The notion of oblique subject and its status in the history of Icelandic
The term ‘oblique subject’ is used in recent descriptions of Icelandic about NPs that behave syntactically like subjects without having nominative case. Data in support of such an analysis can easily be found in Modern Icelandic. Various linguists have assumed that also Old Icelandic has oblique subjects. In this paper I first discuss the notion of oblique subject on a metatheoretical basis. My claim is that oblique subject is not an empirical entity, it is a result of a decision to use it as a descriptive device because it may yield a more economical or elegant description of certain facts about the language. The main body of the paper is a thorough examination of the kinds of data that have been used in support of an oblique subject analysis for Old Icelandic, supplemented by some of my own additional data. It turns out that the set of subject properties of Old Icelandic is different (smaller) than that of Modern Icelandic, and the result of this examination is that Old Icelandic does not exhibit data that call for an oblique subject analysis. The final section of the paper offers an account of the diachronic process that may have led to the kind of structure that justifies an oblique subject analysis of Modern Icelandic. This process is a reanalysis leading to a change in the possible content of the Specifier position of IP, whereby it has become an exclusive subject position. Non-nominative NPs in that position may have kept their oblique case, and become oblique subjects.