LINGUIST List 6.600

Sun 23 Apr 1995

Disc: Affricates

Editor for this issue: <>


  1. , affricates

Message 1: affricates

Date: Thu, 20 Apr 1995 00:05:43 affricates
From: <>
Subject: affricates

For those who require "intersubjective verification" of Alexis' claim
that it is possible to HEAR the difference between [t+S] and [tS], I can
hear it. (I can't address the German facts because I don't know it well
enough, unfortunately.) I have heard at least one French pronunciation
of the bit before the vowel in the country spelled "Tchad" in which this
was VERY different from English pronunciations of the bit before the
vowel in "Chad" -- and this difference seemed to be clearly [t+S] vs.
[tS]. (This seems to contradict what Alexis says about "Tchekoff" --
interspeaker differences? Speaking of which, English speakers I have
heard have [s] at the beginning of "tsunami" (and [t] in "tsetse"!).)
By the way, both of these sound different from geminate affricates, like
the [ttS] in Hausa _waccan_ "that (one), fem.".

As to the phonetic difference, although I have not done any instrumental
work, I would be very surprised if it was not durational differences, at
least of the frication part (probably closure too when it's possible to
tell) -- considerably longer for [t+S]. (Geminate affricates have longer
closures than simplex affricates, and (near?) identical frication durations.)

Also, some dialects of Fula are described as having CONTRASTS among
(tautosyllabic) prenasalized stops, (heterosyllabic) nasal + stop, nasal
+ prenasal, and geminate prenasalized stops. (I may be wrong about one
of these, but there is at least a three-way contrast; I believe that
the variety described in D. W. Arnott's _The nominal and verbal systems
of Fula_ has all four.) Since Fula speakers presumably can hear these
differences, the two-way Polish [t+S]/[tS] contrast ought to be a "piece
of cake"!

At the risk of trodding away from phonetics toward phonology/psychology,
the subject of prenasals brings to mind an incident that may be of
interest to those who do not happen to have taught phonetics using
Ladefoged's _Course ..._ to a class that included (perceptive) native
speakers of Russian and Zulu (or their equivalent for the following
purposes). In the course of discussing "different types of stops" (pp.
165-6), L gives as an example of a stop "with nasal release" the Russian
[dno] `bottom'. The Russian speaker pointed out, with absolutely no
prompting, that this [dn], unlike the first seven examples that it
followed, was a sequence of TWO segments. At which point, the Zulu
speaker noted, equally spontaneously, that the next example, a (Swahili)
prenasalized stop ([ndizi] `banana'), was NOT a sequence of two
segments, at least in Zulu (despite the biliteral orthography!). SOME
people, whether for phonetic reasons or phonological ones, apparently
have no trouble deciding whether or not they are dealing with a sequence
of segments.

I will now turn directly to phonology, in particular, how one might
decide how to represent the above phonetic things phonologically. It
depends in part on whether one takes a "God's truth" or a "hocus pocus"
approach (in Householder's sense) to linguistic analysis. If the
latter, then anything goes: why worry if pure complementary distribution
criteria result in an analysis in which English [h] and [N] (the velar
nasal) are allophones of the same phoneme (the former occurs only
syllable-initially, the latter syllable-finally), regardless of how
counter-intuitive it is to native speakers? We have a compact description
of the phonetic facts (and reduce the number of phonemes by one!) If
the former, and if we're lucky, we'll have an informant with strong
intuitions (and if we're REAL lucky, these intuitions contradict what
we'd expect based on the orthography, like in the Zulu case mentioned
above). If we're not so lucky, and if the "God's truth" we're seeking
is "how native speakers represent these things in their mental
dictionaries", then we'll have to (gasp!) do some work. We can't
assume that traditional linguistic criteria will work; many of them
have been conceived by folks who took (or claimed they did) a "hocus
pocus" approach, after all. We might try to find people who are "fluent
backward talkers" (as Cowan and Leavitt did -- see, most accessibly for
linguists, their 1981 CLS Parasession on Language and Behavior paper) --
ones who can pronounce words with the order of phonemes reversed, and as
fast as simultaneous translators do their trick -- and see if they
behave like "other" sequences of segments [they don't, in English!]. Or
see if children who invent their own spelling systems ever treat them as
if they are "t" plus whatever they use for [S] (which, alas, does not
occur in the name of any English letter, unless you count the post-vowel
part of "h" -- and kids don't seem to), as unfortunately Charles Read
did not do in his 1975 _Children's categorization of speech sounds in
English_. Or look at speech errors (N. Davidsen-Nielsen, 1975 J IPA
[sorry, Jim Fidelholz, I don't KNOW what his/her first name is!]. Etc.,
etc. But knowing the etymology (and Alexis meant this literally, Benji
-- as in David Powers' claim that `"tschuess" ... derives from =EO dieu
or adios ...') is irrelevant to what's in most folks's mental
dictionaries (and, as Alexis notes, irrelevant to ANYbody's phonetics).

Don Churma, Dept. of English, Ball State University, Muncie, IN 47306
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue