Some languages do not mark a distinction between referents a speaker has in mind. Others only mark it in special situations. Spanish, for example, makes a distinction in noun phrases with relative clauses:
"Me voy a casar con una mujer que tiene mucho dinero." (specific)
"Me voy a casar con una mujer que tenga mucho dinero." (non-specific)
Without a relative clause, however the distinction is unmarked:
"Me voy a casar con una mujer." (may be specific or non-specific)
(Note: My Spanish is limited and I owe this example to "A New Reference Grammar of Modern Spanish".)
A similar distinction is made in English:
"I intend to buy a car." (may be specific or non-specific)
And then either: "It's cheap and has low mileage." (disambiguated as specific)
Or: "It should be cheap and have low mileage." (disambiguated as non-specific)
In both English and Spanish, it is the non-specific case that is marked.
The English definite article, "the", often discussed in relation with specificity, subordinates the consideration of whether the speaker has referents in mind to the question of whether the referents are identifiable in context.
"I want to rob a bank." (may be specific or non specific)
Here the indefinite article ("a") indicates that, regardless of the speaker's state of mind, context is not sufficient to disambiguate the reference: There may not be an obvious referent, or there may be more than one.
It seems to me that reason that {le} is not often used is that the distinctions it bears have not, over the course of the last ten years, proven useful to speakers (and writers and translators) most of the time. If the distinction of speaker-oriented specificity were important, for example, and if the neglect of {le} were merely a matter of conformism, I would expect to see more locutions like:
{mi djica lo nu terve'u lo karce poi mi nau pensi tu'a ke'a}
I'm interested in knowing what other people think about all of this. (But less interested, perhaps, in whether or not they're currently thinking about it.)