Structures in Germanic Prosody : A diachronic study with special reference to the Nordic languages

This study provides a reconstruction of the development of the Germanic stress and syllabification system (as reflected in Gothic and Proto-Nordic) up to the completion of the quantity shift in Late Old Swedish. By means of current prosodic theory it is established that a domain of two moras wordinitially is present at all stages of development, in Gothic, Old English and Proto-Nordic as well as in Modem Swedish. It is argued that this domain is the linguistic instantiation of word minimality, referred to as the bimoraic condition. The bimoraic condition is interpreted over different prosodic categories - the prosodic word, the foot and finally the main-stress syllable - at different language stages. This development reflects the transition from a quantity system permitting light and overlong syllables to a system where the main-stress syllable is obligatorily heavy. Various prosodically conditioned changes and processes take place in the early Germanic dialects. New explanations are proposed for several of them in terms of prosodic theory. The vowel/glide alternation (Sievers’s law) in Gothic is derived from regular syllabification of the archiphonemes III and /U/. Syncope in Proto-Nordic (corresponding to high vowel deletion in Old English) is analysed as mora-deletion in metrically weak positions. Vowel shortening and nasal loss are also analyzed as mora-deletion following destressing under stress clash. The long-standing problem of delayed syncope (in Proto-Nordic) or absence of syncope (in Old English) in light stems is explained as a minimal word effect. The deletion rule in the so called second syncope period in Proto-Nordic is a case of vowel deletion (not mora-deletion). The patterning known as vowel balance is analyzed as the result of interaction between the general trend of reduction and the development of a particular balance prosody. In balance prosody one main-stress position (a unipositional foot) dominates two light syllables. This prosody is directly reflected as level stress on the surface. Rules that relate directly to the metrical configuration of balance are vowel strengthening (läsa > läså ’to read’), and vowel levelling (läså > låså). The latter rule is rendered as parametrized projection of features onto the stress unit, and the vowel patterns of vowel levelling are thereby given a principled description. Finally, the quantity shift in Old Swedish is discussed in detail. Balance - argued to be a Scandinavian innovation - is shown to be directly linked to the quantity shift. The loss of balance necessarily leads to the implementation of the quantity shift. Moreover, it is argued that the particular Central Scandinavian lengthening pattern involving both vowel and consonant lengthening (in roughly complementary contexts) is due to the (former) presence of balance. Finally, the theory predicts that the Modem Swedish quantity system emerging after the quantity shift depends on distinctive consonant quantity, rather than distinctive vowel quantity