DO incorporation

Tinkerers Anonymous: Some people can't help making changes to "fix" Toki Pona. This is a playground for their ideas.
Tokiponidistoj: Iuj homoj nepre volas fari ŝanĝojn por "ripari" Tokiponon. Jen ludejo por iliaj ideoj.
janMato
Posts: 1545
Joined: Wed Dec 02, 2009 12:21 pm
Location: Takoma Park, MD
Contact:

DO incorporation

Post by janMato »

I think somewhere I got the wrong impression was that "e" phrases and prep phrases had to be moved to the end of the sentence, including in li chain sentences. Anyhow, it turns out that no one else is following that rule, so I was trying to deal with the consequences for naught. So as it turns out, there's doesn't seem to be a good reason to detransitivize verbs by replacing "e" with "pi", unless maybe you're birdwatching and like calques.

I count ~350 usages of S li V e DO li V e DO ... etc. in the corpus
I was using this regex, incase anyone cares: \bli\b [A-Za-z][A-Za-z0-9 ]* e [A-Za-z][A-Za-z0-9 ]* \bli\b

My conclusion, Wikipedia's phrase grammar is confusing.

e.g.
sona sin en ante li sin e lawa li pana e sona. jan Sonja ***
jan li moli e sijelo li ken moli ala e kon sewi la o pilin ike ala. jan Ke

Prep Phrases in the middle of li chains.
I'm still kind of sure you can't put a prep phrase right after the subject and before the li. I'm still not sure if prep phrases can be put in the middle of a sentence, e.g.

soweli li moku e moku lon kasi suli kepeken kiwen li lape.
The animal ate food in the tree using a rock and then slept.

*** Speaking of "sona sin en ante", here is a canonical example of two modifiers joined by a conjunction modifying a head noun. This pattern also shows up in the more common word for zebra-- soweli (pi kule) walo en pimeja. Previously I would have said the wikipedia phrase grammar doesn't allow for this. Now, humbler, I'm just not sure what the phrase grammar is trying to say.
User avatar
jan Ote
Posts: 424
Joined: Thu Oct 08, 2009 1:15 am
Location: ma Posuka
Contact:

Re: DO incorporation

Post by jan Ote »

janMato wrote:*** Speaking of "sona sin en ante", here is a canonical example of two modifiers joined by a conjunction modifying a head noun. This pattern also shows up in the more common word for zebra-- soweli (pi kule) walo en pimeja. Previously I would have said the wikipedia phrase grammar doesn't allow for this. Now, humbler, I'm just not sure what the phrase grammar is trying to say.
jan Pije. Lesson 13. wrote:Image
The shirt is obviously made up of loje and laso. However, you can't call it len loje laso, because that means "purple shirt" [...]. So, we have to use en to separate the two colors, and then we have to use pi to show that even though there are two different colors, they both modify the word "shirt":
  len pi loje en laso -- shirt of red and blue
Get it? Now let's look at what would've happened if you had not used pi here:
  len loje en laso li pona. -- A red shirt and blue are good.
Without pi, laso is just left sitting there, and it doesn't have anything to modify. Therefore, that sentence is incorrect. You have to use pi to show that both loje and laso modify len.
I would say that it's a good rule.

Let's take two sentences and two rules, "zebra rule" (jan Sonja?) and "shirt rule" (jan Pije):
Zebra rule:
  jan lili ona en sina li pona. = (jan lili (ona en sina)) li pona.
  Child of hers and yours is good.
Shirt rule:
  jan lili ona en sina li pona. = ((jan lili ona) en sina) li pona.
  Her child and you are good.
Zebra rule and shirt rule:
  jan lili pi sina en mi li pona. = (jan lili pi (ona en sina)) li pona.
  Child of hers and yours is good.
According to "zebra rule" the first sentence is at least ambiguos while the second sentence covers the same meaning.
According to "shirt rule" both sentences are unambigous and have different meanings.
janKipo
Posts: 3064
Joined: Fri Oct 09, 2009 2:20 pm

Re: DO incorporation

Post by janKipo »

