Voiced fricatives in Germanic

[So Haleza Grise]

When did a voicing distinction develop in fricatives in Germanic? Look at cognates like English see compared to German sehen. I understand that the rule in (standard?) German is that if the initial <s> is followed by a vowel, then it's voiced. Phonetically, this makes sense. Originally (I understand), in the ancestral language, /s/ and /z/ were allophones. I don't know enough about German phonology to say if German ever developed a distinction between /s/ and /z/ in native vocabulary, or if a /z/ phoneme was just imported from elsewhere.

But why, then does English have initial /s-/ instead of /z-/ in see? Should we assume that [s-] was the original pronunciation in Proto-Germanic or whatever, and then German applied a rule to voice these initials, after the breakup of West Germanic?

A related question is German <st->. Originally (again, I understand), there was no distinction in the ancestral language between /s/ and /ʃ/. Can we assume that in the ancestral language, <st-> (for example in stone) was always [st-] and there wasn't allophonic variation between /s/ and /ʃ/ in this position? Or is it the case that in the ancestral language there probably was variation between [ʃt-] and [st-] here, and both English and German have simplified to one of the two, but picked different options?

[Moose-tache]

The voicing of intervocal single fricatives shows up in enough Germanic languages that it may go back to proto-Germanic, so the German innovation is that initial prevocalic fricatives follow the same pattern. This is an areal change that swept across much of West Germanic, including some English varieties, specifically Yola. The change probably happened in the middle ages.

As for the backing of German s to sh, this may be a knock-on effect of the High German Consonant Shift. Original /t/ becomes /(t)s/ in Standard German, with words like wasser being pronounced with a normal, family-friendly /s/. But when the HGCS began, the sibilants created from /t/ were distinct from the existing sibilants in their point of articulation. Most likely, the former was apical and the latter laminal. A similar situation occurred in Old Spanish, where the sibilants from palatized /t/ in words like plaza were distinct from the sibilants in words like escola. It's very likely that original s in German moved further toward being alveopalatal in an attempt to keep it distinct from the new s-from-t. Again, it's possible something like that happened in some dialects of Spanish. For example, most Spaniards transcribed Native American sh with x, but a few used s, such as DeSoto, who transcribed Native s and sh as z and s, respectively.

[Estav]

As far as I know, Gothic shows no signs of voicing Proto-Germanic *s *ɸ *θ *x *xʷ. Instead, in Gothic [β ð ɣ] functioned as allophones of /b d g/, and Proto-Germanic *z was maintained as a distinct phoneme from *s (setting aside neutralization in word-final position or before an obstruent). So I'm confused by the idea that intervocalic voicing of fricatives could "go back to proto-Germanic". (Verner's law was a type of conditioned voicing that occurred before Proto-Germanic.)

It has been argued that there was variation between dialects of Old English in the voicing of word-initial fricatives.

As in English, German has developed a distinction in native vocabulary between /s/ and /z/ in intervocalic position as a result of simplification of historical geminates.

IPA /s/ isn't a precise sound, it's a broad phonemic transcription. As Moosetache said, there is speculation about the precise phonetic quality of /s/ in German and other languages historically changing based on whether the language had or didn't have a contrast with other sibilant fricatives.

[Linguoboy]

This is an areal change that swept across much of West Germanic, including some English varieties, specifically Yola. The change probably happened in the middle ages.

In particular, compare Dutch zien, Limburgs zeen (but also West Frisian sjen, Low Saxon sehn).

A couple things worth mentioning: One is that the voiced pronunciation is not universal in initial position. This was originally a dialect feature that became incorporated into the formal pronunciation standard. It is completely absent in Upper German varieties.

Moreover, the pronunciation of initial /s/ before consonants varies too. There's even a saying in German, "Sie stolpern über den spitzen Stein" ("They stumble over the pointy stone"), which refers to the tendency of some Northerners to use [s] before /t/ and /p/. Conversely, one of the most salient features of Alemannic is the use of [ʃ] not just in initial position before consonants but also medially and finally. Note, however, that there is a contrast here between the /s/ inherited from Common Germanic and the /s/ which is the result of the affrication (and subsequent deaffrication and degemination) of *t, e.g.:

Isch(d) är? "Is he?"
Issd är? "Does he eat?"

More evidence for a historical aveolopalatal pronunciation of Middle Upper German /s/ comes from Hungarian, which adopted <s> to represent /ʃ/ and <ß> (now written <sz>) to represent /s/ (e.g. Hungarian szatmári sváb [ˈsɒtmaːri ˈʃvɒb] "Sathmar Swabian").

