I recently discovered toki pona on account of a blog I’m subscribed to. I started to study it imagining that for a small amount of effort, I’d be able to read and write yet another foreign language, which I would find amusing. So far, I can read a little-little bit, got in an argument on a mailing list, got in an argument on my blog and now I’m wondering how I got off on the wrong foot here.
Religious and Grammar Wars. Constructed languages suffer from religious wars and grammar wars in a way that natural languages don’t. (Natural language instead suffer from real wars, but that isn’t what I’m interesting in talking about) A constructed language, almost by definition is incomplete. You can’t resolve questions of semantics or grammar.
An argument about the semantic intent of a natural language can be resolved by taking a survey, video taping speakers using their language in a natural setting. Semantics in a constructed language are resolved by translating them into a natural language and then arguing that your translation is better than the next.
Questions of grammar, likewise is less resolvable than in a natural language. Linguists have been trying to reduce speech to transformational rules, sentence diagrams and phrase rules since forever. While it is possible to reduce a language to a machine validatable language, not unlike an algebraic equation or a C++ program, you generally can’t do the same for natural languages. Even MS Word can’t figure out if your English sentence are grammatical. This suggests to me that a language, constructed or not, once it is spoke can be expected to be used in ways that are not neatly machine parsable. And without a system of proving grammatical correctness, the arguments continue.
Is there anyone who can say what is right and wrong? Maybe Zamenhoff’s ghosts, Marc Okrand or Jan Sonja?
Intellectual Property & Control. A constructed language is something of an artistic work. However, it’s already been tested in court with Tolkien’s constructed language isn’t something you can’t easily stop people from speaking in, writing about, writing in, etc. Similarly, programming languages don’t seem to be protected–companies create competing compilers for the same language without fear of reprisals and some companies even give the specifications of their languages to standards bodies. Inventors of programming language know they can’t dictate the grammar, so they turn it over to a neutral party that will promote consistency, which may or may not be followed. Nonbinding standards have a way of being like that.
Given that a language inventor can’t control the language or how it is used, the language inventor is the arbiter only to the extent that they are famous, have time to adjudicate questions, and still are alive and can invest the marketing effort to sell the community on the idea of one form of consistency over another.
Lack of Expert Models. In most cultures people start writing at a college level somewhere between high school and college, about 10-15 years worth of reading and writing. Learning a language can take anywhere from a few to ten years before one achieves a level of competence that is indistinguishable from natives. A constructed language is no different. From inception, it will be a decade before there are natural models to study. Until then, we are all speaking, reading and parsing clumsy, broken toki pona, [those with extremely high opinions of their linguistic talent excluded, of course]. Alas, a whole decade to wait before the first Nobel prize for literature written in toki pona will be awarded.
So what to do? I don’t know about you, but I’m going to try to shift from a criterion of correctness to a criterion of beauty, consistency with the gestalt of the original design and euphony– all of which are criterion that no one expects one to the next to agree upon. If, after reading and writing a large quantity of toki pona, I come to hear in my mind the euphony of “mi lukin lon e ilo” as compared to “mi lukin lon ilo”, then the former is what I will use. The brain’s sense of euphony trumps arguments based on machine parsable grammars.