Final devoicing of /z/ in North American English

Discussion of natural languages, or language in general.
Post Reply
Travis B.
Sumerul
Sumerul
Posts: 3570
Joined: Mon Jun 20, 2005 12:47 pm
Location: Milwaukee, US

Final devoicing of /z/ in North American English

Post by Travis B. »

Lately I have been interested in how widespread final devoicing is in English, and in particular I have been paying attention to the phoneme for which it is most obvious, /z/. To avoid being too biased towards how those directly around me speak, and to avoid looking at the informal register I am used to, I have been listening for it in advertisements and music rather than in the speech of people I know. And what it seems to me is that, in what I hear on the radio, most of the time final /z/ is voiceless or at least markedly less voiced than initial /z/ when not followed by a vowel. Furthermore, if it follows a plosive, that is devoiced as well. Yet I do not see mentions of final devoicing in relatively standard English, where English is typically contrasted with the likes of German, Dutch, Russian, and Polish. Surely this has been documented in journals - which I have not read because they cost money and I do not belong to a university who will pay for them for me - but overall it does not seem to be readily known. Why is this?
Dibotahamdn duthma jallni agaynni ra hgitn lakrhmi.
Amuhawr jalla vowa vta hlakrhi hdm duthmi xaja.
Irdro. Irdro. Irdro. Irdro. Irdro. Irdro. Irdro.

Vijay
Smeric
Smeric
Posts: 2244
Joined: Sat Feb 06, 2016 3:25 pm
Location: Austin, TX, USA

Re: Final devoicing of /z/ in North American English

Post by Vijay »

Maybe it has something to do with the fact that final devoicing also applies to a wider range of sounds in all of those languages?

Travis B.
Sumerul
Sumerul
Posts: 3570
Joined: Mon Jun 20, 2005 12:47 pm
Location: Milwaukee, US

Re: Final devoicing of /z/ in North American English

Post by Travis B. »

That is a possibility, but I am not sure if that is it. I chose /z/ because it is the phoneme I can most easily hear the voicing, or lack thereof, of. It is harder for me to tell whether plosives really are voiced or not. I can tell sometimes when people are using clearly voiced or clearly voiceless realizations of final lenis phonemes, but I have to listen more carefully to do so. Sometimes people will use clearly fully voiced realizations of final lenis plosives, which really sticks out to me, but most of the time I do not really register their voicing per se (but rather hear the preceding vowel length and preglottalization or lack thereof).
Dibotahamdn duthma jallni agaynni ra hgitn lakrhmi.
Amuhawr jalla vowa vta hlakrhi hdm duthmi xaja.
Irdro. Irdro. Irdro. Irdro. Irdro. Irdro. Irdro.

Sumelic
Avisaru
Avisaru
Posts: 385
Joined: Sat Mar 28, 2015 7:05 pm

Re: Final devoicing of /z/ in North American English

Post by Sumelic »

For me, at least, any final devoicing that occurs in this position (I assume it does, but it's hard for me to hear the difference between devoiced and voiced consonants in this position) is entirely allophonic, due to factors like vowel length allophony and possibly consonant length (I'm not sure, but I think /s/ is longer than /z/ in English).

My impression is that the familar European languages standardly described as having "final voicing", like German, Dutch, Polish, Russian etc., are described this way because the contrast between voiced and unvoiced phonemes appears to be more or less neutralized in this position. (I say "more or less" because I've seen a number of studies saying that the neutralization may in fact be incomplete, but they rely on measuring very small differences in average vowel length and subtle stuff like that. But that kind of "difference" doesn't seem to be the same as my experience as a native speaker of English, where it's just obvious that e.g. "sauce" and "saws" sound differerent).

Devoicing phenomena in English are certainly known, but I think they haven't been considered as interesting since they seem more like phonetic implementation details, whereas German final devoicing was traditionally analyzed as a phonological constraint (mentioned in Assessing incomplete neutralization of final devoicing in German, Röttger et al.).