[Travis B.]

This is an areal change that swept across much of West Germanic, including some English varieties, specifically Yola. The change probably happened in the middle ages.

In particular, compare Dutch zien, Limburgs zeen (but also West Frisian sjen, Low Saxon sehn).

IIRC West Frisian has voiceless initial fricatives, not voiced ones, and northern Dutch despite Dutch orthography also has voiceless initial fricatives under Frisian substratum influence. I might be wrong here though.

[alice]

To confuse things further, I've seen initial /z/ and /v/ in Dutch transcribed with the under-ring diacritic for devoicing, which clearly indicates something different from /s/ amd /f/, but I've never understood what.

[Travis B.]

To confuse things further, I've seen initial /z/ and /v/ in Dutch transcribed with the under-ring diacritic for devoicing, which clearly indicates something different from /s/ amd /f/, but I've never understood what.

From hearing sung Afrikaans, orthographic initial <s>, corresponding to orthographic Dutch <z>, sounds clearly voiceless to my ears, which corresponds to my understanding that Afrikaans is primarily derived from northern Dutch. (Also note that Afrikaans initial <v> is also voiceless, and contrasts with Afrikaans initial <w>, which sounds like [v] to my ears.)

[alice]

To confuse things further, I've seen initial /z/ and /v/ in Dutch transcribed with the under-ring diacritic for devoicing, which clearly indicates something different from /s/ amd /f/, but I've never understood what.

From hearing sung Afrikaans, orthographic initial <s>, corresponding to orthographic Dutch <z>, sounds clearly voiceless to my ears, which corresponds to my understanding that Afrikaans is primarily derived from northern Dutch. (Also note that Afrikaans initial <v> is also voiceless, and contrasts with Afrikaans initial <w>, which sounds like [v] to my ears.)

In other words, Afrikaans initial <v> and <w> represent /f/ and /v/?

[Travis B.]

To confuse things further, I've seen initial /z/ and /v/ in Dutch transcribed with the under-ring diacritic for devoicing, which clearly indicates something different from /s/ amd /f/, but I've never understood what.

From hearing sung Afrikaans, orthographic initial <s>, corresponding to orthographic Dutch <z>, sounds clearly voiceless to my ears, which corresponds to my understanding that Afrikaans is primarily derived from northern Dutch. (Also note that Afrikaans initial <v> is also voiceless, and contrasts with Afrikaans initial <w>, which sounds like [v] to my ears.)

In other words, Afrikaans initial <v> and <w> represent /f/ and /v/?

That's what it sounded like to my ears.

Edit: The wiki says that Afrikaans initial <v> is /f/ except in some French and Latin loanwords, where it is /v/, and Afrikaans <w> is /v/ except in some clusters like <kw> where it may be /w/, but at the same time Afrikaans /v/ is considered by some to actually be [ʋ].

[Sol717]

It has been argued that there was variation between dialects of Old English in the voicing of word-initial fricatives.

I am personally sympathetic to this belief, but the mainstream view is that initial fricative voicing in the dialects of Southern England and the the south-western Midlands only occured in Middle English, as it is only then it appears in the orthography; compare Anglian OE scep, fleoman, swin [ˈʃ(t͡ʃ)eːp ˈfleo̯ːmɑn ˈswiːn] to Kentish ME ssep, vleme, zuin [ˈʒeːp ˈvleːmə ˈzwiːn]; c.f. London schep, fleme(n), swope(n) [ˈʃeːp ˈfleːmə(n) ˈswiːn] "sheep, to flee, pig". As London is adjacent to the area affected (and London English may have originally posessed it before immigration from more northerly parts), some traces of this survives in Standard English (vat, vane < OE fæt, fanu). This initial voicing retreated to the West Country due to the heavy influence of the London standard during the Early Modern English stage; it is now moribund even there, hence the stereotyped bucolic "Zummerzet".

[So Haleza Grise]

It has been argued that there was variation between dialects of Old English in the voicing of word-initial fricatives.

I am personally sympathetic to this belief, but the mainstream view is that initial fricative voicing in the dialects of Southern England and the the south-western Midlands only occured in Middle English, as it is only then it appears in the orthography; compare Anglian OE scep, fleoman, swin [ˈʃ(t͡ʃ)eːp ˈfleo̯ːmɑn ˈswiːn] to Kentish ME ssep, vleme, zuin [ˈʒeːp ˈvleːmə ˈzwiːn]; c.f. London schep, fleme(n), swope(n) [ˈʃeːp ˈfleːmə(n) ˈswiːn] "sheep, to flee, pig". As London is adjacent to the area affected (and London English may have originally posessed it before immigration from more northerly parts), some traces of this survives in Standard English (vat, vane < OE fæt, fanu). This initial voicing retreated to the West Country due to the heavy influence of the London standard during the Early Modern English stage; it is now moribund even there, hence the stereotyped bucolic "Zummerzet".

