No abstract

This paper e~camines the nature of scrambling in Japanese in the light of Webelhuth (1989) and Mahaian (1989). Webelhuth proposes that scrambling is uniformly movement to a third type of position, the non-operation/non-A position, and that this position has the binding properties of both A and A" (operator) positions. Mahajan does not recognize the third type of position, and argues that clause-internal scrambling can be either A or A" movement, while "long distance" scrambling is necessarily A" movement. 1 arguc in this paper that these two apparently inconsistent hypotheses are both necessary for the analysis of scrambling in Japanese.As evidence for Webelhuth's hypothesis, I show that unlike wh-movement, scrambling need not establish a semantically significant operator-variable relation. Then, I argue that Mahajan's hypothesis, based on the A/A' dichotomy, is also needed to account for the distinction between clause-internal scrambling and "long distance" scrambling with respect to anaphor binding. Finally, adopting proposal that non-operator/non-A positions are licensed at S-structure but not at LF, I suggest that a modified version of Webelhuth's hypothesis applies at S-structure, and Mahajan's hypothesis applies at LF.

This article proposes the following mechanism of Merge, modifying and incorporating the effect of the head parameter: K = {γ, 〈α, β〉}, where γ ε {α, β} a. γ = α: head-initial, left-headed b. γ = β: head-final, right-headed It is argued that under the parameterized version of Merge, traditional “adjunction” operations (scrambling and heavy NP shift) are characterized as substitution in the sense that they always accompany the projection of the target, whereas traditional “substitution” operations (wh-movement and NP-movement) are analyzed as genuine adjunction. It is then shown that numerous empirical consequences follow from this theory of phrase structure and movement, including (a) the nature and distribution of optional movements, (b) an elegant account of some peculiar properties of specifiers, and (c) a new unification of adjunct and subject condition effects, with a natural explanation of the parametric variation associated with the latter effect.

No abstract

The star formation rate in spiral galaxies is considered to be decreasing continuously with time in a time scale of 10 9 yr. The present star formation activity, on the other hand, occurs in molecular clouds with a time scale of 10 7 yr, and shows various degrees among galaxies. We make a new data set of 1681 nearby spiral galaxies from available databases and study the statistics of the present star formation activity. We analyze far-infrared and optical B-band surface brightnesses of the H II regions and the non-H II regions in M 31 and show that a far-infrared-to-optical B-band flux ratio, f FIR /f B , is a useful indicator of the present star formation activity of spiral galaxies. For the sample galaxies, we make the distribution diagram of log (f FIR /f B ) versus log L B for each morphological type. The distribution of f FIR /f B has a dispersion of one to two orders of magnitude even within the same morphological type of galaxies, implying that the star formation activities of spiral galaxies changes discontinuously in a short time scale. Analyzing the log (f FIR /f B ) versus log L B correlation, we suggest that the most active star formation in galaxies does not continue longer than 10 8 yr. We also construct a universal distribution histogram of log (f FIR /f B ) for each morphological type. The earlier-type spirals tend to show larger variation of the present star formation activity. We discuss the correlation between the present star formation activity and the structures and environments of galaxies. We suggest that the short-term variation occurs primarily due to internal processes which may change with the morphological type of galaxies.