LINGUIST List 6.490

Mon 03 Apr 1995

Disc: German Affricates

Editor for this issue: <>


  1. , Re: 6.472 German Affricates
  2. , Re: 6.472 German Affricates
  3. , 6.472 German Affricates

Message 1: Re: 6.472 German Affricates

Date: Fri, 31 Mar 1995 08:59:38 Re: 6.472 German Affricates
Subject: Re: 6.472 German Affricates

ursula.doleschalWU-WIEN.AC.AT (ursula.doleschal) writes:

)As to the recent postings on German affricates I wonder what the criteria
)are in order to posit a one- or two-segment solution. So far rather strange
)(to my mind) suggestions have been made: "you can hear it", "the affricate
)does not occur word-initially". Can some phonteists inform us, what are the
)usual criteria in their science? (Since I will not have the time to read
)through the relevant literature, but still find this an interesting
)question, the more since German is my native language).

I'm not a phoneticist, but I have a few suggestions:

1. We must assume an affricate where either the stop or the fricative is not
otherwise found in the language. Thus Spanish [ts^] must be analyzed as a unit
affricate phoneme /c^/, since there is no [s^] otherwise in Castillian or
American pronunciation.

2. Beyond that it's a pseudo-problem. Those of us who believe in phonemes
(count me in!) sometimes fail to realize that phonemes, unlike "allophones" or
"realizations" of phonemes, are not physically present in the stream of speech:
they are mental constructs, part of our linguistic equipment, but not something
existing "out there". (Blame American structuralists before Chomsky for this
one: striving at all costs to avoid "mentalism", they placed both phonemes and
allophones out in the physical world and hence regarded phonemes as groups or
classes of similar sounds.) But if they are mental constructs which must be
realized, then it is perfectly reasonable to say that [c^] in English _pitcher_
represents a unit phoneme /c^/, while in [hic^] _hit ya_ it is /ty/.
Moreover, the fact that we analyze [ts^] in German _Wirtschaft_ as /ts^/ tells
us nothing about whether the same sequence in _deutsch_ is /ts^/ or /c^/, i.e.
a sequence of phonemes or a unit phoneme. Indeed, different native speakers
can have differing analyses where, as (in my opinion) here, there is no solid
basis for preferring one over the other.

3. While phonemic affricates make good sense in many instances in many
languages, there seems to be no reason to worry about whether there are
phonetic affricates in *any* language -- even Spanish, where [s^] occurs only
after [t]. It's like the calender: while it makes perfect sense to establish
weeks in terms or organizing what we do and how we perceive time, they have no
basis in reality, which knows only days and couldn't care less whether they are
grouped in units of seven. We might as well worry about whether a phonetic
string contains a verb -- that too is a creature from a different level.

)I always thought that one criterion should be the possibility of separate
)release of the plosive and fricative parts, as has been illustrated by
)David Powers for morphologically complex words (I am repeating his
)examples, inserting a morpheme boundary "-"):

A useful criterion for establishing the phonemic representation of certain
words, though obviously not applicable in all cases.

)By the way, in my pronounciation putting a syllable boundary between the t
)and S of _entscheiden_ sounds very, very weird, in fact I syllabify
)en$tS..... Note that this verb is not very transparent to native speakers
)of German.

Agreed. But several analyses are possible:

 a. affricate /c^/
 b. sequence /ts^/ without word boundary
 c. sequence /t#s^/ with word boundary

While (c) would normally be expected after the prefix (consider the realization
of _enteignen_ as [ent?aiknn] by many speakers; [?] is a glottal stop), the
facts you cite make (a) or (b) much more likely. But none of this says
anything about whether there is a *phonetic* affricate, but rather only about
how native speakers analyze [ts^] in *this* instance.

Alexis Manaster Ramer (amrCS.Wayne.EDU) writes:

)Richard Desrochers and David Powers seem to me to miss the point
)about the distinction between one-segment affricates and two
)segment sequences of stop+fricative. Distributiuonal and
)etymological arguments are irrelevant if the distinction is
)audible (whether there is a contrast in the language in question
)or not).

