LINGUIST List 6.460

Mon 27 Mar 1995

Disc: German Affricates

Editor for this issue: <>


  1. Desrochers Richard, Re: 6.393 Sum: German /tS/ (tsch)
  2. "David M. W. Powers", Affricates--one segment or two

Message 1: Re: 6.393 Sum: German /tS/ (tsch)

Date: Wed, 22 Mar 1995 13:51:15 Re: 6.393 Sum: German /tS/ (tsch)
From: Desrochers Richard <desrochrERE.UMontreal.CA>
Subject: Re: 6.393 Sum: German /tS/ (tsch)

As naive as may seem to be the following arguments, I see no reason to
consider German /tS/ as a single segment. As far as I know (I have no
German grammar or dictionary at hand), the distribution of /tS/ is very
limited, it occurs mainly in word-final position (or within compounds
like Deutschland, does not seem to occur in initial position or even
within the same morpheme in a word. On the other hand, have the sequence
/St/ in word initial position (Stark, Stump, buchStabieren, Stein.
Compare with this /pf/ which is more plausibly a single segment, since it
occurs in a greater variety of contexts (Pferd, A"pfel, Kopf).(and /fp/
does not seem to occur, at least in word-initial and final position)
For me, that settles the question, unless I have overlooked decisive data.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Affricates--one segment or two

Date: Sat, 25 Mar 1995 13:40:43 Affricates--one segment or two
From: "David M. W. Powers" <>
Subject: Affricates--one segment or two

Alexis Manaster Ramer (amrCS.Wayne.EDU) writes:
) The recent query and summary by Larry Trask on German affricates
) makes me think of the fact that, unless I am mistaken, it is
) perfectly easy to HEAR the difference between one- and two-segment
) stop-fricative sequences at least in some cases.

I don't think I responded to the original query, and I'd always assumed
that tsch was a single segment, although in my automatic analysis of
German (and Dutch) it is treated as two segments (but this is
a function of the assumptions and parameterizations used and means little
as "tcsh" was not systematically studied and controled).

Alexis's comment, however, reminds me that
"tschuess" (tcsh=FC=DF) derives from =E0 dieu or adios or adieu (or ad

(I've seen it attributed to both Fr. and L. although that doesn't explain=
final segment, which is historically s+z. As for the "tsch" it seems to =
from a palatalized dental [dy].)

Conversely, with "deutsch*", I can't help suspecting that adjectival -(i)sch
has attached to deut (clear/right) as to Angle (English), etc. I=
n this case "tcsh" derives from two segment [t] and [sh].

There are occurences which are clearly two segment (doing a grep in the s=
tolfi =

wordlist to find other examples of which I show representative below)


(I had a quick look through Dutch too, and all occurences fitted this pat=

The others fit the pattern of a -*(s)ch* ending (of which many are
verbs involving noise - does anyone know anything about the common ending=
and the others are adjectives, as discussed already in relation to Deutsc=


There are also some which aren't absolutely clear to me at first glance, =


(In all of this - means this an affix, * means it is a morph
exhibiting a variety of inflectional or morphophonological extensions,
parentheses mean optional or alternative or standard orthography,
brackets mean phonetic, quotes mean orthographic).

Excluding the cases which cross syllables, it seems to me that a
historical segment pair has become, or is becoming, a single segment,
although it is sufficiently rare that I don't think any of my
segmentation algorithms would classify it that way under any

David Powers
Associate Professor David Powers
SIGART Editor; SIGNLL Chair Facsimile: +61-8-201-3626
Department of Computer Science UniOffice: +61-8-201-3663
The Flinders University of South Australia Secretary: +61-8-201-2662
GPO Box 2100, Adelaide South Australia 5001 HomePhone: +61-8-357-4220
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue