Comment by sterlind | Hacker News Reader

Comment by sterlind | Hacker News Reader

Parent Full thread

Lojban is actually equivalent to first-order predicate logic, I believe!

for instance, "lo cmalu noltru (ku)" is equivalent to "there exists some X, such that X satisfies the predicate 'cmalu noltru', i.e. X is a small type-of prince."

unlike Esperanto however, lojban isn't likely to be teachable to children, or fluently speakable. all words in natural human languages seem to have a maximum of three 'arguments', i.e. parts of speech that valsi bind to. words in lojban may have up to five arguments, so it's unclear whether human brains can accommodate that many slots.

additionally, while lojban is LL-parsable, machines like Parsey McParseFace can now accurately parse human universal grammar. unambiguous parsing was a big feature of lojban, so it's lost a major selling point.

but! learning exotic languages is good for the brain, so it's a worthwhile and rewarding endeavor. I speak a conlang with only 2 speakers, and it has enriched my life, so tilt at those windmills! :)

I don't know very much Lojban but (while I agree that the grammar is meant to be a realization of a formal logic), I'm a bit unsure of your specific example with "lo cmalu noltru":

First, I think "lo cmalu noltru" is a sumti (akin to a noun phrase) and not a complete sentence by itself.

Second, I'm not positive about the existential import of "lo". There are many reasons for this, but to take a kind of extreme example, the reform of the meaning of le and lo at https://mw.lojban.org/papri/BPFK_Section:_gadri#cmavo:_lo_.2... presents the example

lo pavyseljirna cu ranmi danlu gi'e simlu lo ka ge ce'u xirma gi lo pa jirna cu cpana lo mebri be ce'u

'Unicorns are mythical creatures that look like a horse with a horn coming out of their foreheads.'

This doesn't seem to imply that there exists X such that X satisfies the predicate pavyseljirna. Or to take another example, presumably Lojban can express using "lo" the notion that "Counterexamples to Fermat's Last Theorem were never discovered prior to 1993, despite extensive computer searches".

I do think that many Lojban utterances have formal existential import in many cases, and I don't know enough about Lojban to try to describe which ones. :-) (I seem to remember something about a bare gismu implying that the predicate is at some occasion true of something, like "tavli" by itself without arguments meaning something like 'speaking, it's a thing that happens at some point!'. But given the example of talking about unicorns, maybe even this is too strong somehow.)

Edit: ... by the way, what's the conlang with two speakers that you speak?

I don't know any Lojban at all, but I do know formal logic and "there exists" (∃) is usually used to denote existence with respect to some universe of discourse, and not necessarily physical existence. Most mathematical concepts only exist as things that mathematicians talk about, not unlike Unicorns. However, I would read the example sentence as making a statement about all unicorns, of the form "∀ U ∈ Unicorns . U ∈ MythicalCreatures ∧ (∃ H ∈ Horses . looks-like(U, H)) ∧ ∃ H ∈ Horns . (has(U, H) ∧ ∃F ∈ Foreheads . (has(U, F) ∧ coming-out-of(H, F)))" or something to that effect.

The sentence about counterexamples to Fermat's last theorem could be written using an existential quantifier without problem, since it appears in a negated sentence, thus not actually claiming their existence, rather the nonexistence of counterexamples discovered prior to 1993.

However, looking at the definition you linked, it appears that "lo" does not have any specific equivalent in formal logic since "It converts a selbri, selecting its first argument, into a sumti. The resulting expression refers generically to any or some individual or individuals that fit as the first argument of the selbri." which seems to be mostly a grammatical function?

The meaning of "lo" was changed from the original existential meaning to its current grammatical meaning under the recent "xorlo" change, I believe.

https://mw.lojban.org/papri/How_to_use_xorlo

Thanks for the useful discussion on this point.

> machines like Parsey McParseFace can now accurately parse human universal grammar. unambiguous parsing was a big feature of lojban, so it's lost a major selling point.

Slow down there ... Most linguists (or a substantial minority) do not subscribe to the theoretical idea of universal grammar. Parsey McParseFace is only an incremental improvement in decades of statistical parsing. The problem of natural language understanding remains largely unsolved. Like other deep learning models, it is not hard to come up with adversarial examples that will confuse the parser but not any competent human language user. The parser is only as good as the data it is trained on; this data is expensive to acquire but there is never enough. Additionally there are a myriad other kinds of ambiguity in language beyond the sentence-level syntactic ambiguity which is resolved by a statistical parser such as Parsey McParseFace.

In my opinion Lojban provides a good illustration of just how hard it is to remove ambiguity from human language.

There are fluent speakers of Lojban as can be seen in Lojban groups on Telegram: https://t.me/joinchat/BLVsYz20Boixl0xN-0TrPw and https://t.me/joinchat/BLVsYz3hCF8mCAb6fzW1Rw

Reply on news.ycombinator.com