"li"-marker has two proto-semantic atoms: "to be" ("A := B", nomination) and "to do" ("A -> B", action). They are almost identical, because, if you have a name "king", you behave yourself as a king. If you have a name "father", you behave yourself as a father. If you have a name "rocket", you behave yourself as a rocket...etc. "pu" mixes these two meanings. And it gives a permanent internal semantic conflict (nomination vs. action, re-naming vs. re-acting). I guess, it will be good for tp to use a fit marker for each case:
"is" -- if you want to nominate something ("man is good", "man is driver", "man is married")
"li" -- if you want to activate something ("man eats", "man feeds", "man thanks")
Evaluate an exactness of expression, if to use both markers (both proto-semantic atoms):
1) "ona is olin mi" -- "she is my love -- (nomination: is)
2) "ona li olin e mi" -- "she loves me" -- (action: li)
3) "olin ona li ali lon mi" -- "her love lives into me" -- (action: li)
4) "olin ona is pona" -- "her love is good" -- (nomination: is)
5) "ona is musi ala" -- "it is not funny" -- (nomination: is)
6) "ona li musi" -- "it makes fun" -- (action: li)
If you want to express the nomination, use the marker "is" ("jan is pona" -- a good boy). And, when you wish to underline the action, insert the marker "li" ("jan li pona -- the boy makes a good). Such a separation will help tp to become clear semantically. I believe that tp can become a mediator, like Esperanto, in poly-linguistic communities:
nomination: "la donna è bella" -> "jan meli ni is lukin pona" -> "moteris yra graži"
action: "die Frau singt" -> "jan meli ni li toki kalama" -> "la femme chante"
By will of tp-community, in the visible future tp will be able to translate more complex meanings between speakers from any country. First steps to this goal you see above: italian -> tp -> lithuanian, germ -> tp -> fr.
so I am wondering how you have collapsed the DO /.../(who is being fed) /.../
mUU-format contains: m-marker (li: verb), 1st-U is a lead action, 2nd-U is its shade.
li "pana moku" -- "food-giving act" -- feeding (here: "pana" is "action", "moku" is "shade")
li "pana mani" -- "money-giving act" -- investing
li "pana sona" -- "knowledge-giving act" -- teaching
Now about difference between "mU" and "mUU". If a verb slot has only word, it means an act without aim. You may use the thing for any purpose you want (store, catapult, trade, art, ritual...) But, if a verb slot has two words like "pana moku", it means an act with aim. So, what is a mission for food? To be eaten. What is a mission for money? To be spent. What is a mission for knowledge? To be known. The 2nd-U in a verb slot defines a mission. Here is still one way to analyse the sentence "jan li pana moku e ona", to understand: who is being fed? Let us devide this sentence on a left part and a right part. The left part "jan li pana moku" means "man is a food-giver". So, from the left side we have the food-giver, the man who makes a food-giving act. It is his mission. Now we shift our focus on an object: on the right part of the sentence "moku e ona". What is a mission of the food-eater? "to eat a food". It is just the right part ("moku e ona") of our sentence (jan li pana moku e ona). The word "moku" is related as with a "left" provider as with a "right" receiver. In whole, the sentence "jan li pana moku e ona" describes a trip of food (from a hand to a stomach).
Finally, compare three examples to see semantic flexibility, which tp can express (mU vs. mUU vs. mUUU):
: "jan li pana
e moku tawa ona" (here, feeding is one of variants) may be fed: 0-100%
: "jan li pana moku
e ona" (here, feeding is embedded in a verb) must be fed: 100%
: "jan li pana moku kili
e ona" -- this is about 100%-vegetarian feeding
'en pini' is perfect passive, apparently
I would like to change "en"-marker on a new trial variant "jen", because, you are right, "en" is too close with pu. So, jen-marker from now. "jen pini" means that something is finished (maximized on to 100%). And:
"jan lili jen pini moku
" -- "kid is foodily maximized on to 100%
... stomach is filled", i.e. "kid is fed".
In this set, 'moku' is always assumed to mean "food", without a marker ('en ijo'?). And in the next sections it is similarly assumed to mean "eating" and "edibility". How do you know which is which?
Almost all languages are conventional, including tp. I guess, by default "moku" means "food". If we want to point some feature, we add an affix: "jen ijo", "jen awen", "jen ken" or any other like affixes of Esperanto.