A Dissertation
Presented to the Faculty of the Graduate School
of Cornell University
in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy
by
Haraldur Bernhardsson
August 2001

Ever since the discovery of Verner's Law in 1875, there have been persistent questions regarding its outcome in Gothic. Beside unambiguous examples of Verner's Law, there are numerous instances in the attested form of Gothic where its voicing effects fail to appear, contrary to expectation based on historical comparison. The handbooks offer mostly two lines of reasoning to explain this. On the one hand, it is claimed that the voicing effects of Verner's Law were subject to widespread elimination by analogy; this is the predominant view. On the other hand, some argue that the accent development in Gothic was different from that of the other Germanic languages, and since Verner's Law was conditioned by accent, this affected its outcome.

This dissertation presents an examination of the relevant data and existing explanations, as well as the characteristic features of Gothic phonology that could affect the outcome of Verner's Law, such as Thurneysen's Law and word-final devoicing. The investigation relies heavily on historical comparison, involving material from the other Germanic languages, as well as other Indo-European sources. It emerges that the explanations offered in the literature for the defective outcome of Verner's Law in Gothic are wholly inadequate. In particular, a small core of examples are identified where the absence of the expected Verner's Law voicing simply cannot be explained by way of analogy.

In all of these examples, it appears that the Verner's Law voicing of a sibilant or a velar has been reversed. A further examination reveals that there is a correlation between the outcome of Verner's Law in root-final position and the nature of the following syllable. Based on these findings, it is argued that in pre-Gothic there was a phonological rule of dissimilation which devoiced a sibilant or a velar fricative voiced by Verner's Law if it was followed in the next syllable by a voiced obstruent. This rule not only accounts for all the problematic examples, but also has significance beyond Gothic, for it raises questions about the development of the Proto-Indo-European voiced aspirates in Proto-Germanic and the Germanic voiceless velar fricative.

The dissertation can be obtained from ProQuest UMI Dissertation Publishing.