>> Does this seem like a reasonable approach?
>
>Well, yes, as far as it goes. The problem is that (a) the notion of a
>direct object is so indelibly linked to an accusative case that it may be
>difficult to get people thinking about a direct "object" in the dative or
>genitive case--and it does sound a bit odd;
My Real Live Linguist^TM source says that most objects are marked for accusative, but some are marked for other cases. That doesn't sound that hard to explain.
>and (b) "object" disguises the
>fact that the grammatical function of the word used as a direct complement
>of a verb is structural and bears no necessary relationship to external
>reality.
True.
Jonathan