Three-way parts of speech (again)
Posted: Sun Aug 28, 2016 5:24 pm
I don't really want to change tp, but the perennial problem of interpretation has arisen again in a number of cases, so I'm going to review (and organize somewhat) the sorts of solution that have come along -- if we don't mind making significant changes in tp . I've picked a simplified version of each type to show the general idea. The basic problem is that, even though we know the home Part of Speech of each word and know the role(is) it plays in a given sentence (its functional POS), we still don't know what it is meant to mean (semantic POS). For the most part, the choices at the end come down to nouns, verbs, and adjectives (a more thorough approach would have to deal with the lesser parts, of course)
Move one: save the syntax and change the morphology. (Inspired by Giannhs Kenanidhs' Sostematiko, though different in appearanceand realization).
The simplest move here is to mark each word overtly to show which of the semantic roles it is playing in a given location. For this sample I use the system of adding a stop at the end: /p/ for nouns, /t/ for verbs and /k/ for modifiers. Thus the notorious 'jan li moku' is variously 'janp li mokup' "Man is food", 'janp li mokut' "Man eats" and 'jan li mokuk' "Man is edible". The even more boggling 'moku li pona' runs through all the possibilities
mokup li ponap "Food is a good thing/virtue"
mokup li ponat "Food cures/ improves"
mokup li ponak "Food is good"
mokut li ponap "Eating is a virtue"
mokut li ponat "Eating is a way to health"
mokut li ponak "Eating is good"
mokuk li ponap "Edibility is a desirable trait"
mokuk li ponat "Edibility leads to good things"
mokuk li ponak "Edibility is good"
The ever annoying 'ona li pana e tono tawa mi; becomes either 'onap li panat e tomop tawat mik' or 'onap li panat e tomop tawa mip' (prepositions as prepositions go unmarked).
Obviously some simplifications are possible. In particular, if a word is just doing its job in a position it doesn't need marking. But "doing it job" can mean doing what is typical of that position (noun a head of NP, verb as head of predicate, modifier as modifier) or what its home POS is. These can have very different result: on the first, 'moku li pona' would be "Food is improving, 'mokup li ponat'; under the second "Eating is good" 'mokut li ponak'. Which one (if either) to pick is a later question.
I hope you get the idea, though individual cases may take some thinking through to hit on just the right combination of forms for the meaning intended.
Move two: save the morphology but change the syntax (based on what I think I understand of janpona120's discussion at viewtopic.php?f=29&p=14972#p14972)[It turns out that I didn't understand it at all and that his proposal is something quite different -- I don't know what.]
There are obviously a lot more variants here than in move one. I pick what seems to me to be the simplest and closest to traditional tp. The idea here is that every sentence is built up of pairs of components hierarchically arranged, each of two meaningful parts and a central connector. The main pairs are
head pi modifier
subject li predicate
verb e direct object
predicate {prep} prepositional object ({prep} being a list closed at a given time: lon, tawa, tan, kepeken, poka, and maybe a few more as of now)
sentence {conjunction} sentence ({conjunction} similarly a closed list, now just en, anu and la.)
Any of these pairs may appear in any position in any other pair and one or the other member of a pair may be missing in a given occurrence. There are probably a number of rules of elision (the mi/sina subject is not one of them, apparently) but we will skip those for display purposes. Again, it is probably possible to drop markers for naturally behaving cases, with the previous caveat about which "natural" is meant. For now, no markers are dropped. Notice that, in simple cases at least, a following 'li' or a preceding 'e' marks something as nominal (and maybe as either agentive or patientive [yuck! ptui!]). Similarly, a preceding 'li' or a following 'e' marks a word as a verb. A preceding 'pi' marks a word as a modifier and, clearly, there will need to be some elision in strings of modifiers or we will be 'pi'ing all over ourselves.
The standard test cases are
jan li moku "Man eats" (a standard source might be 'jan li moku e ijo')
jan li e moku "Man is food" (related to standard 'ijo li moku e jan', but Gawd knows how).
jan li pi moku "Man is edible" ('jan li ijo{moku}' with a different 'pi' rule)
moku li pona "Food cures"
moku li pi pona "Food is good"
moku li e pona "Food is a good thing"
moku e li pona "Eating cures"
moku e li pi pona "Eating is good"
moku e li e pona "Eating is a good thing"
pi moku li pona "Edibility cures"
pi moku li pi pona "Edibility is good"
pi moku li e pona "Edibility is a good thing"
ona li pana e tomo pi li tawa pi mi vs ona li pana e tomo tawa mi (assuming that prepositions, like 'e', compel what follows immediately to be a noun. Note, the two 'li's are not a problem since matching predicates here are joined by explicit 'en', one of the basic pairs).
