coi xirma, doi xirma
From Jboske:
CLL explicitly states that {coi xirma}, {doi xirma} are ambiguous between {coi/doi la xirma} and {coi/doi le xirma}. Since these two are nonequivalent, and since Lojban generally is not ambiguous (it claims that it is syntactically unambiguous but not semantically unambiguous, but in practise it strives to be semantically unambiguous and its claim is meant to be that it is not free from vagueness), should we see this is an error or brokenness in CLL?
It's really an issue about whether CLL should be allowed to subvert the underlying principles of the language, since one can disambiguate with {coi/doi la xirma} and {coi/doi le xirma}.
There is a further wrinkle, though, which is that I take {doi/coi le xirma} to mean "I hereby address/greet a certain horse". (It's a silly meaning, but not a silly construction: I might stand up before a class of students, wishing to address certain ones of them, and say {doi le tadni}.) CLL glosses {coi xirma} as "hello horse". But {coi la xirma} would be "Hello Horse", while "hello horse" would be "coi do noi ke'a xirma". So CLL seems inconsistent and hence 'broken' (assuming that consistency is necessary condition for nonbrokenness).
I therefore think that {coi/doi xirma} should be equivalent only to {coi/doi la xirma}, which also makes for a simpler rule: la/lai can be omitted following DOI/COI.
I've never thought about this much, but I agree on the basis of "coi djan" being at least semantically equivalent to coi la djan --greg.
So "doi pendo" is not "friend" ("you, person, who are my friend") but "Friend" ("you, whom I am giving the name 'Pendo'")? --la .filip.
The argument is symmetry versus usefulness. As Philip just hinted, the useful default of doi is doi le. I agree that it does have to be one and not the other, and vote for the less symmetrical but more useful doi le. If you want a cmene, you know where to find them. I'm not even convinced all Lojbanists realise we can have names in Lojban which are bridi rather than morphological cmene (cmeseltai), and they certainly aren't prototypical cmene anyway. -- nitcion
I agree about choosing the most useful, but the most useful interpretation of doi pendo is doi do noi pendo, not doi le pendo. In the case of doi le pendo the hearer first of all identifies the referent and then understands that the referent is being addressed. --And Rosta
*shrug* as long as you don't want the default to be doi la pendo, we are not in real disagreement -- nitcion