113
Mostly, ablaut forms due to I—>U-alternation do not conform to the de-
fault rule. As ablaut forms are never diphthongal and as there are two sub-
types of monophthongal U-forms, viz. «-form and «-form, this means that the
forms in question are «-forms (cf. sprech-^sproch). However, there is a mi-
nority of stems like
BINDL, SINGL,
and
TRINKL
that are not subject to this spe-
cial treatment. The default rule takes over again, and I—►U-alternation results
in «-forms (bund, sung, trunk, which are ‘participle stems’). This applies to
stems in which the stem vowel is followed by nd/ng/nk. As for these ‘na-
sal+consonant-stems’ (or ‘NC-stems’, for short), the special regulation that
requires «-forms is suspended. Its results would not fit into German phonotac-
tics.18
As regards stems that show change of complexity type, ablaut forms due
to U-+U-alternation join the pattern of I—►U-altemation. In sum, ablaut forms
due to I—>U-altemation or U—»U-altemation do not conform to the default rule
if possible; they are, as a rule, «-forms. (Here the proviso ‘if possible’ ac-
counts for the exclusion of NC-stems.) No other special stipulations concern-
ing ablaut vowels are needed. Given this, it suffices to specify the quality type
and the quantity type of an ablaut form in order to fix its stem vowel.
2.5 Change of quantity type
Quantitative ablaut (i.e., change of quantity type between stem forms) in
German is a poorly understood phenomenon. Reference grammars record
patterns of quantitative ablaut on a case-by-case basis (e.g., Curme 1922: 302,
305) but no uniform overall picture emerges (cf. Segeral 1995: 80-84). Gen-
erative rules have also been proposed but the systematic role of quantitative
ablaut has remained obscure.19
As a first step towards uncovering the regularity of quantitative ablaut,
two kinds of quantitative change may be distinguished:
18 With these stems, the vowel is followed by a nasal+consonant-cluster or by the velar
nasal (which, from a phonotactic point of view, ‘counts’ as a nasal+consonant-cluster in
German) but h i hardly ever occurs in this context, at least in ‘native’ stems. Hence the
‘aberrant’ ablaut formation (that is, the appearance of /u/ ‘instead o f hi) found with NC-
stems is regarded as an interference crossing the proper regularities of ablaut (Durrell 1980:
21, Barbour 1982: 343).
19 See Wurzel (1970); Bittner (1995: 191) offers an amended version of Wurzel’s rule (as
regards descriptive adequacy) but he does not arrive at a clarification of the relationship
between qualitative and quantitative ablaut, let alone of the function of quantitative ablaut.
114
Short—>long-alternation. A short base alternates with a long ablaut form.
Examples: /a/—*/i:/ (fall-fiel), /e/—*/a:/ (sprech-sprach), /a/—*/u:/ (schaff-schuf)
Long^> short-alternation. A long base (i.e., a diphthongal or a long
monophthongal base) alternates with a short ablaut form.
Examples:
/i:/—>/o/ (gieß-goss),
/a
u/—»/o/ (sauf-soff),
/at/—
»/i/ (reit-ritt)
(For convenience, I shall also use the terms lengthening and shortening.) The
majority of stems do not allow a change of quantity between their forms, that
is, long bases correspond to long ablaut forms (e.g., bleib-blieb; stehl-stahl-
stohl); and short bases correspond to short ablaut forms (e.g., bind-band-
bund). The question arises, which types of stems do allow quantitative change
at all?
In forms of strong verb stems, short vowels are not allowed to appear be-
fore voiced obstruents.20 To be sure, this fact does not constitute a peculiarity
of verbal stem formation. As is well known, sequences of short vowel plus
voiced obstruent are highly marked in German from a phonotactic point of
view (Heidolph et al. 1981: 980, 985).21 When they do occur at all (e.g., in
forms of nouns such as
EBBEW, FLAGGEW)
this usually points to non-native
origin (Paul 1916: 273, 301, from Low German). It does not come as a sur-
prise, then, that strong verb stems such as
SCHIEBL,
h e b
l ,
l ie g
l ,
or
g r a b l
do
not allow quantitative change. In their forms, vowels are long throughout as
they are followed by voiced obstruents.22 Thus stems are exempt from quanti-
tative ablaut if ablaut vowels are immediately followed by voiced obstruents.
The same holds when stem forms end in a vowel
(cf. SEHL).
On the other hand, stem vowels may be either long or short when they
are followed by voiceless consonants; cf., e.g., bases in short vowel plus /$/,
/{/, Is/ (sprech, stech, treff, schaff.ess, mess etc.) and in long vowel or diph-
thong plus /g/, /f7, /s/ (kriech, riech, trief, schlaf, lauf, ruf, gieß etc.). In fact, it
is stems that have forms such that the segment immediately following the stem
20 I follow a traditional practice in using the distinction
voiced/voiceless,
which might be
replaced by
lenis/fortis
(Kohler 1995: 157). Nothing depends on a specific analysis of
German consonantism in the present context.
21 Not only in German, of course; see Kohler (1995: 159
et passim
) and, in particular with
respect to diachrony, d’Alquen (1979), and the literature quoted there.
22 This assumes a phonological level where ‘final obstruent devoicing has not yet applied’.
The final obstruents in question may be voiceless due to final devoicing when stem forms
such as
schob
of
SCHIEB0
appear in verb forms such as
schob, schobt, schobst.
However,
final devoicing is not relevant to ablaut regularities (Halle 1953: 46). (No specific account
of final devoicing need be presupposed in the present context.)