Double vowels in German orthography

[alice]

Is there any particular reason for them, and how many words actually have them? Off the top of my head I can only think of:

- Paar, Saar, Haar
- See, Idee, Meer (?), Schnee
- Boot

ON another matter, See and Schnee have cognates in Gothic saiws and snaiws, but then why Möwe for maiws? Is this by analogy with Löwe?

[Jipí]

Off the top of my head I can only think of:

- Paar, Saar, Haar
- See, Idee, Meer, Schnee
- Boot

Yes, it's only a few words that must be learned. I have never heard an explanation as for why those words are written with double vowels instead of Vh. Compare their MHG forms, if that may be any help (not all words are borrowed yet):

- pâr, ?, hâr
- sê, -/- (if this was already /ide:/ in OF I'd expect *îdê/idê if it was borrowed. BMZ doesn't attest it, though.), mer, snê
- -/-

You forgot Saal 'hall, (assembly, ball) room', MHG sal; and Saat 'seed', MHG sât.

but then why Möwe for maiws? Is this by analogy with Löwe?

Duden says the etymology of Möwe (MLG mêwe, Du. meeuw, OE mǣw, ON már) is not clear, but it's most likely onomatopoeia (Fris. meau, mieu). The word was spelled Mewe up into the 18th century.

--
EDIT: More words with double vowels are Seele 'soul', MHG sêle; and leer 'empty', MHG lære (< OHG lâri; æ = ê from umlaut)

[alice]

Ah, danke schön. It's nice to discover a genuine mystery.

You forgot Saal 'hall, (assembly, ball) room', MHG sal; and Saat 'seed', MHG sât.

I've never actually used those words in German, either written or spoken, which probably explains why All the others, yes, but not these.

[hwhatting]

If we don't count loanwords that have double vowels in the original orthogtaphy (like Lee "leewards", French loans like Matinee etc.), there's still a few that haven't been listed: Moor "moor", Seele "soul", Aar "eagle" (literary/obsolescent), Aas "carrion", Fee "fairy", Reede "roadstead", Teer "tar", Tee "tea", Kaffee "coffee", Beet "(flower)bed". Then there's Maar "lake in the crater of an extiguished volcano" which, AFAIK, is limited to such lakes in the Eifel region. Double vowels are more frequent in geographical and family names. German orthography is a mish-mash of distinct regional orthographical traditions, where the central German traditions won out against Northern and Southern traditions; IIRC, double-vowels for length is more typical for the Northern and western traditions, which may explain why it is rarely used.

[linguoboy]

If we don't count loanwords that have double vowels in the original orthogtaphy (like Lee "leewards", French loans like Matinee etc.), there's still a few that haven't been listed: Moor "moor", Seele "soul", Aar "eagle" (literary/obsolescent), Aas "carrion", Fee "fairy", Reede "roadstead", Teer "tar", Tee "tea", Kaffee "coffee", Beet "(flower)bed".

Fee doesn't belong there, as it represents French fée.

As for Möwe, my understanding is that this is a relatively recent loan from Low Saxon, which has retained the /v/. (Cf. Modern Northern Low Saxon Meev.) Though I suppose that only begs the question, since NLS in turn has See and Snee, not *Seev and *Sneev.

[Qwynegold]

Seele Schneider!

[hwhatting]

Fee doesn't belong there, as it represents French fée.

You're right. It doesn't feel like a loanword, but that doesn't change the fact.

As for Möwe, my understanding is that this is a relatively recent loan from Low Saxon, which has retained the /v/. (Cf. Modern Northern Low Saxon Meev.) Though I suppose that only begs the question, since NLS in turn has See and Snee, not *Seev and *Sneev.

I assume that's because it must have been me:v@ until at least the late middle ages or early modern times, i.e., must have kept a schwa word-final that protected the /w/. That would be supported by the HG form with schwa and by Polish mewa, which is a loan from LG These forms alone could be just adaptation of /me:v/ to HG and Polish declensional types, but together with the preservation of final /w/ they seem to point to a MLG (or thereabouts) meve.

And we forgot one more word: Aal "eel".

[Skomakar'n]

Slightly off topic, I just want to say that I'm bothered by only certain clusters with /ʃ/ being etymologically incorrectly spelled with ‹sch›. Why isn't it either universally applied, or not applied at all?

[Jipí]

You mean like in /St/ and /Sp/, which never get spelled <schp> and <scht>?

[jal]

You mean like in /St/ and /Sp/, which never get spelled <schp> and <scht>?

I think he means the opposite, sometimes /S/ is spelled <sch> while not coming from /sx/(if that's where it comes from)?


JAL

[MisterBernie]

I think he means the opposite, sometimes /S/ is spelled <sch> while not coming from /sx/(if that's where it comes from)?

It does. The first source of Standard German [ʃ] is [sk] > [sx] > [ʃ] everywhere. The second source is [ˈsCV] > [ʃCV] (plus some, but not all [rs] > [rʃ]), in which case the spelling changed for all <sC> clusters except for initial <st> and <sp>. As for the why, I don't know either.

[linguoboy]

You mean like in /St/ and /Sp/, which never get spelled <schp> and <scht>?

I think he means the opposite, sometimes /S/ is spelled <sch> while not coming from /sx/(if that's where it comes from)?

That seems less plausible to me than Guitarplayer's suggestion.

I think it comes down to predictability: orthographic <s> is always /ʃ/ initially before stops, regardless of the origins of the word. Spelling it <sch> here would simply be redundant.

Conveniently, however, this suits the pandialectal nature of the standard orthography. There are northerners with /s/ in this position. (They "stolpern über den spitzen Stein", as the popular saying has it.) And then there are southwesterners with /ʃ/ before stops within the same morpheme in every position. (Unfortunately I don't know any snappy saying for them, only stale jokes playing on the homophony of hast du? "have you?" and hascht du "do you smoke hash?" in Alemannic.)

[Jipí]

Yeah, but what about other initial s+C not turned into ʃ+C in the North? I know pathetically little about German dialects, I'm afraid.

[linguoboy]

Yeah, but what about other initial s+C not turned into ʃ+C in the North? I know pathetically little about German dialects, I'm afraid.

Examples? Unless you're talking about actual Platt instead of accented Hochdeutsch, I can't think of any.

[MisterBernie]

I think it comes down to predictability: orthographic <s> is always /ʃ/ initially before stops, regardless of the origins of the word. Spelling it <sch> here would simply be redundant.

Well, it would be predictable if it weren't for recent loans. If that rule applies, it's only for the stops /p/ and /t/ (<sk> is /sk/, e.g. <Skala>, <Skelett>, <Skepsis>), but even then, for words like <Star, Stereo, Stola> both [st] and [ʃt] exist, and it remains to be seen whether they'll be fully integrated into German phonology.

[Skomakar'n]

You mean like in /St/ and /Sp/, which never get spelled <schp> and <scht>?

Yeah. Neither ‹Stein› nor ‹Schnee› etymologically have anything but /s/, but are both now pronounced with /S/, so why is only one of them spelled with the non-etymological ‹sch›? >.< It's such an ugly inconsistency.

[linguoboy]

You mean like in /St/ and /Sp/, which never get spelled <schp> and <scht>?

Yeah. Neither ‹Stein› nor ‹Schnee› etymologically have anything but /s/, but are both now pronounced with /S/, so why is only one of them spelled with the non-etymological ‹sch›? >.

Because the rule is historical before /n/ but not before /t/. As MisterBernie points out, there is variation and the rule could be on its way out for /st/ and /sp/ clusters as well. But as it currently stands, it's acceptable to say Stereo with [ʃ] (that's how I do say it, in fact) but using this for the first s in Snackshop strikes me as very odd indeed.

Moreover, given the relative proportion of words with initial /ʃt/ to those with /st/, it would make more sense to me to spell the latter with ßt than to use scht for the former. Much more parsimonious.

[Skomakar'n]

You mean like in /St/ and /Sp/, which never get spelled <schp> and <scht>?

Yeah. Neither ‹Stein› nor ‹Schnee› etymologically have anything but /s/, but are both now pronounced with /S/, so why is only one of them spelled with the non-etymological ‹sch›? >.

Because the rule is historical before /n/ but not before /t/. As MisterBernie points out, there is variation and the rule could be on its way out for /st/ and /sp/ clusters as well. But as it currently stands, it's acceptable to say Stereo with [ʃ] (that's how I do say it, in fact) but using this for the first s in Snackshop strikes me as very odd indeed.

Moreover, given the relative proportion of words with initial /ʃt/ to those with /st/, it would make more sense to me to spell the latter with ßt than to use scht for the former. Much more parsimonious.

Does every dialect have /S/ in Schnee? If not, I'd find it much nicer to have ‹s› in all of these words, and just have the /Sn/ clusters (and similar things) be irregular exceptions, since not everyone has /sn/ anyway (if that's the case).

[linguoboy]

Does every dialect have /S/ in Schnee? If not, I'd find it much nicer to have ‹s› in all of these words, and just have the /Sn/ clusters (and similar things) be irregular exceptions, since not everyone has /sn/ anyway (if that's the case).

The only varieties I can think of which don't are considered dialects of Low Saxon, not of High German. The isogloss in question is the Panninger Line.

[Jipí]

Neither ‹Stein› nor ‹Schnee› etymologically have anything but /s/, but are both now pronounced with /S/, so why is only one of them spelled with the non-etymological ‹sch›? >.< It's such an ugly inconsistency.

The second [source of Standard German [ʃ]] is [ˈsCV] > [ʃCV]

It's actually very regular a change, at least if you look at Middle High German > New High German, AFAICT.

[linguoboy]

The second [source of Standard German [ʃ]] is [ˈsCV] > [ʃCV]

It's actually very regular a change, at least if you look at Middle High German > New High German, AFAICT.

Just noticed this: why "['sCV]"? That implies that it doesn't happen in the clusters /sCL/, which is clearly erroneous. As far as I can tell, the rule isn't the least bit sensitive to what follows the initial /sC/ cluster. As Guitarplayer says, it's a very regular change.

[MisterBernie]

It's actually very regular a change, at least if you look at Middle High German > New High German, AFAICT.

A linguist agrees with you.

So, to sum up:
[s] – <s>
Suhle, Sand, Saal
[ʃ] – <sch>
Schule, Schande, Schaal
[ʃP] – <sP>
Stuhl, Stand, Stahl
Spule, spannt, Spalier
[ʃP̄] – <schP̄>
Schlinge, schlecken, schlimm
schmal, schmecken, schmiergeln
schnell, Schnecken, Schnur
schwul, schwand, Schwall

+
[sk] – <sk>
skandieren, Skala
+
[s/ʃP] - <sP>
Stereo, Star, Stola, ?Stalin, Spaghetti*

* IME, <Star> celebrity has [staːɐ̯] as the preferred pronunciation, whereas for <Spaghetti>, [spaˈɡɛti] would be perceived as snobby.

[MisterBernie]

Just noticed this: why "['sCV]"? That implies that it doesn't happen in the clusters /sCL/, which is clearly erroneous. As far as I can tell, the rule isn't the least bit sensitive to what follows the initial /sC/ cluster. As Guitarplayer says, it's a very regular change.

Yeah, sorry, the V is superfluous (or should've been at least ['sC(L)V]), I put it in as a sort of signifier for "initial cluster in a stressed syllable" to differentiate between /sC/- and /-sC/ in Standard German.

[linguoboy]

Yeah, sorry, the V is superfluous (or should've been at least ['sC(L)V]), I put it in as a sort of signifier for "initial cluster in a stressed syllable" to differentiate between /sC/- and /-sC/ in Standard German.

Isn't that why we have a symbol for juncture (#)? Or does that make you feel like you're proposing a phoneme of dubious status?

On second thought, I suppose that notation would run afoul of the retention of /ʃ/ in inflected and derived forms, e.g. bestellen vs. am besten.

[MisterBernie]

Isn't that why we have a symbol for juncture (#)? Or does that make you feel like you're proposing a phoneme of dubious status?

I'm not that used to writing down sound changes, hence why I didn't think of #.

On second thought, I suppose that notation would run afoul of the retention of /ʃ/ in inflected and derived forms, e.g. bestellen vs. am besten.

This why I included the [ˈ], to say 's in a cluster at the onset of a stressed syllable' (for Standard German, I'm well aware that sC > ʃC applies elsewhere, too, in southern varieties).