True, and yet not entirely relevant either. I agree that an audible
distinction indicates that the more separate pronunciations do not have
a phonemic affricate. But it does not therefore follow that the more
"together" pronunciation must be analyzed as an affricate. Presence or absence
of a boundary between the two parts may be responsible. For instance, no one
has (to my knowledge) proposed that German has three phonemic degrees of vowel
length. Yet phonetically it does:

 Short: glichen [glIc,n] 'resembled'
 Long: kriechen [khRi.c,n] 'to creep'
 Extra long: Kniechen [khni:c,n] 'little knee'

The extra long pronunciation occurs primarily in stressed final position, e.g.
in _Knie_ 'knee'. This is entirely predictable and hence nonphonemic. Adding
the diminutive suffix does not change this. How can the vowel remain
"word-final"? Well, if we analyze the suffix as /#xn/ or the like, that's
what we have. The word boundary /#/ explains why /x/ in this suffix is
always realized as [c,] regardless of the preceding vowel, leading to such
monstrosities as _Kuhchen_ [khu:c,n] 'little cow' beside _Kuchen_ [khu.xn]
'cake'; they can be analyzed as /ku:#xn/ and /ku:xn/ respectively.

Note that the analysis above does *not* tell us whether the long vowels must be
analyzed as long vowels or as two-vowel sequences. So too, the "together"
pronunciation of a stop+fricative sequence does not tell us whether the
phonetic sequence must be analyzed as a sequence of phonemes or as a unit
phoneme, either in this instance or in other similar instances.

Leo A. Connolly Foreign Languages & Literatures University of Memphis Formerly "Memphis State University"
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Re: 6.472 German Affricates

Date: Fri, 31 Mar 1995 07:51:57 Re: 6.472 German Affricates
From: <>
Subject: Re: 6.472 German Affricates

I've been reading this discussion on German
affricates with some puzzlement. It's an old
chestnut, or course, one that probably will
never be finally resolved (if there even is
a resolution). Personally I still for the life of
me cannot accept Alexis' claim that the difference
between [tS] and [t-S] is an audible on in
German and that German clearly has [tS].
I for one can't hear the difference.

However, I don't have to cite my own poor ears:
no lesser an authority than William Moulton,
one of the grand old men of American structuralism
and probably one of the best experts on
German phonetics/phonology and an extremely
keen observer of what's "audible", didn't seem
to hear this. In his book _The sounds of English
and German_, chap. 5, Moulton specifically
analyzes English has having an affricate [tS] (=
[c^], i.e. c with hacek), but German as having
cluster [t-S]!

Now, just because Moulton didn't catch this
doesn't mean it isn't there, but given the fact
that at least some (and probably many) reputable
observers of German sounds haven't discerned the
audibility of the distinction in question, I
believe it falls on Alexis' shoulders to bolster
his claim. If it's audible, then certainly this
must have some sort of acoustic correlates that
can reliably be measured. Is there any such
evidence available? Short of that, it seems
Alexis will just have to teach the rest of
us to hear what he has no problem hearing.

Before closing, let me make it clear that I'm
not trying to be argumentative here and that
I don't have any particular stake in the issue,
one way or the other. All I'm concerned with is
the claim that one simply has to listen and the
issue will be decided. It'd be nice if that were
true, I'd really enjoy that (finally something
that's easy to determine!). However, like my
biblical namesake ("doubting Thomas") I need
to be shown. That's all. If this can be heard,
then surely we can get some corroboration
from other sources ("intersubjective verifiablity").
Thanks. Hope I don't stir up the proverbial
hornets' nest here!

tom shannon
german department, uc berkeley
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 3: 6.472 German Affricates

Date: Fri, 31 Mar 95 22:07:42 ES6.472 German Affricates
From: <>
Subject: 6.472 German Affricates

Kenneth L. Pike's book PHONEMICS (University of Michigan Press, 1947)
presents in Chapter 12 detailed procedures for determining whether a sequence
of two segments is a single unit phoneme or a sequence of phonemes.
This book is still available from the University of Michigan Press.
Ernest McCarus
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue