seems to be aware of the historical background: “While this fact is well documented (e.g. on Wikipedia), and it’s also documented that substitution by sz used to be more common, I have never seen an explanation for why the change in rules happened, […]” – Wrzlprmft Dec 2 '15 at 20:40

If you read a capitalized text, you don't expect to substitute letters in your head: you just read the letters.

If you see -SS-, you read it in the same way as -ß-: as a voiceless s [s].

If you see -SZ-, you read it as a combination of -s- and -z-: [sts].

So STRASZE wouldn't be read as it should and this confuses. The SS inSTRASSE is read in the same way as it should be. Alas, the preceding vowel is could be read differently (short instead of long).

• any idea as to the "why"? I think that was the actual question – Emanuel Feb 12 '14 at 11:48
• Actually if you go by "Die neue Rechtschreibung" you would never read Strasse the same as Straße. When we read old books (Goethe, Schiller, Kafka) in school I often stumpled on words which I usually see with a ß. The same counts for the opposite. When I see Schloß, I won't read it Schloss, it will sound like Kloß. It's really a big difference if you're not used to the Alte Rechtschreibung or if you want to pronounce words clearly. – Leo Pflug Feb 12 '14 at 13:50
• @LeoPflug That is indeed correct. The preceding vowel is read differently. The consonant (group) itself is read in the same way. – Toscho Feb 12 '14 at 18:22
• This argument does not hold water, as you do not "just read the letters" in German (or really any language for that matter). By your logic, it should be spelled "SCHTRASSE". Actually, no, even that is nonsense, as SCHT would have to be pronounced /st͡sht/ and not as /ʃt/. In short, you are perfectly fine with "substituting letters in your head" elsewhere, including elsewhere in this very word. You have happily accepted whatever conventions you were presented with. If the convention were to replace ß with LMP, then you would be just as perfectly happy with writing and reading "STRALMPE". – RegDwight Feb 13 '14 at 22:03

Blame the Swiss.

The ⟨ß⟩ is the result of two spelling conventions converging. I’ll describe the more prominent one first:

1. At least in blackletter type, word-final s always took the round form ⟨s⟩ whereas pretty much everywhere else the lowercase letter retained the default long form ⟨ſ⟩.
German orthography has quite systematic rules for doubling of consonant letters, so double s should work by the same principles as ⟨ff⟩, ⟨ll⟩, ⟨mm⟩, ⟨nn⟩, ⟨pp⟩, ⟨rr⟩ and ⟨tt⟩. All of these can only appear at the end of a word or as the only letters between vowels, whereas (voiced/soft) ⟨bb⟩, ⟨dd⟩ and ⟨gg⟩ require the vowel following them. Together with typesetter’s rules, we got ⟨ſſ⟩ in word-medial and ⟨ſs⟩ in final position. For the latter, it was common in calligraphy and typography to form a ligature. This tradition survived when Fraktur etc. were abandoned, because it had come to be treated as an orthographic rule, although it really was a typographic rule.
This convention was finally abolished in the 1996 orthographic reform. It was the one that originally informed the substitution of ⟨ß⟩ by ⟨ss⟩.
• Special treatment is not unique to s, by the way: ⟨ck⟩ is a special way of writing *⟨kk⟩ and ⟨tz⟩ is used instead of *⟨zz⟩. Some may go further and consider ⟨ch⟩ a replacement of *⟨hh⟩ (e.g. in sehen > Sicht), but unlike all other double consonants it can be separated from the preceding vowel by an ⟨r⟩, ⟨l⟩ or ⟨n⟩ (because it replaced a former single ⟨k⟩).
2. A solution was needed to represent /s/, because single ⟨ſ⟩ (or ⟨s⟩) was already used for /z/ in initial and medial position and a double consonant like ⟨ſſ⟩ (or ⟨ss⟩) would have regular, but unwanted implications.
Single ⟨z⟩ (usually looking more like ⟨ʒ⟩, especially in final position) could have been employed for this purpose, although it represents /ts/ in initial (onset) position and after other consonants in codas, because ⟨tz⟩ (or ⟨tʒ⟩) in native and ⟨c⟩ in loan words was employed for that sound in medial and final position – but it wasn’t and when ⟨z⟩ replaced Romance ⟨c⟩ that alternative was gone. Another possibility might have been to indicate altered vowel quality in a different way, e.g. by adding an /h/, but firstly that would have to be dropped sometimes (e.g. essen, aßen *⟨ahsen⟩, *⟨aassen⟩; äsen; aasen) and secondly double consonants generally follow a simple vowel without interruption.
The digraph that evolved instead was “Eszett” ⟨ſʒ⟩ (in classic shapes). Grimm’s Deutsches Wörterbuch and Bundeswehr telex cables always used ⟨sz⟩ in place of ⟨ß⟩, for instance.
It survived in rare cases of adopted initial /s/ where it replaced former ⟨ſc/sc⟩. The digraph had long been replaced by ⟨ß⟩ when this phonographic rule was kept in the recent reforms. However, not least due to multilingual mechanical typewriters, Switzerland and Liechtenstein had abolished ⟨ß⟩ decades earlier – in favor of ⟨ss⟩ in all cases.
• In simple onsets, German natively only supports /ʃ/ and /z/, not /s/. In codas, it has only /ʃ/ and /s/ (due to “Auslautverhärtung”). Complex onsets require /ʃ/ except for /ts/ which is the only native one that has another consonant preceding an S sound.

