Home
This dissertation demonstrates that it is not always the case that scope phenomena reflect the structural relation of c-command at LF, and establishes operational tests for teasing apart scope phenomena that are on the basis of LF properties from those that are not. It enables us to overcome (much of) the problem of judgmental fluctuation often observed in the discussion of scope phenomena in the literature, making it possible to seriously aspire to the attainment of repeatability in generative grammar, which has not been possible in the field for the principled reason that quite distinct types of phenomena have been conflated into one, not only in the area of scope phenomena but also in other areas having to do with interpretations.
Chapter 2 examines the scope interaction between a subject QP and an object QP in the configuration of [ … QPSub [ … QPObj … ]], where the QPSub and the QPObj are clause-mates, and demonstrate that the inverse scope obtains only if three conditions are met, but the availability of the surface scope is not subject to such conditions. Chapter 3 argues that the surface scope may emerge on the basis of LF properties while the inverse scope does not, and that the latter requires both the QPSub and the QPObj to be in A-positions at LF while the former does not, supporting the view that a QP may or may not undergo covert movement. It is concluded that the inverse scope involves an extra-grammatical operation (= MINOR), and the characteristics of the inverse scope observed in Chapter 2 are attributed to MINOR.
Chapters 4-5 further confirm the conclusion reached in Chapter 3. Chapter 4 establishes independently one of the implications in Chapter 3 that a wide scope reading of a QP a over a QP b can emerge on the basis of LF properties only if a c-commands b prior to covert movement (cf. Huang 1982). Chapter 5 presents further evidence for the two sources of scope interaction, based on the scope interaction between a QP and a wh-word. In particular, it is argued that functional reading may emerge through LF compositional computation while pair-list readings must be due to MINOR.