(More personal introspection: to me, /z/ at least doesn't sound devoiced when preceded and followed by voiced sounds; I am thinking about the phrase "Dad's done" as compared to "that's done". The /sd/ sequence of the second sounds very much like /st/ to me, as in "stone", but the /zd/ sequence of the first really doesn't.)

I believe I've read somewhere that French also has interesting voicing and devoicing processes, but in general they aren't well-known because like in English, they usually don't neutralize contrasts (aside from a few obvious cases that mainly arise from syncope such as "médecin" with /ts/, "je suis" > "chuis"). Here's the abstract of a paper I found about voicing/devoicing assimilation in French /s/ and /z/: https://www.ncbi.nlm.nih.gov/pubmed/22127548

Travis B.
Sumerul
Sumerul
Posts: 3570
Joined: Mon Jun 20, 2005 12:47 pm
Location: Milwaukee, US

Re: Final devoicing of /z/ in North American English

Post by Travis B. »

I agree why this is commonly overlooked is probably due to any final devoicing in English being purely phonetic, thanks to vowel length allophony, preglottalization, consonant length, etc. still marking fortis versus lenis for final obstruents. This is like how most people have not heard of vowel length allophony or preglottalization themselves, despite them being key parts of the perception of obstruent and plosive fortisness versus lenisness respectively.

About clusters with /z/, I find that for me clusters of obstruents are normally voiceless unless emphasized, so /zb zd zg/ for instance normally come out as [sp st sk]. But I cannot say that this is typical by any means.
Dibotahamdn duthma jallni agaynni ra hgitn lakrhmi.
Amuhawr jalla vowa vta hlakrhi hdm duthmi xaja.
Irdro. Irdro. Irdro. Irdro. Irdro. Irdro. Irdro.

Travis B.
Sumerul
Sumerul
Posts: 3570
Joined: Mon Jun 20, 2005 12:47 pm
Location: Milwaukee, US

Re: Final devoicing of /z/ in North American English

Post by Travis B. »

I wonder what the actual distribution of final devoicing is in English; e.g. my parents (both from southeastern Wisconsin) devoice very inconsistently, often having full-voiced obstruents even utterance-finally, whereas I hear local commercials with clear and consistent devoicing of final sibilants and final sibilant-containing clusters and little obvious voicing of other final obstruents, to when I heard this one program which had someone on it from Oklahoma who was speaking in length who had extremely obvious final devoicing (like it almost seemed like final /b d dʒ g/ were being aspirated, because they almost sounded like a careful /p t tʃ k/ minus preglottalization to me).
Dibotahamdn duthma jallni agaynni ra hgitn lakrhmi.
Amuhawr jalla vowa vta hlakrhi hdm duthmi xaja.
Irdro. Irdro. Irdro. Irdro. Irdro. Irdro. Irdro.

User avatar
jal
Sumerul
Sumerul
Posts: 2633
Joined: Tue Feb 06, 2007 12:03 am
Location: Netherlands
Contact:

Re: Final devoicing of /z/ in North American English

Post by jal »

Travis B. wrote:so /zb zd zg/ for instance normally come out as [sp st sk]. But I cannot say that this is typical by any means.
Well, given the prevelance of "Asbergers" for "Aspergers" on the 'net, I think it's pretty common to confuse the two.


JAL

Travis B.
Sumerul
Sumerul
Posts: 3570
Joined: Mon Jun 20, 2005 12:47 pm
Location: Milwaukee, US

Re: Final devoicing of /z/ in North American English

Post by Travis B. »

jal wrote:
Travis B. wrote:so /zb zd zg/ for instance normally come out as [sp st sk]. But I cannot say that this is typical by any means.
Well, given the prevelance of "Asbergers" for "Aspergers" on the 'net, I think it's pretty common to confuse the two.
That is a different matter, specifically that the fortisness/lenisness of plosives is neutralized after sibilants. But in Standard English you can have both [sp st stʃ sk] and [zb zd zdʒ zg] clusters, as the key thing is that the fortisness/lenisness of the sibilant and the plosive must agree. So the innovation is the merger the realization of these two sets of clusters except by preceding vowel length.
Dibotahamdn duthma jallni agaynni ra hgitn lakrhmi.
Amuhawr jalla vowa vta hlakrhi hdm duthmi xaja.
Irdro. Irdro. Irdro. Irdro. Irdro. Irdro. Irdro.