In conclusion, where the ligature has lost orthographic status almost 20 years ago, it was rightfully replaced by ⟨ss⟩, because the long form isn’t in use any more. That part was compatible with what the Swiss were already doing anyway. The remaining uses of ⟨ß⟩ really should be substituted by ⟨sz⟩ where necessary! Alas, that was not often required to be done in the 20th century by enough writers, and many people already got the main ss/ß rules wrong too often (because the typographic rule wasn’t well understood once its use wasn’t obvious any more with the demise of ⟨ſ⟩), so it’s not a big surprise the non-mandatory, vague recommendations for replacement were simplified in practice.

The Swiss probably ditched the ⟨sz⟩ replacement because having a single variant was perceived as simpler and a double letter was the regular choice for the more frequently occurring case while creating hardly any (semantic) ambiguities for the less important case. The reformers chose not to abolish ⟨ß⟩ completely (as had been proposed by some), but to promote it to a proper (lowercase-only) letter. It’s a pragmatic choice, even though a bad call by linguistic purity, to keep the fallback compatible with widely established practice within the largest affected user base, because most writers would have the ⟨ß⟩ available anyway.

I’m pretty sure that the reason was not an unfamiliar appearance of the digraph ⟨sz⟩, because it already appears (alongside stranger things) in loan words and ⟨tz⟩ works in much the same way. One could argue that ⟨tz⟩ should be hyphenated as ⟨z-z⟩, like regular double consonant letters and ⟨ck⟩ previously, which became ⟨k-k⟩, but ⟨sz⟩ would still have to be broken into ⟨s-s⟩, which would better match a *⟨cs⟩ digraph.

Corollary: Although the shape of lowercase ⟨ß⟩ builds upon the ⟨ſs⟩ ligature in most typefaces, a newly designed uppercase variant ⟨ẞ⟩ – often an ugly ⟨B⟩ lookalike – may well be based loosely upon an ⟨SZ⟩, ⟨Sz⟩, ⟨Sʒ⟩, ⟨ſz⟩, ⟨ſʒ⟩ or even ⟨SƷ⟩, ⟨ſƷ⟩ or ⟨ſZ⟩ ligature.

• I have a few comments on this: 1) At least in blackletter type, word-final s always took the round form ⟨s⟩ whereas everywhere else the lowercase letter retained the default long form ⟨ſ⟩. – That’s not exactly correct, there are spellings like Maske. However, it does not matter for your further argument. – Wrzlprmft Dec 2 '15 at 6:31
• 2) It can still be seen in rare cases of adopted initial /s/, i.e. basically just ⟨Szene⟩ and its derivates.Szene is pronounced with /ts/ or even /sts/ rather than /s/ in the beginning. Also, using ⟨ſz⟩ (not ⟨ß⟩) only came up during the reform of 1901. Before, it was Scene, lasciv or Disciplin. – Wrzlprmft Dec 2 '15 at 6:34
• 3) Although the shape of lowercase ⟨ß⟩ builds upon the ⟨ſs⟩ ligature in most typefaces, a newly designed uppercase variant ⟨ẞ⟩ – often an ugly ⟨B⟩ lookalike – may well be based loosely upon the ⟨SZ⟩ digraph, e.g. ⟨Ƨ⟩, ⟨Ʒ⟩, ⟨Ʃ⟩, ⟨Ƽ⟩, ⟨S̷⟩. – Actually, modern capital ß designs avoid the SZ digraph forms (as they would diminish readability) in favour of forms keeping elements from the modern lowercase ß. It’s older capital ß designs (mostly from 1900–1950) that use SZ digraph forms. – Wrzlprmft Dec 2 '15 at 6:38
• 4) I fail to see your main line of thought. You are describing a lot of things, but you do not make a clear connection to the actual question. Also, it’s often not clear about what time you are roughly talking, when talking about past events. – Wrzlprmft Dec 2 '15 at 6:43
• 5) because the typographic rule wasn’t well understood once its use wasn’t obvious any more with the demise of ⟨ſ⟩ – similar to ⟨st⟩ hyphenation restrictions – Actually, the st hyphenation restrictions did not even make sense with the long s (ſ), at least in my opinion. – Wrzlprmft Dec 2 '15 at 6:45

