LINGUIST List 3.187

Thu 27 Feb 1992

Disc: Devoicing

Editor for this issue: <>


  1. Henry Kucera, Re: 3.169 Queries: Chinese, Dictionary, Final Devoicing
  2. , Re: 3.169 Queries: Chinese, Dictionary, Final Devoicing

Message 1: Re: 3.169 Queries: Chinese, Dictionary, Final Devoicing

Date: Sat, 22 Feb 92 11:46:32 ESRe: 3.169 Queries: Chinese, Dictionary, Final Devoicing
From: Henry Kucera <>
Subject: Re: 3.169 Queries: Chinese, Dictionary, Final Devoicing

 The two pronunciations of final (normally) voiceless consonants in Russian
and Czech (as voiced and voiceless, depending on the morphology) is certainly
well established. It occurs in cases when the distinction between homophones
 (e.g. Russian "rot" and "rod") needs to be conveyed. Since the spelling in
these languages is morphophonemic, this kind of articifical pronunciation is
simply a variety of spelling pronunciation under specific circumstances.

 It should be also noted that in Czech and Russian (in contrast to German, for
 example), the voice-voiceless opposition in final word position is "suspended"
-- as Trubetzkoy called it-- BUT the actual realization of a morphophonemically
voiced consonant can be either voiced or voiceless, and vice versa, depending
on the environment that follows the word or preposition boundary. For a
discussion of these facts in generative phonology, cf. Halle, The Sound
Pattern of Russian, initial chapters. For Czech, where the situation is very
complicated, a description is given in my Phonology of Czech (out of print).

Henry Kucera
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Re: 3.169 Queries: Chinese, Dictionary, Final Devoicing

Date: Mon, 24 Feb 92 09:15 EST
From: <KINGSTONcs.umass.EDU>
Subject: Re: 3.169 Queries: Chinese, Dictionary, Final Devoicing

Regarding Alexis Manaster-Ramer's query about final devoicing.
Two points:

1) Thai is well-known for having just a single series of stops syllable-
finally, neutralizing an initial contrast between prevoiced, voiceless
unaspirated, and voiceless aspirated stops (though prevoiced stops
are only found at the bilabial and alveolar but not palatal and
velar places of articulation). The usual phonetic description of the
final stops is voiceless unreleased, though some maintain that they
also have constriction or closure of the glottis (I don't recall the
references on this last point, but can dig them up if asked). However,
the stops are written (most of them anyway, see below) with the symbols
used in initial position to represent prevoiced stops, and not with
what would seem the "logical" choice, the symbols for the voiceless
unaspirated stops (there's an illogic to that choice, too, what I
won't go into). Furthermore, Mary Haas, whose understanding of Thai
I would hesitate to question, also insisted that the choice of symbol
represented some kind of phonetic reality, citing the pronunciation
of a consultant of the loan word _cheed_ as (approximately) [tShd^]
where [] is a mid back unrounded vowel, and the final stop is voiced,
with a voiced release. The force of this example is that it's the
English word "shirt" which of course has a voiceless final consonant
in the source language. One caveat, this pronunciation is cited as
occurring when the consultant was emphasizing the proper pronunciation
of the word. [Perhaps irrelevant notes on Thai orthography: Final stops
are also written with symbols that represent other kinds of consonants;
in initial position, e.g. stops with with other laryngeal articulations,
fricatives, and even [r], but these are less common and are largely
restricted, I think, to learned vocabulary, which are largely imports/
adpatations from Pali.

2. There have a series of investigations, largely by people at Indiana
(Dan Dinneson, Jan Charles-Luce, Bob Port, Louisa Slowiaczek), which
puport ot show that final neutralization of voicing in German, Polish,
and Catalan is phonetically incomplete, in that speakers produce stops
which are underlyingly [+voice] differently than those which are underlyingly
[-voice]. While I believe that much of the early work on this possibility
was seriously marred by flaws in method (mostly a matter of drawing the
speakers' attention to the possibility of a difference), the more recent
work has been more careful in this regard, in particular a paper by
Port in, I believe, Journal of Phonetics, in the last two years.

John Kingston
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue