This paper presents an account of st-(middle) verbs in Icelandic that treats -st as the head of a functional phrase-object agreement (AgrO)- and argues that, consequently, st-formation is a syntactic rather than a lexical process. Supporting evidence includes the fact that -st is a clitic rather than an inflectional or derivational element attached to the verb. This analysis has the advantage that st-verbs, which on the surface appear disparate in their interpretation (e.g., reflexive vs. passive), can be unified under a single syntactic operation. The differences in interpretation fall out from argument structure properties of individual verbs. A consistent mapping between morphology and syntax is obtained.*

Shannon (1990) argues that the notions of lexical aspect (Aktionsart) and transitivity are a central part of perfect auxiliary variation. This paper builds on Shannon's (1993, 1995) analyses of perfect auxiliary selection in German and other related languages. Shannon accounts for perfect auxiliary variation by means of prototypes for high transitivity and high mutativity as well as the notion of Aktionsart. This paper examines the transitivity/mutativity parameters and the concept of Aktionsart using the examples of perfect auxiliary + past participle found in the Old Saxon Heliand. An examination of some 150 examples reveals that the choice of auxiliary can be explained by Aktionsart and a definition of transitivity that is characterized by conceptual imagery.

An examination of the verb forms in Brother Hermann's Middle High German verse epic, Leben der Gräfin Iolande von Vianden (c. 1283 in the neighborhood of Luxembourg), is undertaken in the categories of loan words, athematic verb forms, class shifting in strong verbs, the ge-perfective prefix, and apocopation and addition of final n. The last category is developed to situate Iolande, which consists of approximately 5,963 lines of rhymed couplets, in the context of Old High German and Middle High German historical development.

Evidence suggests that Germanic languages resisted the spread of historical umlaut processes. We propose that examples of such superficially varied umlautless residues all yield to a single coherent phonological account. Specifically, these vocalic assimilations show strong preferences for reducing more extreme differences in place of articulation between trigger and target while failing to assimilate articulatorily closer vowels, so that triggering /i, j/ first and most consistently mutated /a/ and last and least consistently mutated /u(:)/. This vocalic cline interacts with the consonantal material intervening between trigger and target, so that obstruent clusters (especially velar or labial) are prone to inhibit umlaut. In both cases, the smaller the phonetic difference between target and trigger, the less likely umlaut is to occur. While umlaut failure has long been consigned to the margins of theories of umlaut, we argue that it is crucial, providing a snapshot of how umlaut unfolded across western and north Germanic.