Again, these are just a small and elementary sample. For direct comparison, I leave you "The man who came to dinner stayed a month" in both versions
jan pi kamat pi tawa mokut (or 'mokup', it's not clear) li awent lon tenpop munk.
jan pi li kama tawa li moku li awen lon tenpo pi mun.
Move one: save the syntax and change the morphology. (Inspired by Giannhs Kenanidhs' Sostematiko, though different in appearanceand realization).
The simplest move here is to mark each word overtly to show which of the semantic roles it is playing in a given location. For this sample I use the system of adding a stop at the end: /p/ for nouns, /t/ for verbs and /k/ for modifiers. Thus the notorious 'jan li moku' is variously 'janp li mokup' "Man is food", 'janp li mokut' "Man eats" and 'jan li mokuk' "Man is edible". The even more boggling 'moku li pona' runs through all the possibilities
mokup li ponap "Food is a good thing/virtue"
mokup li ponat "Food cures/ improves"
mokup li ponak "Food is good"
mokut li ponap "Eating is a virtue"
mokut li ponat "Eating is a way to health"
mokut li ponak "Eating is good"
mokuk li ponap "Edibility is a desirable trait"
mokuk li ponat "Edibility leads to good things"
mokuk li ponak "Edibility is good"
The ever annoying 'ona li pana e tono tawa mi; becomes either 'onap li panat e tomop tawat mik' or 'onap li panat e tomop tawa mip' (prepositions as prepositions go unmarked).
Obviously some simplifications are possible. In particular, if a word is just doing its job in a position it doesn't need marking. But "doing it job" can mean doing what is typical of that position (noun a head of NP, verb as head of predicate, modifier as modifier) or what its home POS is. These can have very different result: on the first, 'moku li pona' would be "Food is improving, 'mokup li ponat'; under the second "Eating is good" 'mokut li ponak'. Which one (if either) to pick is a later question.
I hope you get the idea, though individual cases may take some thinking through to hit on just the right combination of forms for the meaning intended.
Move two: save the morphology but change the syntax (based on what I think I understand of janpona120's discussion at viewtopic.php?f=29&p=14972#p14972)[It turns out that I didn't understand it at all and that his proposal is something quite different -- I don't know what.]
There are obviously a lot more variants here than in move one. I pick what seems to me to be the simplest and closest to traditional tp. The idea here is that every sentence is built up of pairs of components hierarchically arranged, each of two meaningful parts and a central connector. The main pairs are
head pi modifier
subject li predicate
verb e direct object
predicate {prep} prepositional object ({prep} being a list closed at a given time: lon, tawa, tan, kepeken, poka, and maybe a few more as of now)
sentence {conjunction} sentence ({conjunction} similarly a closed list, now just en, anu and la.)
Any of these pairs may appear in any position in any other pair and one or the other member of a pair may be missing in a given occurrence. There are probably a number of rules of elision (the mi/sina subject is not one of them, apparently) but we will skip those for display purposes. Again, it is probably possible to drop markers for naturally behaving cases, with the previous caveat about which "natural" is meant. For now, no markers are dropped. Notice that, in simple cases at least, a following 'li' or a preceding 'e' marks something as nominal (and maybe as either agentive or patientive [yuck! ptui!]). Similarly, a preceding 'li' or a following 'e' marks a word as a verb. A preceding 'pi' marks a word as a modifier and, clearly, there will need to be some elision in strings of modifiers or we will be 'pi'ing all over ourselves.
The standard test cases are
jan li moku "Man eats" (a standard source might be 'jan li moku e ijo')
jan li e moku "Man is food" (related to standard 'ijo li moku e jan', but Gawd knows how).
jan li pi moku "Man is edible" ('jan li ijo{moku}' with a different 'pi' rule)
moku li pona "Food cures"
moku li pi pona "Food is good"
moku li e pona "Food is a good thing"
moku e li pona "Eating cures"
moku e li pi pona "Eating is good"
moku e li e pona "Eating is a good thing"
pi moku li pona "Edibility cures"
pi moku li pi pona "Edibility is good"
pi moku li e pona "Edibility is a good thing"
ona li pana e tomo pi li tawa pi mi vs ona li pana e tomo tawa mi (assuming that prepositions, like 'e', compel what follows immediately to be a noun. Note, the two 'li's are not a problem since matching predicates here are joined by explicit 'en', one of the basic pairs).
Again, these are just a small and elementary sample. For direct comparison, I leave you "The man who came to dinner stayed a month" in both versions
jan pi kamat pi tawa mokut (or 'mokup', it's not clear) li awent lon tenpop munk.
jan pi li kama tawa li moku li awen lon tenpo pi mun.