This weirdness about objects with {-moH} has always bothered me. I mean, if Klingon is simply loose with what objects are, then why to we bother talking about the prefix trick when we say something like {chab HInob}? If objects are so loose, why even call it a trick? It’s just yet another loose object. It doesn’t deserve having a specific term to describe it or rules about it’s use. The only special thing about it is that it was revealed to us before all the looseness with {-moH}.
It's called the "prefix trick" because I called it that. I coined the phrase. I did so scornfully, because it seemed to me that the rule was created as a lame excuse for errors by Okrand in which he just translated English without thinking, and didn't consider the very straightforward rule that the verb prefix must agree with the object.
In fact, I still think Okrand was covering for sloppy, Englishy translations, but I also think this new information was something more: a sign that Okrand was thinking more carefully about the grammar of Klingon, and deepening it in response, deepening it beyond the simple "an object is an object is an object" structure it had at first. Klingon clearly had need of more complicated predicates, as was being demonstrated to him by KLI members who kept questioning his sentences and expressing frustration with the limitations of Klingon. I think he was also being asked to translate more and harder pieces, and he discovered that he needed to make Klingon more flexible to do this.
The "looseness" we're talking about is not a property of -moH. It's a property of Klingon objects. Consider the old sentence, qaja'pu' I told you. What is the object of the sentence? Apparently, it's an elided SoH, because the prefix qa- tells us so. So everyone said "The object of ja' is the person spoken to." But (a) that doesn't jibe with the gloss given in TKD, where the person being spoken to would be the indirect object, not the direct object, and (b) we have since seen ja' take other objects, like lut. The solution is simple: the "object" of a verb might be a direct object (acted upon) or an indirect object (receives the result of the action), depending on how the verb is used.
It just so happens that verbs with -moH encounter object "looseness" more often because verbs with -moH add a new entity into the semantics of the sentence: the causer. The old entities might still be present, so you have to think carefully about what is being acted upon and what is receiving the result of the action. But this "looseness" isn't a property of -moH itself; it's a general property of Klingon sentences that might change things up depending on what verb you're using and how you're using it.
I still call it the prefix trick partly to remind myself
of my foolish days when I thought Klingon sentences were rigid,
their syntax and semantics invariable, and partly because that's
what everybody calls it now, and we don't have another phrase for
it. But it's not a trick; it's just a fact of verb prefixes, that
while they must agree with an object, they don't
necessarily have to agree with a direct object.
-- SuStel http://trimboli.name