Grammar GeekDawn, I'm sorry, but I can't follow you. What do you mean, a "nonsensical direct-object relation"?
Hm, my post was quite confusing, wasn't it? Common problem with me.
First, I agree with Nona and you about "wife" being perfectly acceptable. I also think the "Mum" example is a good one, but it doesn't go all the way. Nona said:
Nona the Britif I talk
about my mum when she was a little girl, I still call her my mum, even
though she clearly wasn't a mother then. If I am talking about
someone's wife or husband, I still call them that, even when we are
discussing a time when they weren't.
That's very true. But in the case of "married his wife", there is an additional factor, one that the time-tense structure Nona was talking about overrides. And this is the problem, I think, that Yoong Liat and Feebs might have (though, of course, I can't know for sure).
It's the act of marrying results in the status wife. It's not, I think, that she wasn't the wife then; I doubt Yoong Liat would have problems with sentences like "My wife was born in 1971." The problem is one of logical relation, rather than time. If the current reference ("I'm talking about my wife.") does not override the verb-object relation (verb causes status implied by object-noun), the effect of "I married my wife," can be akin of looking at one of Escher's paintings.
Actually, now that I thought about it, it's the other way round: the "verb-object-relation" doesn't override the reference system, even though analysis might suggest to some (as is demonstrated in this thread) that it should. The question, "Why doesn't it?" is interesting (probably something to ask cognitive linguists, rather than grammarians, but interesting, nevertheless.)
That probably doesn't make any more sense than my other post. 