Thanks for reminding me of that lesson in Pije's set. The argument for the shirt rule is totally decisive as far as I can see (I remember a while back trying to sort out the problems with 'pi' and 'en' and deciding that, if pushed far enough -- which wasn't terribly far -- no system would work perfectly, but the one with 'pi' setting off an modifier en modifier structure went further (beyond the first step in many cases, but notice 'jan lili pi meli sama en mije ona li pona.') In general across languages, if you don't have RHE markers, you have RHE ambiguity (and if you do have the markers you have long boring strings of them at the ends of sentences and various other places).
janMato
Posts: 1545
Joined: Wed Dec 02, 2009 12:21 pm
Location: Takoma Park, MD
Contact:

Re: DO incorporation

Post by janMato »

So (stylistically)/(grammatically) speaking, which is better:

sona sin en ante li sin e lawa li pana e sona.
or
sona pi sin en ante li sin e lawa li pana e sona.

and I suppose the question could have been sidestepped with

sona sin en sona ante li sin e lawa li pana e sona.
janKipo
Posts: 3064
Joined: Fri Oct 09, 2009 2:20 pm

Re: DO incorporation

Post by janKipo »

The 'pi' one both stylistically and grammatically, assuming you mean to talk about a new and different idea, not a new idea and something else. The two-noun solution is also good ideologically, but then, a two-sentence solution would be better ideologically. And that is going too far. Notice the ambiguity potentially here: while in this case, the two modifiers act independently of one another, in the red and blue shirt case, they do not. That is, you can't separate the "the red and blue shirt" into "the red shirt and the blue shirt" tp (like most of the languages I know, except the LoCCan) does not distinguish the "logical" "and" from "mingling" "and." Context seems to have worked for this across human history, so I guess it will do no harm in tp.
User avatar
jan Josan
Posts: 326
Joined: Sun Oct 18, 2009 12:41 pm
Location: ma tomo Nujoka
Contact:

Re: DO incorporation

Post by jan Josan »

janMato wrote:I think somewhere I got the wrong impression was that "e" phrases and prep phrases had to be moved to the end of the sentence, including in li chain sentences.
As I understand it you can have "S li V e DO li V e DO" structures. The only problem is when you want "S (li V li V) e DO" because each DO seems to refer back to the last verb only.
jan li moli e sijelo li ken moli ala e kon sewi la o pilin ike ala. jan Ke
This first 'e' work IF we are allowed to use two sentences with two subjects split with a 'la'. I think the jury is still out on this...
Prep Phrases in the middle of li chains.
I'm still kind of sure you can't put a prep phrase right after the subject and before the li. I'm still not sure if prep phrases can be put in the middle of a sentence, e.g.
soweli li moku e moku lon kasi suli kepeken kiwen li lape.
The animal ate food in the tree using a rock and then slept.
I'm not sure about this either, I think I have been challenged when attempting this in the past...
*** Speaking of "sona sin en ante", here is a canonical example of two modifiers joined by a conjunction modifying a head noun. This pattern also shows up in the more common word for zebra-- soweli (pi kule) walo en pimeja. Previously I would have said the wikipedia phrase grammar doesn't allow for this. Now, humbler, I'm just not sure what the phrase grammar is trying to say.
I also wonder what this says about conjuctions in the DO. It doesn't make sense to me that we can say:
soweli pi (kule) walo en pimeja li lon.
but not:
mi lukin e soweli pi (kule) walo en pimeja.
janMato
Posts: 1545
Joined: Wed Dec 02, 2009 12:21 pm
Location: Takoma Park, MD
Contact:

Re: DO incorporation

Post by janMato »

jan Josan wrote:As I understand it you can have "S li V e DO li V e DO" structures. The only problem is when you want "S (li V li V) e DO" because each DO seems to refer back to the last verb only.
I don't mind that. This wouldn't be the first time TP had syntactic ambiguity. It could be an intransitive followed by a transitive, or both action happen to the DO, or only the 2nd action happens to the DO and possibly more things I haven't thought of. Simplicity in production comes at the cost of complexity in parsing and understanding.

