There are very few predicates with types "xN (property of xM)" with M > N:
| makfa |
x1 (property of x3) is magic or supernatural to x2 performed by x3
|
That's the only one where the property skips a place. I don't think x1 needs to be a property of x3 though, couldn't we say:
lo nu lo ractu cu canci cu makfa lo verba la djan
I think makfa1 could be something that x3 makes happen, not necessarily something in which x3 takes a part.
| vamji | x1 (property of x2) is what x2 is worth of to x3 for use x4 (property of x2) |
This one I think is wrong. vamji1 is not a property of x2, it is something that x3 is willing to give up or endure in exchange for obtaining or keeping x2.
| krali | x1 (property of x2) is a legal/moral entitlement of individual x2 in legal system x3 (located) |
| marde | x1 (property of x2) are the ethices of performer x2 about situation x3 (located) |
| cfila | x1 (property of x2) is a flaw or defect in x2 causing x3 (located) |
| jinzi | x1 (property of x2) is an innate or natural property of x2 |
| kampu | x1 (property of x2) is common among x2 |
| steci | x1 (property of x3) is specific or particular to member x2 among x3 |
| tengu | x1 (property of x2) is a texture of x2 |
| jibri | x1 (property of x2) is a job of person x2 |
| frili | x1 (property of x2) is easy for x2 |
steci3 should be "property of x2". These seem fine, although they could have been defined in reverse order for the most part. Here we seem to be missing some though: nandu/tcaci/snuti/cimde/vrusi?/tarmi?
tseli2 says (property of x2) instead of (property of x1).
| sisku |
x1 searches for x2 (property of x3) among x3
|
I don't agree with that one, I think sisku2 should be an object not a property.
| mupli |
x1 (any type) is an example of x2 (property of x3) among x3 |
| fadni |
x1 (any type) is ordinary in x2 (property of x3) among x3 |
| traji |
x1 (any type) is superlative in x2 (property of x3) among x3 |
| cnano |
x1 (any type) is an average in x2 (property of x3) among x3 |
| rirci |
x1 (any type) is rare or unusual in x2 (property of x3) among x3 |
These should all have x2 (property of x1)
| cpedu |
x1 requests or asks for x2 (property of x3) from x3 in manner x4 (property of x1) |
| stidi |
x1 suggests x2 (property of x3) to x3 |
| javni |
x1 (located) is a rule mandating x2 (property of x3) in x3 |
Shouldn't flalu follow the pattern of javni?
I think those are all. There are no properties in x3 or x4 that send to a later place.
mu'o mi'e xorxes