Abstract
This paper discusses the results of a nationwide study of a syntactic change underway in Iceland. The new construction appears to contain passive morphology, auxiliary vera ('to be') & a passive participle which can assign accusative case to a postverbal argument. The study was designed to test the hypothesis that the innovative construction involves the reanalysis of passive morphology as a syntactically active voice construction with a phonologically null impersonal subject. Such a reanalysis parallels the completed development of the -no/to construction in Polish & the autonomous form in Irish. I show that verbal morphology can be ambiguous between active & passive voice, & speculate about the reasons why this change is happening in Icelandic but in none of the other Germanic languages. Tables, References. Adapted from the source document