Note. Some of these flaws may go back to Pije, whose revisions were not quite as advertised.
Pronunciation: using 'more' is probably bad for English, where 'or' reglarly gets odd pronunciation. 'toe' is probably safwer. But the boundaries are so fluid here that it probably doesn't matter as long as /o/ doesn't slide into /a/ or /a/ into /e/.
ambiguity. Strictly speaking, tp words are not ambiguous; they are vague. The fact that they can be precided in various ways in different contexts or that they correspond to several different words in English (or whatever) doen't make them ambiguous, as long as here is a coherent (very broad, perhaps) meaning that covers all the cases. But the point here is comparative, so it is true that one tp word covers a large number of English (etc.) words and that is the point.
Verb (I won't mutter about the constant use-mention muddles throughout. I just accept that this kind of messiness is standard among tpers (and pretty much conlangers generally) and drop a reminder in from time to time. E.g.: “Because toki pona lacks 'to be', the exact meaning is lost. 'moku' in this sentence ...” ) It isn't by the way, the lack of the verb “to be” but the lack of a distinctive mark for noun use and verb use (and adjectival use, since 'mi moku' could mean “I am edible” as well). But none of that is at issue here; it is enough to warn of the variety of readings possible, which you do. (Why is the lack of tense markers ambiguity and of number vagueness?) We can argue about whether 'li' separate subject and verb or starts the predicate (but is dropped after 'mi' and 'sina' standing alone). Sorry I keep getting off on techy details, but it is better not to commit to early on these issues, as “separates” does, since it makes it hard to explain compound sentences later, where it doesn't do that.
DO. “whatever is getting action done on itself” is not at all clear, since it suggests DO's cooperation (which is rare) and covers things like the objects of prepositions, which aren't DOs.
(I admit, I couldn't figure out why 'wile' was singled out for special treatment nor why 'wile verb' was brought in before introduced formally, until I saw the bit about placing DO relative to infinitive (as it were) and remembered you were working this for German more than English.)
Compound Sentences. There are actually three ways (at least, we skip over 'anu' which isn't well worked out) to do this. You omit 'Subj en Subj Pred' (and similarly with conjoined prepositions and prepositional objects). I suppose these are dealt with separately even though they are grammatically the same (or many of them are). It's arguable whether it is easier to stick 'li' in front of each predicate except when preceded by 'mi' or 'sina' alone or to have it always there but dropped in the exceptional cases.
Adjectives. Mainly about idioms, which is a good way in. Verbs are often used as adjectives – with regular shifts in meaning (worth noting?) . Nouns are, too, although no mention is made of this before they are so used. There is no upper limit to the number of modifiers in a noun phrase, though, in practice fewer is better. And the terminal positions of 'mute', 'ni' and personal pronouns is part of a more general pattern of adjective order (and notice, these three cases can shift among themselves, with varying meanings).
Adverbs No number restriction here either, but they tend to be fewer than nouns.
'nasa' as a verb, with regular shift in meaning. Example should probably wait until explained.
Prepositions. Why introduced as verb? I see the old “do something to” line is at work again (see earlier). Going to a house affects the house at least as much as looking at something affects it. 'tawa' is just a preposition, so its object behaves differently in the sentence from that of a (transitive) verb. Probably should talk about sentential prepositional phrases before using them out of the blue. The problem with those terminal PP is quite general, not confined to 'wile' and 'lon' – as would have been brought up if the PPs had been discussed already. The expansion move is a good one to teach, as it is almost always helpful when muddles occur. 'kepeken' is just another perposition, and behave just like 'lon' and 'tawa'. The use of 'tawa' (and other prepositions less often) as transitive verbs is just part of a regular process that covers nouns and adjectives as well and could be dealt with all at once, rather than several times over. 'tawa' isn't often used a preposition, it is a preposition, but is often used as a verb and an adjective and occasionally a noun. That's just how tp works. And so we get back to the terminal PP problem with 'tawa', the most common offender. Again, one warning should server for all here. (While you're running trough cases of 'tawa', you might throw in one like 'mi tawa tomo e poki' “I moved the box to the house”.)
'kama' (Infinitive) doesn't really need its own section if modals have been discussed already (as they have not yet), since it is not special except in its wide use.
Spatial nouns. Don't confuse 'poka' preposition “with” with 'poka' noun “side” by saying 'mi moku poka jan pona mi' means “I ate beside my friend” rather than “with him”; the former is 'lon poka pi jan pona mi' (which, amazingly, you can't say yet). The confusion encourages the already notable pressure to make all the spatial words prepositions (which will happen eventually, but needs to be prepared for).
Negation. 'jan ala li toki' is unambiguous, 'jan li toki ala' is ambiguous: it also means “Someone is silent”.
Note, 'ala' alone is also a good “No” answer.
Unofficial words. Some discussion of how to tpize names is probably in order, since people do want to do it and do not (given some of the products) have a clue.