Huh, that is fascinating.

[Estav]

It has been argued that there was variation between dialects of Old English in the voicing of word-initial fricatives.

I am personally sympathetic to this belief, but the mainstream view is that initial fricative voicing in the dialects of Southern England and the the south-western Midlands only occured in Middle English, as it is only then it appears in the orthography; compare Anglian OE scep, fleoman, swin [ˈʃ(t͡ʃ)eːp ˈfleo̯ːmɑn ˈswiːn] to Kentish ME ssep, vleme, zuin [ˈʒeːp ˈvleːmə ˈzwiːn]; c.f. London schep, fleme(n), swope(n) [ˈʃeːp ˈfleːmə(n) ˈswiːn] "sheep, to flee, pig". As London is adjacent to the area affected (and London English may have originally posessed it before immigration from more northerly parts), some traces of this survives in Standard English (vat, vane < OE fæt, fanu). This initial voicing retreated to the West Country due to the heavy influence of the London standard during the Early Modern English stage; it is now moribund even there, hence the stereotyped bucolic "Zummerzet".

Isn't it clear that visible effects of a sound change in spelling can only provide a terminus ante quem, not a terminus post quem for the sound change? In modern German, word-initial [z] (in accents that have it) is represented by ⟨s⟩ up to the present. I based my previous message on Roger Lass's "Old English fricative voicing unvisited", which discusses the question.

[Sol717]

Isn't it clear that visible effects of a sound change in spelling can only provide a terminus ante quem, not a terminus post quem for the sound change? In modern German, word-initial [z] (in accents that have it) is represented by ⟨s⟩ up to the present. I based my previous message on Roger Lass's "Old English fricative voicing unvisited", which discusses the question.

I'm aware of Lass's paper and am, as I said, "sympathetic to [the] belief" that West Saxon initial fricatives were voiced, though I think the evidence is insufficient to make a definitive judgement. This is why I talk about the "mainstream belief" that initial fricatives weren't voiced; it sometimes seems like a crude retrojection of the modern standard pronunciation when I'm in a unsympathetic mood, just like the supposition that (early) Old English <sc> was [ʃː], not [ʃt͡ʃ] or [sc].

[Estav]

I'm aware of Lass's paper and am, as I said, "sympathetic to [the] belief" that West Saxon initial fricatives were voiced, though I think the evidence is insufficient to make a definitive judgement. This is why I talk about the "mainstream belief" that initial fricatives weren't voiced

I see. I had wanted to add a summary of Lass's arguments but found I couldn't easily make one as he seems pretty uncertain too. I'm not sure whether I am inclined to defer to the mainstream in this case so I am curious about any other evidence that pertains to this.

; it sometimes seems like a crude retrojection of the modern standard pronunciation when I'm in a unsympathetic mood, just like the supposition that (early) Old English <sc> was [ʃː], not [ʃt͡ʃ] or [sc].

I'd wondered about the [ˈʃ(t͡ʃ)e] in your previous post. In word-initial position, there's evidence that Old English sc- was palatalized before back as well as front vowels, right? My assumption was that this was a result of an initial stage with [sx]- (as in Dutch) undergoing coalescence and fronting to [ʃ]... although I don't know exactly why fronting would occur in that context.

[Sol717]

I see. I had wanted to add a summary of Lass's arguments but found I couldn't easily make one as he seems pretty uncertain too. I'm not sure whether I am inclined to defer to the mainstream in this case so I am curious about any other evidence that pertains to this.

I'm not aware of anyone else who discusses this topic in-depth; I've checked a few works that cite Lass, but they don't go beyond a brief mention. I believe this may be because there isn't else to discuss because the evidence is so sparse.

I'd wondered about the [ˈʃ(t͡ʃ)e] in your previous post. In word-initial position, there's evidence that Old English sc- was palatalized before back as well as front vowels, right? My assumption was that this was a result of an initial stage with [sx]- (as in Dutch) undergoing coalescence and fronting to [ʃ]... although I don't know exactly why fronting would occur in that context.

I believe if that was the case, we'd have spellings such as <sh> in OE, but those sorts of spellings aren't attested until the ME period. A explanation more congruent with the spelling evidence is that [k] fronted to [c] or something under the influence of the preceding sibilant, though I'm not aware of any other instance of this kind of sound change.