I do see here that it would be handy to have something like an "en" or "anu" that specifies the nature of the join between the verbs, or the "e" phrases when there is more than one.
jan Josan wrote:
jan li moli e sijelo li ken moli ala e kon sewi la o pilin ike ala. jan Ke
This first 'e' work IF we are allowed to use two sentences with two subjects split with a 'la'. I think the jury is still out on this...
But from jan to sewi, it is only one sentence. One sentence can have multiple verbs.
I also wonder what this says about conjuctions in the DO. It doesn't make sense to me that we can say:
soweli pi (kule) walo en pimeja li lon.
but not:
mi lukin e soweli pi (kule) walo en pimeja.
Both look okay to me because the "en" is after a "pi"

This on the other hand is wrong:
* mi lukin e soweli en akesi
* mi lukin e soweli anu akesi (and I wouldn't know how to fix it !)
? mi lukin e soweli anu e akesi.
mi lukin e soweli anu mi lukin e akesi. (conservative and possibly correct, but seems un-necessarily verbose)
mi lukin e ni. ni li akesi anu soweli. (Again, conservative and possibly correct, but what a circumlocution.)
janKipo
Posts: 3064
Joined: Fri Oct 09, 2009 2:20 pm

Re: DO incorporation

Post by janKipo »

The rejection of 'en' in the DO in the lessons seems to apply explicitly only to joining two NPs as DO -- you should use 'e'. The question of joining two modifiers (including two NP modifiers) is not mentioned explicitly, iirc. So is tp an ant language or a grasshopper one, is everything not expressly allowed forbidden, or is everything not expressly forbidden allowed? I think we are nearer the latter, so use 'en' for modifiers in DO and for anything in Subj and PP (well, maybe not for the prepositions themselves). This leaves the question whether it can be used in VP for compounded verbs (no English example of which comes to mind, so this may be a pointless question, though tp might find a use for such) or for compounded modifiers ("He ran back and forth"). Again, I'd say so, though with a lot less confidence.
Again, note the difference between the logic "and" and the compounding one. I think the compounding one should be allowed even in DOs between head nouns, if that is appropriate. And similarly, a compounding "and' could go anywhere (legal) in the 'la' clause, while the logical one might be suspect or even wrong (though we do need a way of extending both what goes before a 'la' and what comes after-- even beyond the ambiguous double 'la's).
It seems to me that, once we allow DOs in iterated 'li's we are saying that what goes there is a predicate and thus PPs ought to be allowed as well. But I can imagine objections. Of course all of this arises because we try to say too much in one breath rather than breaking it down into its parts, which is the tp way.
Last edited by janKipo on Wed Mar 03, 2010 5:08 pm, edited 1 time in total.
janKipo
Posts: 3064
Joined: Fri Oct 09, 2009 2:20 pm

Re: DO incorporation

Post by janKipo »

Josan's point about the 'la' phrase is that, although it is one sentence structurally, it is probably two logically (in the particular case of the parents of a child and the like we could argue to the opposite effect). Of course we can get two sentence conjoined (in effect) in the 'la' phrases by adding on another 'la' phrase:
p> (q > r) = (p & q) > r (of course, (p > q) . r is something very else but looks the same in tp).
'anu' seems to work with different rules, since it can be placed (I think) even between sentences and certainly anywhere in a sentence. So you can't go from what works with it to 'en' -- although the reverse, using 'pi' for disjoined modifiers, looks necessary. I think 'mi lukin e soweli anu akesi' is OK and 'e soweli anu e akesi' is definitely not.
janMato
Posts: 1545
Joined: Wed Dec 02, 2009 12:21 pm
Location: Takoma Park, MD
Contact:

Re: DO incorporation

Post by janMato »

janKipo wrote:So is tp an ant language or a grasshopper one, is everything not expressly allowed forbidden, or is everything not expressly forbidden [is] allowed?
Forgot a not. It is interesting how one can understand by context despite ambiguity introduced by a speech error, even when the ambiguity is between polar opposites. It is another example of how ambiguity on it's own isn't necessary a problem in a language's design.

The original writers of the phrase grammer (morpheme addict/and/or la pingvino) aren't updating it fast enough to fill in the problems that are being found in it. And it is on my wish list to figure out how to turn this into a BNF notation and find a utility to validate sentences the same way a computer program compiler validates C# or Pascal or what have you.
Of course all of this arises because we try to say too much in one breath rather than breaking it down into its parts, which is the tp way.
But that is style. Hemingway wrote with short(er) sentences, but it was hardly a matter of grammar-- an important distinction, imho.
Post Reply