lojban freaks like to talk a lot of shit about how their language is ~syntactically unambiguous~. and yes, lojban is syntactically unambiguous, in the sense that LISP is syntactically unambiguous: computers can parse it, and humans can't.
is there a way to bridge this gap? can we make a language that is syntactically unambiguous without being to the human mind what a demagnetizer is to a hard drive? and most importantly, is doing so even remotely morally defensible?
for the sake of fun, i'm going to ignore that last question. what follows is my answer to the first two.
Ne Ta-me ko is a language. it's a very tidy little language with a minimalistic grammar and a no-bullshit orthography. it's built to explore a single concept: center branching.
a "problem" with natural languages, or so the monolingual computer scientists behind lojban would have us believe, is the fact that many complex clauses are characterized by ambiguous constructions. for instance, in the English sentence "matilda hit the man with the chair," it is unclear whether matilda hit a man who was carrying a chair, or whether she hit a chair-underspecified man using a chair as an implement of violence. (this is also an example of semantic ambiguity, occuring mainly because the preposition "with" can modify both nouns and clauses, with different meanings, and offers no indication which it attaches to). this phenomenon creates such catastrophic ambiguity in natural languages that not a one has ever bothered to devise strategies to mitigate it.
lojban rectifies this situation by [editor: insert boring spiel about dumb nerdlang here]
we're going to solve it in a way that is both more aesthetically pleasing, and more easily parsed by the human brain. i think.
as the salient aspect of the language is its unique grammar, phonology is mostly irrelevant, but a reference pronunciation is given here for the sake of convenience. this should not be taken as an endorsement of attempts to speak this language.
|stop||p b||t d||k g|
|fricative||f v||s z||x gh|
consonants are evenly distributed between three places (labial, dental, velar) and three manners (nasal, stop, fricative) of articulation. stops and fricatives have an additional voicing contrast. alveolar fricatives palatalize (/s/ → [ʃ], /z/ → [ʒ]) before high vowels; alveolar stops become homorganic affricates. for clarity, syllables like ‹ti› or ‹de› may be written ‹tsi› or ‹dze›, but this is not obligatory. /n/ palatalizes as [ɳ] and /ŋ/ as [ɳ͡ŋ]
a key feature of Ne Ta-me ko is that all lexical items (and many more items besides) are split into two "tags." an example is the verb Ta-me, "speak", which is composed of the tags Ta and me. why is the capitalization signifcant? what do the dashes mean? we'll get to that in a moment.
these tags are inextricable aspects of the word itself, and cannot be separated into constituent meanings. they can, however, be separated grammatically. let's take a look at this in action. below, the syntax of a simple sentence is given.
you may notice some unusual notation in these phrase structure rules. we'll get to that in a moment. first, let's talk about semantics. the most basic form of clause is comprised of a verb phrase and, optionally, a noun phrase. Ne Ta-me ko is an ergative-absolutive language. The absolutive argument is coded within the VP; the ergative, if any, is placed after it.
"now hang on a minute," you may be rightfully interjecting, shaking your fist and pointing your finger. "what the hell is up with these brackets? this isn't a phrase structure grammar, this is gibberish!
let's take a look at some examples, dear reader, and see if you can't figure out exactly what those brackets denote.
are you beginning to get a sense of how this works? in the grammars of natural languages, constituents can only appear in sequence, leaving the depth and structure of their embedding up to inference, like a LISP without parentheses. in Ne Ta-me ko, however, both the end and beginning of each constituent is explicitly tagged, like HTML, except without redundancy — in "Ta Ngi-si me" above, neither "ta" nor "me" alone carry the lexical weight of "speak". if one were to instead say "Ta Ngi-si go", the meaning would instead become "despise English."
the purpose of the strange orthography should now become clear. an uppercase initial letter indicates an opening tag; a lowercase initial letter indicates a closing tag. a hyphen indicates an opening tag followed immediately by a closing tag. of course, none of these conceits are strictly necessary. one might as well write "ta ngisi me nga xova be" or even "ta ngi si me nga xo va be" and be equally well understood; however, these conventions provide a useful hint to the eye about the underlying grammatical structure, filling the gap left by the lack of tone in written language.
you can imagine any number of alternative orthographies, of course. one could substitute brackets for capitalization, writing instead "ta[ ngisi ]me nga[ xova ]be." one could even adapt S-expressions to the purpose, expressing the same sentence as "(ta (ngi si) me) (nga (xo va) be)". the reference orthography is advantageous, however, in that it requires fewer characters than either of these forms.
Ne Ta-me ko is a very flexible language, allowing the speaker to express only that which is relevant to the situation at hand. a sentence may be as simple as an unadorned verb, e.g. "Ta-me", meaning "speak". any and all arguments may be omitted if they can be inferred from context.
relativization is Ne Ta-me ko is very straightforward. a relative clause describing the noun phrase is simply embedded within the body of the noun.
the astute reader may note from (9) in particular that the relativization structure assigns no particular role to the noun phrase relativized. it may be ergative, absolutive, genitive, or even simply topical. while this may seem at first blush profoundly ambiguous, you will discover, as have the speakers of Japanese, that it is an extraordinarily flexible and expressive construct. the liberation of relativization from grammatical role affords us many simplifications of grammar, including the elimination of complementization as a distinct construct.
due to the nature of relativization, complementization can be accomplished with no further grammar. the clause can simply be placed within a generic word such as Ne-ko "thing", Dzi-tsi "event, happening", Gho-za "case, eventuality", or Fi-zi "problem, incident".
to be continued