Okay, thinking deeply about recursion in toki pona. I've seen in at least two places that it is claimed recursion is a defining characteristic of human languages.
http://en.wikipedia.org/wiki/Recursion# ... n_language
Ignoring Everett and Piraha--who I think is a crypto-racist who'd be surprised if the Piraha could wipe their asses without a Portuguese tutorial--all human languages recurse and this is one of the stumbling blocks that keep all species except humans from speaking human like languages. Toki pona almost doesn't recurse (I think jan Kipo would say by the phrase grammar it doesn't recurse except in la-phrases). We seem to routinely accidentally recurse (pi chains where the 3rd modifies the second), la chains that are not AND's of the list of "la-conditions", prepositional phrases that modify nouns in another prepositional phrase. [Am I missing any?]
I've found it more productive to think of toki pona as a series of styles of speaking (there is jan Pije style, jan Sonja style, jan Wiko style, jan Kipo style, jan Mato style, etc). I'd be hard pressed to pull quotes to prove everyone has been recursing, but we probably all have--our brains are wired to recurse effortlessly. Even if I did pull quotes, a sentence built with recursion can be parsed without recursion to yield a different meaning, b
It seems that there is a re-cursion free style of toki pona and this language would be a candidate for animal language studies. It has less than the magic ~200 words and doesn't recurse. 200 words is about the typical upper limit vocabulary for dogs, animals and abused children that have past their 15 birthday without exposure to speaking humans.
Anyhow, it's too bad we don't have any Neanderthals anymore, because the current research on parrots and chimps is rather pessimistic for teaching them any language, where as there were probably several species of apes that spoke languages somewhere in between full blown recursive human language and animal signaling.
Useful phrase for the cat
soweli pona!
soweli o awen lon ni.
soweli o pana e mu wawa!
soweli o pana ala e ko jaki lon supa anpa pi tomo mi.
tu la soweli mi li wile la wile lukin e sitelen tawa pi jan Harry Potter?
soweli o moku ni li tawa mani mute. mi li esun ala e mana tawa moku sin.
lemme know if I have any accidental recursion in there, I don't want to be unfair to the cat
The Toki Pona for Animal Communication Thread
The Toki Pona for Animal Communication Thread
Last edited by janMato on Sun Jan 24, 2010 7:48 pm, edited 1 time in total.
Re: Recursion-- Could parrots,say, speak toki pona?
Well, more recursions than just in 'la' if we allow noun or PP in modifier. So there are some. But the Hell breaks loose when you recurse on S in the middle of things. And we don't need to (we don't even need sentential 'la' as NMS shows). In any case, tp could work without recursion (as would any language, but it is harder to drop when we're used to suing it), so it might be handy for experiments.
sin la soweli mi li wile ala wile (an AHA moment == not a missing subject but a missing 'a').
More useful cat phrases
soweli o kalama ala
soweli o, sina ken ala tawa insa ala
soweli o weka tan sinpin mi
soweli o tu ala e noka pi supa mi (recursion -- indeed all complex modifiers are recursion)
soweli o, tenpo moku li lon ala kin
(Cat food doesn't go to much money, it comes from it -- I think).
soweli o anpa tan supa
sin la soweli mi li wile ala wile (an AHA moment == not a missing subject but a missing 'a').
More useful cat phrases
soweli o kalama ala
soweli o, sina ken ala tawa insa ala
soweli o weka tan sinpin mi
soweli o tu ala e noka pi supa mi (recursion -- indeed all complex modifiers are recursion)
soweli o, tenpo moku li lon ala kin
(Cat food doesn't go to much money, it comes from it -- I think).
soweli o anpa tan supa
Re: Recursion-- Could parrots,say, speak toki pona?
Subset of Toki pona for animals
Since I've no one to talk tp to, I've resolved to start speaking tp to my cat. From time to time I speak to my cat in English, so why not in toki pona. But first, I thought I should give him a chance at understanding me by simplifying things further.
If animals can only understand symbols (ie. can't break sentences into parts), well this is waste of time. But what if they are only missing some of the machinery of language?
S1 = [o] n [a pi n a] li [prep] n [a pi n a]
s2 = [o] n [a pi n a] li v [e n a pi n a]/[prep n a pi n a]/[nothing]
s3 = [o] n [a pi n a] la S1/S2
where [ ] means optional and / means "on or the other"
Sentences of maximal length.