Well I’m German native speaker and I don’t see any evidence for the assumption in the original post that it became more and more popular to use SZ in the meaning of a capitalized ß. As correctly put by Wrzlprmft the naming of the letter has historical meanings originating from a century where a different script than the nowadays used Roman font was in use for German language. Reading old German writers (Goethe, Schiller, Kafka were brought forward in other comments) you will also find a lot usages of the letter y in words that are now mostly written with i.

The usage of SZ is simply incorrect (perhaps apart of some not representative regions in eastern Austria/western Hungaria) and reading the word MUSZE would take me more than 2 seconds to correctly identify it as Muße.

One of the reasons is that people usually read words by syllables and there is a difference between SS and SZ in the separation of syllables as SS is never (!) separated into different syllables while SZ is (S-Z). Thus I would read MUS-ZE which simply doesn’t maps to any known German word in my mind in comparison to MU-SSE, which does.

• Ich weiß nicht, ob ich deinem Argument richtig folge; meinst du, dass geschriebenes MUSSE nicht als MUS-SE getrennt werden darf, oder meinst du, dass du bei silbendeutlichem Sprechen Mu-ße sprichst, und nicht etwa Muß-ße? – Jan Nov 30 '15 at 12:20
• “The usage of SZ is simply incorrect” — Such prescriptivist dogmatism is rarely useful. The alte Rechtschreibung recognised “SZ” as an entirely correct (albeit unusual) spelling of capital-ß. The neue Rechtschreibung leaves this rule out. That doesn’t make its use incorrect though, it just removes it from the amtliche Rechtschreibung. – Konrad Rudolph Nov 30 '15 at 13:16
• Furthermore, I’m not sure why you are doubting the question’s premise: it’s well established that substitution of “ß” with “SZ” used to be common. In fact, it was the rule in the 1901 Rechtschreibung. Substitution with “SS” came later. – Konrad Rudolph Nov 30 '15 at 13:22
• Digraphs have to be learned, but since ⟨ß⟩ is usually available, especially in handwriting, there was hardly a need for anyone to get used to ⟨sz⟩ like they did for ⟨ng⟩, ⟨ck⟩, ⟨ch⟩, ⟨sch⟩, ⟨tz⟩, ⟨sp⟩, ⟨st⟩, ⟨ph⟩, ⟨ie⟩, ⟨ei⟩, ⟨au⟩, ⟨eu⟩ etc. Instead, the Swiss got used to ⟨ss⟩ (sometimes) working differently than all the other doubled consonant letters (which are also digraphs, actually), and now other German speakers have to do the same (some already did). – Crissov Dec 2 '15 at 9:24

The letter ß is "called" S-Z because, in the old German Fraktur typeface, it was literally an S-Z, as follows:

ſȝ

You can easily verify this in any old German text written in Fraktur. As you can see, the first letter is a "long-S", ſ, in use in all European languages until the 17th-18th centuries and in German (Fraktur) until the 20th century, used for initial and middle (but not final) S. The second letter is one of the forms of "z", ȝ, which is still very common in all languages.

In the modern Roman German alphabet, the equivalent of this ligature is ſs. The typographers who adapted German to the Roman type decided to replace the second "z" with the round-s "s". Fraktur also had a round-S, used at the end of syllables. But Fraktur did not use it for this ligature.

Wikipedia says that conferences in 1879-1903 settled on ſs as the Roman equivalent of Fraktur's ſȝ. Somehow the "z" didn't work out. But nonetheless, if you look at this picture, sometimes even the modern Roman font of German continues to use the old ſȝ. So it's not completely true that in modern German the old ſȝ variant (the "real" ESS-ZETT) has gone away.