Orð og tunga - 01.06.2014, Síða 68

56
Orð og timga
These proportions (and also a look at true positives) show that in this
corpus all three variants are well documented. Tlie 3-1-2 order is
very noticeable in the south-eastern regions of the German-speaking
language area. In Austria, it is even the dominant (but not absolute)
variant. As leading grammar books such as the Duden grammar con-
sider themselves grammars ofusage, it is surprising that the 3-1-2 vari-
ant has not been marked as a standard variant yet, even more so, as it
is also highly common in the adjoining areas of Liechtenstein, South
Tyrol and the south-east of Germany (i.e. large parts of Bavaria).
Is this variation and this particular distribution a result of recent
developments in standard German? As mentioned above, three-verb
clusters are relatively rare, and we cannot expect high numbers of
instances in the three smaller corpora. The Kaiserreichkorpus gives no
hints as to regional variation in nineteenth century printed standard
German, there is evidence for the 1-3-2 order only. This may support
the monocentric tendencies after 1870, which von Polenz mentioned.
But note that the corpus is very small, so that the lack of data for vari-
ants may simply be due to limitations of the corpus size.
For the view 'from below', the Pfeffer Corpus and the Emigrant
letter corpus, render a different picture. In spite of the low numbers,
it quickly becomes clear that the areal distribution of the three-verb
cluster variants resembles the spread of the individual variants in
present-day standard German. Moreover, there has been a continu-
ity of usage in serialization patterns since the nineteenth century. In
southern Germany and Austria the 3-1-2 order is the only variant
that occurs in the Emigrant letter corpus (5 matches), while in north-
ern Germany also the initial position of the finite verb (1-3-2) shows
up and is highly dominant (> 70% within the variable). The distribu-
tion in the Pfeffer corpus is very similar, with the only difference that
the 1-3-2 order seems to have become stronger in the southern ar-
eas. With all due caution, we may conclude then, firstly, that the areal
distribution of the three variants in the Variantengrammatik corpus is
matched by the pattern that we find in the non-standard corpora and,
secondly, that there has been a stable pattern in non-standard German
over the last 150 years.
From the results of the corpus study, it becomes evident that an
explanation for the variation in verb clusters must be grammatically
as well as historically and sociolinguistically founded. Grammatically
speaking, prosodic-induced focusshifts (esp. focus on the lexical verb)
might have an effect on word order in verbal clusters, according to