Sentence diagrams of maximum depth.
No recursion. Recursion is most obvious in [n1 a pi n2 a pi n3 a], where n3 modifies n2 instead of n1. Similarly [prep1 n1 [prep2 n2]], where the prep2 phrase modifeis n1.
No pronominal anaphora. No ni, ona, sama, jan in the sense of pronouns.
No metaphor, especially not prepositional metaphors (e.g in time, on time).
Little or no intersentence coordination, i.e. each sentence would have to make sense as if it were the only sentence.
Less polysemy- moku means cat food, not food in general, jan means people, not the indefinate pronoun, etc.
First meaning of idioms are animal world specific = jan lawa = alpha (fe)/male
Fewer words- maybe 70.
Black and white, but no color words (animal eyes are as color sensitive as ours)
And of course testing will be tricky, else we'll get clever hans effects where the animal ignores the language and infers what you want from paralinguistic evidence.
The "n. adj li n. adj" pattern alone can generate 12 million-- so the language is large enough to generate more sentences than an animal can memorize.
And we also probably will have trouble deciding what to ask the animals, since they seem to be more like specialized chess playing machines than general purpose computers. The might be able to discuss the finer points of catching mice, but if they can't reason about circles and colored blocks, they won't be able to select the red block from a selection of toys.
Similarly, for the things cats and dogs do care about (food, other animals) they have other senses at their disposal. So it would be tricky to evaluate their responses to sentences that are not commands.
sin la moku pi kiwen suno li lon poki supa.
recently, there is tinned food in the cupboard.
On my todo list is to work out what the meaning shifts of the 124 words would be for cats.
ilo monsuto = vacuum.
Instead of "jan li wile sona alasa tawa moli ala." "one must know to hunt to survive", from the cat's stand point, it should be "soweli li wile sona alasa tawa moli ala"-- while still being the sense of the indefinite pronoun.
Since I've no one to talk tp to, I've resolved to start speaking tp to my cat. From time to time I speak to my cat in English, so why not in toki pona. But first, I thought I should give him a chance at understanding me by simplifying things further.
If animals can only understand symbols (ie. can't break sentences into parts), well this is waste of time. But what if they are only missing some of the machinery of language?
S1 = [o] n [a pi n a] li [prep] n [a pi n a]
s2 = [o] n [a pi n a] li v [e n a pi n a]/[prep n a pi n a]/[nothing]
s3 = [o] n [a pi n a] la S1/S2
where [ ] means optional and / means "on or the other"
Sentences of maximal length.
Sentence diagrams of maximum depth.
No recursion. Recursion is most obvious in [n1 a pi n2 a pi n3 a], where n3 modifies n2 instead of n1. Similarly [prep1 n1 [prep2 n2]], where the prep2 phrase modifeis n1.
No pronominal anaphora. No ni, ona, sama, jan in the sense of pronouns.
No metaphor, especially not prepositional metaphors (e.g in time, on time).
Little or no intersentence coordination, i.e. each sentence would have to make sense as if it were the only sentence.
Less polysemy- moku means cat food, not food in general, jan means people, not the indefinate pronoun, etc.
First meaning of idioms are animal world specific = jan lawa = alpha (fe)/male
Fewer words- maybe 70.
Black and white, but no color words (animal eyes are as color sensitive as ours)
And of course testing will be tricky, else we'll get clever hans effects where the animal ignores the language and infers what you want from paralinguistic evidence.
The "n. adj li n. adj" pattern alone can generate 12 million-- so the language is large enough to generate more sentences than an animal can memorize.
And we also probably will have trouble deciding what to ask the animals, since they seem to be more like specialized chess playing machines than general purpose computers. The might be able to discuss the finer points of catching mice, but if they can't reason about circles and colored blocks, they won't be able to select the red block from a selection of toys.
Similarly, for the things cats and dogs do care about (food, other animals) they have other senses at their disposal. So it would be tricky to evaluate their responses to sentences that are not commands.
sin la moku pi kiwen suno li lon poki supa.
recently, there is tinned food in the cupboard.
On my todo list is to work out what the meaning shifts of the 124 words would be for cats.
ilo monsuto = vacuum.
Instead of "jan li wile sona alasa tawa moli ala." "one must know to hunt to survive", from the cat's stand point, it should be "soweli li wile sona alasa tawa moli ala"-- while still being the sense of the indefinite pronoun.