This thesis deals with the loss of OV word order in the history of Swedish. The principal aim is to provide an explanation of this development in light of the modern Germanic situation, where some languages (Nordic, English) have VO and some (continental West Germanic) OV word order. Other important questions concern the syntactic structure of OV and VO, and the relationship between OV and asymmetries in the word order of main and subordinate clauses.

The main conclusions and proposals can be summarized as follows. Older Swedish displays the characteristics of a type 3 language (T3). T3 implies being neither OV nor VO, but having verbal heads that can simultaneously identify post- and preverbal arguments, within a universal system of syntax that only accepts right-branching structures (the Basic Branching Constraint). The change from T3 to VO is due to verb movement in both main and subordinate clauses, a proposal more than well supported by word order differences within the Germanic language family. It is furthermore argued that the explanation is ultimately due to human parsing, VO being the most efficient word order in a system with general verb movement.

The development of OV and subordinate clause word order in the history of Swedish is also relevant when evaluating the Rich Agreement Hypothesis, the idea of a correlation between rich subject-verb agreement and verb movement to I°. It is argued that RAH is an essential part of subordinate clause word order in the history of Swedish, but should make different predictions for VO, OV and T3 languages and thus needs to be reformulated.