This dissertation investigates the change from the nearly free relative order of verbs in subordinate clauses in Early New High German (1350-1650) to the more fixed order of Modern Standard German. Chapter 2 presents a corpus study of nearly 3,000 subordinate clauses from 30 texts from a broad range of dialects from the 14th to the 16th century, the most comprehensive overview of ENHG verb clusters to date. Several factors that influence verb order are identified: syntagm type, prefix type, extraposition, focus, and sociolinguistic factors. Chapter 3 breaks this data down by dialect and individual text, showing that most of these factors have similar effects across the dialects and tracing the decline of particular orders and favoring factors over time. Chapter 4 examines these orders in contemporary German, concentrating on the effect of focus on verb order. A survey with speakers of Austrian dialects and Swabian shows that although the Standard German orders are preferred, the non-standard orders may occur under the appropriate focus conditions. A magnitude estimation experiment demonstrates that variation in the Standard German werden-modal-infinitive construction is also sensitive to focus. In Chapter 5, the data from the previous chapters are used to demonstrate that the more traditional SOV approach to the structure of German is slightly preferable to the SVO hypothesis and that non-SOV surface orders are derived by rightward movement. Additionally, a principle is proposed to account for the relationship between focus and word order: a non-normal word order indicates a marked focus interpretation. Chapter 6 discusses the implications of this research for the history of the German language and for language change in general.