OV and V-to-I in the history of Swedish
This paper relates the loss of V-to-I in Swedish, which occurs in the beginning of the 17th century, to changes in the OV system that take place a century before. Prior to ca. 1500, OV is predominantly derived by movement of an argument (Arg) within a single VP (Sg-movement). Such a derivation generates OV patterns where Arg either precedes a non-finite but follows a finite verb (VfArgVnf) or precedes a single finite verb (ArgVf). In the 16th century, the preference for Sg-movement of Arg is reanalysed as a preference for movement to the highest VP (H-movement). H-preference favours ArgVf and the previously marginal ArgVnfVf, but disfavours the former dominant VfArgVnf. Unlike VfArgVnf, both ArgVf and ArgVnfVf have Vf in situ, i.e. they provide cues for new acquirers of the language that V-to-I does not apply.