Addressing. This restriction on commas does not seem to hold. It is often omitted (or worse) with vocatives. It is regularly used, by some folks with 'la' and occasionally to set off terminal PPs and to clarify other messy constructions. The worse mentioned above is the collapse of vocative + imperative into a single 'o', which is cute but doesn't correspond to speech (while the comma does). It also gets confused with the optative (hortative), if that is done simply by replacing 'li' with 'o' (as opposed to the old form that just put 'o' in front of a regular sentence). This whole area, which was in fairly decent shape a while back, is now a mess due to unthought-through changes arbitrarily introduced (apparently).
'sin' has regularly come to mean “new” in a different sense than was originally intended, apparently. I don't think that can be undone (nor should it be, since we need it). It is never used in the sense of “more”.
Compound nouns. This description of 'pi' makes virtually no sense. Most importantly, there is no requirement that the second chunk be a noun+ adjective; it can equally well be an adjective +adverb or a PP or any other sort of complex modifier. See your own examples with 'ala'.
As a result, some of your common mistakes are not mistakes at all. 'jan pi wawa pi pona mute li kama' “A person of very good strength came” is quite correct, and very different from either 'jan pi wawa pona mute li kama' “A person of many good strengths came” or “Many people of good strength came” or 'jan wawa pona mute li kama' “Many many goodly strong people came”. (It occurs to me that you have never really laid out the way that modification works in tp).
Using 'pi' or simple modification for “about” went out long ago, though seems to be coming back (maybe because of this book and Pije's) I suppose what you want to prevent is 'toki pi jan' although that is not clear.
'anu' can be used to make assertions as well as ask questions. Apparently the only difference is in the punctuation. It can also be used between NPs in subjects and objects, both direct and prepositional, and, apparently, between sentences (though just how to punctuate that is not clear – it mainly seems to be like 'taso') and presumably between predicates and verbs and modifiers, though these are trickier.
'en' can't be used between sentences, true, but it can be used between prepositional (but not direct) objects as well as subjects and between modifiers (with required 'pi', of course). All of this is about transformational 'en', the result of collapsing two sentences into one (as are also the earlier compound sentences). None of this applies to mixture “and” which is generated lexically: “That ball is red and blue” is 'sike ni li loje en laso', 'li loje li laso' would simply be false. And so on for other cases.
'taso' is also a limited scope preposition: “except” which is pretty much restricted, so far as I can see, to modifying group nouns: 'ali (pi) taso jan San' “Everybody except John” and the like. I don't quite see the distinction between the adjective and adverb uses here – 'taso' is like 'kin' and can be stuck on anywhere to make a logically complex claim (“this and nothing else”). There is another adjectival use, to mean “alone, solitary” which is more remote from the central notion here.
'kin' goes after the stressed word, to mark and addition or correction. So, “I went too” is, in context. 'mi kin li tawa' (English – and German? – are sloppy with “too” and “only” and the like) 'mi tawa kin' says either that I did something different or more in going, neither of which is correct.
I skip my rant about 'pilin pona', which ought to be 'pona pilin'. Just note that, except for 'pona' and 'ike', other feelings – like cold, say, are adj+'pilin'.
I see that both your description of 'en' and of 'pi' go away in the practical situation of a red and blue shirt. Maybe you shouldn't have said the restrictive things you did originally.
'sitelen' 'sitelen tawa … li kin pona' makes no sense, what would stressing 'li' do? I assume the 'kin' goes just before the 'li', at the end of the NP.
'ko' isn't used much but one should't dismiss it so casually, since it includes gelatin, flour, glue, and countless other things no one has yet had occasion to talk about in tp. And it's a great verb for “to squash”.
'uta e' is used for “kiss”
or 'mi wile pilin e selo mi” “I need to/want to scratch.”
Numbers. 'mute' and 'ali' are a terrible idea for specific values. Either give up big numbers or do the job right. These don't help at all.
Well, “only pronouns after” if 'ni' is counted as a pronoun – and 'ali' and 'ala', of course.
'jan mute kin' also 'jan pi mute mute' (and, in the works 'jan mute mute mute' but let that slide.)
'weka' is an adjective meaning “away, far” and so on. The other meanings follow automatically. This presentation is basically backwards.
'la'. If you think 'ken la' is hard (it does mean “possible” after all), try 'kin la' “moreover, however”.
You don't do superlatives as promised (Yes, it is obvious, but...) note also the occasional use of say 'pona ali' for “best”.
I didn't look at the texts to closely, but just in skimming I saw most of the things I notice earlier in the text and accidentally stopped on 'waso li ken ala lukin la ona li kin ken jo e kili', presumably 'ona kin li'.
Ah, just found a more complete guide to tpization.
The confusion among place names, people names and language names persists. It would be good to either deal with each separately or agree on where to get the common word for all.
The English to tp section is as useless as usual – the next project, perhaps.
Still, this is in many ways better than pu and occasionally even improves on Pije. Thanks.