Exercise 8.6 A. Here are the structures for (1a) and (1b):
B. The nominative case on "Emrys" in (1a) and (1b) is checked by the tensed verb in I under the head-spec configuration. The objective case on "ddraig" in (1a) and on "Fangor" in (1b) is checked under the head-comp configuration. The objective case on "Megan" in (1b) appears to be checked in the head-spec configuration by the preposition "i," which seems to be acting like the English preposition/complementizer "for" that occurs with infinitives. This anlysis implies that the subject of an infinitive moves to Spec,IP even though the subject of tensed verbs does not and that this movement occurs for case-checking.
C. One reason why verb-raising would not be lost in Celtic, as opposed to Mainland Scandinavian, might be that the evidence for verb-raising in Celtic is so strong. In every tensed sentence the verb occurs before the subject and if there is a object, the subject occurs before the object. This word order requires verb-raising, given our other assumptions about phrase structure, so the Celtic child learner has a large amount of evidence for verb-raising. Because Mainland Scandinavian is SVO and V2, only subordinate clauses with pre-VP adverbs provide evidence for V-to-I raising. This amount of evidence was apparently not enough to guarantee that children would learn this syntactic property when the morphological evidence disappeared.