krtisfranks wrote: > there may be complications to the way that Lojban handles lerfu which I > have not previously considered (such as an analog to the difference between > PA and numbers).
One such complication is the fact that concatenation with any immediately previous lerfu string which is unconverted will occur automatically, according to my current mental model of the language and how this word should work (which is derived by analogy from the same model for xo'e and xo'ei which I have constructed). I do not see this as too much of a problem though because: On its own, this word would just potentially refer to any context-allowed lerfu (producing a 1-string which then is interpreted as referring to something, namely an operand, much as xo'e produces a numerical 1-string which is then understood as a (Lojbanically single-digit) number). But if there is an open letteral immediately preceding it, then it just concatenates into that, thereby forming an (n+1+m)-string which will then be interpreted to have a single referent (so, still an operand, in this intended context); in this way, the elliptical letteral is just 'modified' in its effect so that the pool of referants of the whole string is restricted to the possibilities such that the potential referents are named by strings of length n+1+m with the first n and the final m letters matching the explicit ones specified by the resulting string (much as "ci pa xo'e" produces, after interpretation, a single number which belongs to a pool of potential referents such that all of them are three digits long in the base, the first two digits are "3" and "1" in that order, and the last digit has free variation bounded only by context and the restrictions of the first two digits (so, if "317" is not an option for some reason, then the number cannot be called such)).
A potentially bigger complication is the fact that operators and all sorts of other things (even in the context of mekso) can be labelled by letters or can have lerfu strings refer to them. For example, a function f is of the first kind, but su'i (addition) may be referrenced by the lerfu string sy ("s") and thus sy/s in combination with su'i is of the second kind. This can potentially cause unintended consequences, confusion, incorrectness, or could incur a typing error. I do not think that this is resolved by the fact that ellipticals mean what the speaker wants (or, really, what the audience thinks that they want) them to mean. I think that it would be a fundamental flaw in the grammar simply because the elliptical lerfu could mean, say, another operator instead. In that case, there would be nesting an the second (potential) operator (the one referenced by the elliptical lerfu) would have implicit elliptical operands. The functionality of possibly referring to an operator or, indeed, anything which may be labelled in mekso, is necessary/useful and guaranteed by the grammar thus far, so we cannot avoid it here by artificially imposed restrictions (admittedly, operators usually have some sort of modifier which makes them clearly operators, but that is not sufficient help, I believe). Thus, I think that this is a deal-breaker. We need another word for elliptical operands, and a separate one for elliptical operators.
|