One problem with toki pona is that there is only one word order. The subject can't come after the verb nor the object before nor the prepositional phrases anywhere but at the end. tp does not even have a device for changing the predicate to allow for rearranging the order of terms, like passives (jan Sonja's reason for not having passives makes a good point but has several unintended harmful consequences). tp does have some material that would allow some movement: predicates are marked off with 'li' and objects with 'e', a prepositions are their own marks (although, of course, they can be countless other things), so only a shifted subject is a problem technically. We could, of course, invent a word for subject marking (always dropped at the beginning of sentences), but that seems contrary to the program (unless it is what 'pu' is about).
For now, it seems best to continue, noting that with a little ingenuity, many people have manged top produce sentences with satisfying word order that goes against what one would have expected. It is even possible to fake a reasonable passive using 'kama' (and 'tan'). Some folk have even swapped DOs and PPs with only mild objections, but no one has yet tried an object before the verb -- nor a PP. But, if literature is to arise, I think some of these things will have to arise and we should plan for them.
pakala wan pi toki pona li nasin wan taso
Re: pakala wan pi toki pona li nasin wan taso
I think this can cause problems for pragmatics, where you are supposed to talk about important stuff first, old info then new info (or some order like that). I'm not sure if it violates a grician maxim, but putting less important info first, seems to imply we have to violate a grician maximjanKipo wrote:One problem with toki pona is that there is only one word order.
mi moku e telo pimeja wawa.janKipo wrote:The subject can't come after the verb nor the object before nor the prepositional phrases anywhere but at the end.
? telo pimeja wawa li moku tan mi.
? kepeken ilo palisa moli la mi pilin e ni: ni li nasa. jan nasa en jan ike li moli e jan lili mute.
Unless jan Sonja has a change of heart, pu is a contraction of lipu mute.janKipo wrote: tp does not even have a device for changing the predicate to allow for rearranging the order of terms, like passives (jan Sonja's reason for not having passives makes a good point but has several unintended harmful consequences). tp does have some material that would allow some movement: predicates are marked off with 'li' and objects with 'e', a prepositions are their own marks (although, of course, they can be countless other things), so only a shifted subject is a problem technically. We could, of course, invent a word for subject marking (always dropped at the beginning of sentences), but that seems contrary to the program (unless it is what 'pu' is about).
? mi moku kepeken e ijo pi kiwen suno e pan.janKipo wrote:Some folk have even swapped DOs and PPs with only mild objections, but no one has yet tried an object before the verb -- nor a PP.
Not sure it changes the game much. (except with respect to pragmatics)
Ah, an old example, set nominal phrases that use a modifying PP
? kule lon palisa luka li pona tawa mi.
* e pan li moku mi.
There isn't a good starting separator for subjects because in the regular sentence it is a "null" marker.
Re: pakala wan pi toki pona li nasin wan taso
I'm never sure how far Gricean constraints go, but it certainly is sometimes clearer and more forceful to use odd orders to make points.janMato wrote:I think this can cause problems for pragmatics, where you are supposed to talk about important stuff first, old info then new info (or some order like that). I'm not sure if it violates a grician maxim, but putting less important info first, seems to imply we have to violate a grician maximjanKipo wrote:One problem with toki pona is that there is only one word order.
I would say either just 'moku mi' or ' moku tawa mi' , but both seem to have 'moku' meaning "food", not "eat", that is, not a present activity (of course, the original form doesn't have to be a present activity either, but that is the usual way of reading it).mi moku e telo pimeja wawa.janKipo wrote:The subject can't come after the verb nor the object before nor the prepositional phrases anywhere but at the end.
? telo pimeja wawa li moku tan mi.
? kepeken ilo palisa moli la mi pilin e ni: ni li nasa. jan nasa en jan ike li moli e jan lili mute.
Using conditions for topic and other stress needs is a real possibility and probably more forceful than just shifts in order (which are not always clear as to import).
Not sure what to make of first example here, 'kepeken' doesn't take 'e' when a preposition (as here, apparently, since eating silverware is unlikely). I suspect this comes out most naturally as 'mi moku pi kepeken ilo pi kiwen suno (or walo or mun) e pan' and that may be exactly what is needed to front PPs: make them modifiers of the verb. Similarly, we should have 'kule pi lon palisa' but that is not a sentence PP at all since the goodness is not on the fingers. The third, without the 'e' plays on the oddity of 'moku' again and makes perfectly good sense , though not obviously an activity (could be though: 'moku mi' can be "what I am eating"). The question is whether we can get plausible techniques in general, and it does seem we probably can,? mi moku kepeken e ijo pi kiwen suno e pan.janKipo wrote:Some folk have even swapped DOs and PPs with only mild objections, but no one has yet tried an object before the verb -- nor a PP.
Not sure it changes the game much. (except with respect to pragmatics)
Ah, an old example, set nominal phrases that use a modifying PP
? kule lon palisa luka li pona tawa mi.
* e pan li moku mi.
There isn't a good starting separator for subjects because in the regular sentence it is a "null" marker.
Re: pakala wan pi toki pona li nasin wan taso
oops, stray e.
? [mi] [moku] [kepeken ijo pi kiwen suno] [e pan].
Habits are hard to break, I guess my brain felt I needed an e because the verb had happened recently.
Now continuing from toki lili
But the bare subject is unmarked. But the 2nd subject is marked, e.g. soweli en kala li tawa wawa. So I can see kala is a subject. But if I move just soweli anywhere in the sentence it breaks the rules and doesn't make sense anymore. Unless en could be used as a marker for subjects.
e.g.
* li moku kepeken ijo pi kiwen suno e pan en mi.
which in a parallel universe with order insensitive TP parses as
[li moku] [kepeken ijo pi kiwen suno] [e pan] [en mi].
Verb-eat, PP-with silverware, DO-bread, Subj-I
The information is fully recoverable in the above scenario. It's kind of like the flexibility of cases-- pick any order and let word cases sort it out. Except this is more like, use always clitics and order the phrases in any order you like.
? [mi] [moku] [kepeken ijo pi kiwen suno] [e pan].
Habits are hard to break, I guess my brain felt I needed an e because the verb had happened recently.
Now continuing from toki lili
new interpretation of en? does it mean and, or does it mean, start of a subject phrase?
The thought occurred to me when reading jan Kipos post about "re-ordering" the basic sentence pattern. Marked phrases can be moved around, e.g. prop phrases can be moved to the subject slot, post verb-pre-e slot and it makes sense, but merely violates rules.Can I learn about this? idk about its new meaning (I thought it was just for 'coordinating head nouns' (jan Sonja en jan Pije li pona)
But the bare subject is unmarked. But the 2nd subject is marked, e.g. soweli en kala li tawa wawa. So I can see kala is a subject. But if I move just soweli anywhere in the sentence it breaks the rules and doesn't make sense anymore. Unless en could be used as a marker for subjects.
e.g.
* li moku kepeken ijo pi kiwen suno e pan en mi.
which in a parallel universe with order insensitive TP parses as
[li moku] [kepeken ijo pi kiwen suno] [e pan] [en mi].
Verb-eat, PP-with silverware, DO-bread, Subj-I
The information is fully recoverable in the above scenario. It's kind of like the flexibility of cases-- pick any order and let word cases sort it out. Except this is more like, use always clitics and order the phrases in any order you like.
Re: pakala wan pi toki pona li nasin wan taso
As I said, 'en' is not just for subjects but for the noun part of prepositional phrases and for the mixture sense in verbs: sike ni li laso en loje (stripes, background and figures, swirls, whatever) and elsewhere, even 'jan li tawa e poki tawa en tan tomo ona' So, 'en' is a bad choice (though the argument would work but for these ugly facts). Nothing else suggests itself much (I suppose we could use 'la', but I like that for putting the "if" last, if we're moving stuff around. (Speaking of 'la', we need to make the time signs more regular, figure out what else goes there, find what prepositions will work for what tense markers -- another piece of displacement, perhaps along with displacing PP to before 'la'.)
But, as I noted above, we do seem to have a lot of things we can use already: passives in 'kama ... tan ...'. PPs as verb modifiers, and so on. We just need to build a corpus and some further experiments.
But, as I noted above, we do seem to have a lot of things we can use already: passives in 'kama ... tan ...'. PPs as verb modifiers, and so on. We just need to build a corpus and some further experiments.