Pragmatics of LF intervention effects: Japanese and Korean Wh-interrogatives

This paper presents a pragmatic account for what have come to be known as logical form (LF) intervention effects, based on a wide range of data from Japanese and Korean. Despite their appearance, these effects are not due to structural constraints operative at LF but rather to a less-than-perfect realization of the information structure of interrogative sentences. The potential interveners, which seem to be a random collection of various expressions, are classified as Anti-Topic Items, since they cannot bear the topic marker - wa/(n)un. Although the non-Wh material in a Wh-question must belong to old information, the interveners fail to be interpreted as background material because of their Anti-Topicality when they precede Wh-phrases. The cancellation of the intervention effects with scrambling is derived from the prosodic phrasing that scrambling creates. Moving a Wh-element over an intervener places the intervener in the position of post-focus reduction—a prosodically reduced portion of the sentence. With post-focus reduction, the intervener becomes a part of old information. It will be shown that the proposed analysis not only accounts for core cases of intervention effects, but also makes correct predictions concerning the matrix-subordinate contrast in intervention effects and the special status of negative polarity items (NPIs).