Show description

Notice that correspondingly, in the present case, *The car broke the window with the lampstandard is unacceptable, since, on Fillmore's account, the car cannot have been derived by promotion in this instance. e. in which Caruso originates as an adnominal within the instrumental. But (55) could also within Fillmore's 40 Grammars of Case framework be an instance involving an underlying A (with a co-referential adnominal within the instrumental). (55) thus has two possible underlying representations, abbreviated in (56): (56) a.

All this is in conformity with the generalisation just formulated. The question now arises: how is this generalisation to be captured? ) and the semantic properties of the individual lexical items, the generalisation has an obvious embodiment: the passive transformation. (60) and (62) and (61) and (63) have respectively (at least partially) in common a deep structure in which the superficial subject and object of the active versions are the deep subject and object respectively. a). a). But the framework of 'The case for case' does not include such deep structure configurations and such a passive transformation.

This would mean that the A/imperativisation correlation can no longer be observed: only some A's allow imperativisation. But this is not surprising in view of the pragmatic character of whatever determines 46 Grammars of Case both what can be an ergative and what can be imperativised. Alternatively, in the spirit of the proposals we have just made, the 'Force' instances may represent a node with multiple CRs, specifically perhaps an ABL(ATIVE) (Kirkwood's source/cause) and an erg, which conjunction also seems appropriate with the subject of a verb like transitive benefit, and, if we follow a suggestion of Huddleston (1970), perhaps 'unintentional agents' .