Disclaimer: This post does not discuss a natural language, instead this post is about a constructed language, commonly contracted to conlang. You may have heard of Tolkien’s Quenya, Star Trek’s Klingon, and George R. R. Martin’s Dothraki and High Valyrian, and these are all examples of a conlang.

The question begs: What is the minimum number of root words which are needed to represent simple ideas in a language which are culturally universal? The answer, according to a translator and linguist back in 2001, was about 120 root words. Their name was Sonja Lang. Designed to shape the thought processes of its users, Toki Pona aims to express maximal meaning with minimum complexity, drawing inspirations by Taoist philosophy and in a Zen-like manner.

Toki Pona consists of only 14 phonemes, split into nine consonants (k, l, m, n, p, s, t, w, j) and five vowels (a, e, i, o, u), where the ‘j’ sounds like the letter ‘y’ in English. Syllables are all in either V, CV, VN (N stands for final nasal), or CVN. What I can compare this conlang to in terms of phonemes and syllable structure is the Polynesian languages like Samoan, Hawaiian and Tokelauan.

In terms of root words, Toki Pona draws influence mainly from English, Tok Pisin, Finnish, Georgian, Dutch, Acadian French, Esperanto, Croatian, Mandarin Chinese and Cantonese. Many of these derivations are transparent, as Toki Pona root words sound like cognates of their respective words in the languages mentioned. For example, the Toki Pona word for eye, oko, is identical to the Croatian word oko, Spanish ojo, Italian occhio, and English ocular.

Numbers in Toki Pona are interesting, mainly due to it only having three words for numbers; one is wan, two is tu, and many is mute. Formerly, the word luka (literally hand) is used to represent five, and even though it was deprecated in the latest official version of the conlang, its use to represent the number is way more than its use to represent the original meaning of hand. Ala, the negation particle, is also used to represent zero. So, to make higher numbers, the root words are used in combination. For example, to say four, we say tu tu. Knowing this minimal system, it gets impractical to express larger numbers, like 483. I have a recent version of the ebook which contains a rather hilarious paragraph supporting this impracticality, I quote “However, Toki Pona is simply not intended for such high numbers. … If you can’t cope with not being able to talk about the number 523 in a language, then Toki Pona is not for you. Maybe you’d be happier with Latin.”

The grammar of Toki Pona is simple yet extremely ambiguous. In fact, Wikipedia has summarised its syntax in just ten rules, and that was 54 pages of ebook excluding the appendix. These are:

Syntactic rules

1. A sentence may be

(a) an interjection

(b) of the form [sub-clause] [vocative] subject predicate

(c) of the form [sub-clause] vocative predicate

(The interjection may be a, ala, ike, jaki, mu, o, pakala, pona, or toki.)

2. A sub-clause may be

(a) [taso] sentence la

(b) [taso] noun phrase la

(“If/during sub-clause, then main-clause“)

3. A [vocative] is of the form

[noun phrase] o

4. A subject is of the form

(a) mi or sina

(b) other noun phrase li

(mi mute and sina mute require li to form a predicate.)

5. A predicate may be

(a) simple noun phrase [prepositional phrase]*, or

(b) verb phrase [prepositional phrase], or

(c) predicate conjunction predicate (that is, a compound predicate)

(The conjunction may be anu (or) or li (and). The latter is merely subject li predicate 1 li predicate 2 etc., with the first li dropping out if the subject is mi or sina)

6. A noun phrase may be

(a) noun [modifier]*, or

(b) simple noun phrase pi (of) noun plus modifier* (if there is a modifier which only applies to the 2nd noun), or

(c) noun phrase conjunction noun phrase (that is, a compound noun phrase)

(The conjunction may be anu (or) or en (and). A ‘simple’ noun phrase is one which does not have a conjunction.)

7. A prepositional phrase is of the form

preposition noun phrase

8. A verb phrase may be

(a) verbal

(b) modal verbal

(c) verbal_x ala verbal_x (both verbals are the same)

(d) modal_x ala modal_x plus verbal (both modals are the same)

(The modal may be kama (coming/future tense), ken (can), or wile (wants to).)

9 A verbal may be

(a) verb [modifier]* (this is an intransitive verb)

(b) verb [modifier]* plus a direct object* (this is a transitive verb)

(c) lon or tawa plus a simple noun phrase

(Some roots may only function as transitive or intransitive verbs.)

10. A direct object is of the form

e simple noun phrase

Yes, this is basically all there is in Toki Pona grammar. With so few rules and so few root words, it is not unsurprising ambiguities pop up most of the time. For example, consider the sentence “mi moku”. It can mean “I eat, I ate, I will eat or I am food.” The lack of tenses increases the ambiguity, on top of the lack of the verb to be. Thus, a simple “mi moku” can have at least four ways of interpretation.

What happens when one wants to express slightly more complex words like “to cry”? They compound the root words, of course. In this case, “to cry” is “pana e telo oko”, quite literally “to let out water of the eye”, which pretty much conserves the meaning of the original word. When forming such compound words, it is important to preserve the generic idea of the original word, and adhere to the grammatical rules which apply in Toki Pona. In fact, Toki Pona itself is a compound word, translating to “the language of good”, “to say nice things”, “speak well” etc.

The only words which have no way of using compound words to express are names and languages. In which case, separate rules apply to make these words Toki Pona. Phonetic guidelines, and Tokiponization guidelines are all found within the ebook. Thus, the Toki Pona word for Britain is Piten, Finland is Sumi (from Finnish word for Finland Suomi), and Egypt is Masu (from Arabic word for Egypt miSr). Note that when typing this language out, only the names are capitalised, all other words are in lowercase.

This language, to conclude, is simple with good reason. Being able to express concepts in the simplest and most minimal means possible, albeit extremely vague, is one question begging an answer, and so far, this has been the answer. At just about 120 root words, Toki Pona is able to express the simplest of concepts and ideas known to humankind. Speaking outside of simplicity and practicality and lack thereof, Toki Pona tested the principles of minimalism, the Sapir-Whorf hypothesis of linguistic relativity and pidgins. Toki Pona has garnered a small following since its inception, with 100 claiming to be native speakers in 2007, and today, well, I do not have reliable statistics for that.

For our next post, we will be discussing a conlang which sits on the other side of the spectrum of ambiguity and complexity, first publicised in 2004.