User avatar
GamerGeek
Sanci
Sanci
Posts: 42
Joined: Sat May 27, 2017 10:03 am
Location: The Universe
Contact:

Re: Final devoicing of /z/ in North American English

Post by GamerGeek »

Travis B. wrote:About clusters with /z/, I find that for me clusters of obstruents are normally voiceless unless emphasized, so /zb zd zg/ for instance normally come out as [sp st sk]. But I cannot say that this is typical by any means.
My onset clusters are always /sb sd sg/ and my coda clusters are always /sp st sk/.

Vijay
Smeric
Smeric
Posts: 2244
Joined: Sat Feb 06, 2016 3:25 pm
Location: Austin, TX, USA

Re: Final devoicing of /z/ in North American English

Post by Vijay »

I'm pretty sure I remember some of the linguistics professors at a conference I went to in grad school saying that there isn't really any such thing as a voiced fricative because voicing isn't sustainable in a fricative. I don't think I remember what their argument for this was.

Travis B.
Sumerul
Sumerul
Posts: 3570
Joined: Mon Jun 20, 2005 12:47 pm
Location: Milwaukee, US

Re: Final devoicing of /z/ in North American English

Post by Travis B. »

Vijay wrote:I'm pretty sure I remember some of the linguistics professors at a conference I went to in grad school saying that there isn't really any such thing as a voiced fricative because voicing isn't sustainable in a fricative. I don't think I remember what their argument for this was.
I don't buy that, since at least here prevocalic lenis fricatives voice much more strongly than prevocalic lenis plosives - actually, the only time plosives voice consistently is intervocalically and in intervocalic clusters with other sonorants (except /dʒ/, which much of the time does not even voice then), whereas initial lenis fricatives have very obvious voicing.
Last edited by Travis B. on Thu Jun 01, 2017 10:23 pm, edited 1 time in total.
Dibotahamdn duthma jallni agaynni ra hgitn lakrhmi.
Amuhawr jalla vowa vta hlakrhi hdm duthmi xaja.
Irdro. Irdro. Irdro. Irdro. Irdro. Irdro. Irdro.

Vijay
Smeric
Smeric
Posts: 2244
Joined: Sat Feb 06, 2016 3:25 pm
Location: Austin, TX, USA

Re: Final devoicing of /z/ in North American English

Post by Vijay »

Maybe it's also worth pointing out that final devoicing of obstruents is very common just in general, cross-linguistically.

User avatar
GamerGeek
Sanci
Sanci
Posts: 42
Joined: Sat May 27, 2017 10:03 am
Location: The Universe
Contact:

Re: Final devoicing of /z/ in North American English

Post by GamerGeek »

Vijay wrote:Maybe it's also worth pointing out that final devoicing of obstruents is very common just in general, cross-linguistically.
There's even a fancy word for it: (not final devoicing) desonorization. Then again, sonorization refers to voicing. It still sounds fancy.

Vijay
Smeric
Smeric
Posts: 2244
Joined: Sat Feb 06, 2016 3:25 pm
Location: Austin, TX, USA

Re: Final devoicing of /z/ in North American English

Post by Vijay »

GamerGeek wrote:
Vijay wrote:Maybe it's also worth pointing out that final devoicing of obstruents is very common just in general, cross-linguistically.
There's even a fancy word for it: (not final devoicing) desonorization. Then again, sonorization refers to voicing. It still sounds fancy.
"Desonorization" sounds confusing IMO, as if it was referring to sonority.

Post Reply