Tuesday, December 04, 2007

Toki Pona and Minimal Vocabulary

The other day, Edward of Tower of Confusion ran across the author of Toki Pona, an interesting seeming constructed language. As a general rule, I'm dubious of conlangs: Language is organic, a thing that arises and evolves to meet the needs of a speech community. Even if you started with a conlang, for it to truly take off it would almost have to evolve beyond the intentions, even potentially in opposition to the intentions, of its creator. (See the Omniglot's discussion of EFL today for an example of how a language can get away from those to whom it seemingly belongs.) However, I do like thought experiments, and so I wandered over to the Toki Pona site and have done most of the lessons at the linked site (the official Toki Pona site's language lessons are being revised). A couple of observations:

1) The language does feel a bit more natural than others by virtue of the fact bits are a touch arbitrary, like not using "li" (subject marker) after "mi" - I and "sina" - you, even though they're used after everything else. On the one hand, this makes sense - if you and I are talking, we can clarify confusions since we're both present so the same level of formal grammatical distinction isn't necessary. On the other hand, it breaks a pattern. If we all started talking Toki Pona tomorrow, I think there's a good chance this rule would fall by the wayside. Ditto for only using "en" - and - in subjects, but not for objects (where the object particle "e" is put before each direct object and "tawa" before each indirect object if I read it right). In a way, this is nice, though, because forcing the learner/speaker to stick to these rules prompts new ways of thinking about language and clarifying exactly how the elements of the sentence relate. In other words, as a thought experiment, Toki Pona is great for thinking about language in new ways.

2) The really neat thing about Toki Pona is its vocabulary: There are only 118 dictionary words. And you can do a lot with them. Notes Edward:
It is designed to be a simple language with simple vocabulary. Yet it turns out that with such a small set of vocabulary, it is quite sufficient enough to express a lot of complicated ideas. In fact, when Sonja created the language, she wondered why the vocabularies in our natural languages have to be so complicated.
He later asks:
So why do we need such complicated vocabularies in our languages. This just made life difficult for language learners and lovers like us.
This is a question that often troubles me, and I'm forever looking for that tiny nucleus of a language that will be enough to get by with. But as I said before, language is organic. It grows the way it grows. And the way it grows is to meet the needs of those who speak it. If people and societies could constrain themselves for going in new directions, doing new things and dreaming new dreams, language could be greatly simplified and standardized. But whenever we change or alter our frames of reference, language shifts because not only are we talking about new things, but we are living in worlds different from the worlds our ancestors inherited when they worked their language into a form that worked for them.

Someone who has never studied Mandarin before and picks up a book on beginning to write the language could be forgiven for thinking all words are one syllable and represented by one character. But take the word for train, "huoche." It means "fire-car". Fine, you say, it's a nice little word that does the job. But here's the thing: If you look at someone and say "hen duo ren de che" - many-people-car - instead, people aren't going to nod in recognition. They're going to stare at you a little wide-eyed, then assume you're a simpleton who hasn't yet learned the word - and it's one word, thank you - huoche. That is, huo, che and huoche are three distinct words, just as surely as are rail, way and railway.

In Toki Pona, we are told, there is no word for friend. Rather, you say "jan pona" - good person. But say you want to mention your friend's house. The word for house is "tomo." To say "my house," you say "tomo mi." But to say "friend's house," it's "tomo pi jan pona" - with the "pi" to show that the "jan pona" goes together. Otherwise, someone might decide, eg, that "tomo jan" means people house, decide that can stand for hotel, and then mistakenly think that "tomo jan pona" means "good people house" or "nice hotel." These examples are, of course, conjecture. Being new to Toki Pona, I'm unaware of a compound, "tomo jan," but one might exist. The point is that once these compounds start to standardize, we find that Toki Pona may have 118 words on its own terms, but it's got a lot more lexical items. And once people start automatically putting things together, the game's up on simplicity because all new speakers then have to learn that of course "jan pona" means "friend" and not just, eg, nice person.

* * *

It's my intent to keep looking at Toki Pona, and I'm eager to learn enough to start reading through the blogs and cartoons with it. And it's given me some new ideas for continuing my quest to figure out just how little of a language you can learn and still know enough to communicate in it. But at the same time, my few days with Toki Pona have given me, I think, an answer to Edward's question about why we have to have all these words that make life so hard on language learners - we need them! For language to approximate life to a reasonable degree, it's going to have to be approximately as complex (or is "complicated" the better word here?).


Blogger goulo said...

A lot of people (half jokingly?) say toki pona is very easy to learn because "it only has 118 words". I have found it's not easy, partly because of exactly what you say: there are conceptually many (compound) words, even though there are only 118 core roots.

Still, it helps that the word for friend has some reasonable relation to the roots (jan pona) instead of being an arbitrary new root. In that sense, it is probably easier to learn than most languages. Constructing words out of a core set of roots seems nice, even if sometimes the compounds are mere hints to the meaning. toki pona probably takes the number of roots to such a small extreme that it may be counterproductive in terms of ease of learning them, since each root has so many possible interpretations; I'm not sure.

I heard an interesting discussion on this theme about Esperanto, which has a similar advantage, e.g. that from the root lern- we have "lerni, lernanto, lernejo" which are more evidently connected than "learn, student, school", even if "lernejo" could theoretically also mean a room in your house where you study instead of a school. I.e. from first principles it may not be apparent that lernejo=school, but it's a pretty reasonable bet, and the meaning of "lernejo" in context will be clearer more often than the meaning of "school" would be for someone who doesn't know the word "school".

12:42 AM  
Anonymous Kelly said...

Like yourself, I'm not too keen on conlangs (even the more 'official' ones like Esperanto) but I found that site very interesting. I like how they derived words from so many different sources (not just the major languages) and how new words are created. The fact that the language uses so few basic words and manages to create a fairly wide vocabulary is quite impressive.

1:22 PM  
Blogger gbarto said...

It is indeed nice how many sources Toki Pona draws on. It really does feel international, especially given how few words there are. And there's something to be said for how large a vocabulary it allows given how few building blocks there are.

The thing about conlangs, though, is you're left wondering why it was taken as adequate to make some things out of building blocks while others were given their own independent word. When you're making a conlang, do you really need good/bad, hot/cold, tall/short, etc, or is it enough to have the affirmative concept and a negative form? How do you decide which need independent expression and which don't? Is evil good's opposite? its absence? a phenomenon in its own right? And can you make a conlang that doesn't make its speakers think like its creator unless you let such distinctions multiply, thus disrupting the simplicity?

One big thing I will say for Toki Pona: It's not a language for universal communication (Esperanto) or logical thought (Lojban). It seems to be a thought experiment for thinking about how we perceive the world, how we use language to describe it and, frankly, how we can overdetermine our world by the endless creation of lexical items that let us name things without truly considering what they are. Taken on these terms, both its strong points and its weak points turn into strong points. It's fun and interesting what comes up when you apply it to thinking about/talking about something new.

7:50 PM  

Post a Comment

Links to this post:

Create a Link

<< Home