Discursive Logic
On top of the formal ingredients, we should imagine that most conversational Lojban takes place in a discursive modal logic, not Loj or Selb. This is where words like {lo} and {le} can be properly defined. Consider:
lo mlatu ku mlatu
This bridi does not have explicit quantifiers. Whether it is true depends on which universe of discourse we use to interpret it. For example, the empty universe should, in some sense, not satisfy it, because {lo mlatu} should refer to something which the speaker doesn't have in mind yet but which still exists. The act of clarifying referents leads to argument over the contents of the universe of discourse, and so deserves its own treatment apart from mathematics and hard sciences.
Restrictive and non-restrictive clauses
It's not clear what the difference between {noi} and {poi} is, but it is a discursive difference: there is no change in logical content, but something is denoted to listeners.
Modals
Historically assigning any semantics to {fau} is difficult; what is the difference between this utterance:
broda fau lo nu brode
And this utterance:
broda .i lo nu brode ku fasnu
We cannot rely upon any open sumtcita to help, since {fasnu} is unary. Perhaps there is a discursive difference; perhaps the former utterance also implies some connection or relevance between the two bridi. To be clear, the latter utterance also has an implied connection, the "story-time convention", as manipulated with {ki}.
Signs: sinxa, la'e, lu'e, lu'i
As in CLL, we can define {la'e}
and {lu'e}
as shorthand for {sinxa}
.
Note that symbols and interpretations are things, and thus are subject to
quantification.
ro da poi sinxa ke'a zo'u:
lo sinxa be da ku broda de
========================== (lu'e-intel) [CLL 6.10]
lu'e da lu'u broda de
ro da poi ke'a sinxa zo'u:
lo se sinxa be da ku broda de
============================= (la'e-intel) [CLL 6.10]
la'e da lu'u broda de
I recall discussion about whether sinxa3 is implicitly {mi}
or {mi'a}
or
some similar connection to the speaker. However, I am not sure whether this is
a cultural convention, and at any rate, leaving sinxa3
unbound allows for more
flexible usage.
Additionally, {lu'i}
and {lu'a}
can be defined as {lo'i}
-like expansions
of {sinxa}
which include a set quantification. These definitions are only
hinted at in CLL, and were explored by la xorxes.
ro da zo'u:
lo sinxa be lo cmima be da ku broda
=================================== (lu'a-intel)
lu'a da broda
lo sinxa be lo se cmima be da ku broda
====================================== (lu'i-intel)
lu'i da broda
Ellipsis: zo'e, co'e, do'e
An ellipsis represents a selbri or terbri which the speaker has not yet
uttered. As such, they should be registered as discursive evidence, rather than
given semantics like {zi'o}
.