I came across a post by Radius Solis on the ZBB which touched a topic I've generally not cared to think about, but being in need of some procrastination, I decided to think about it.

Here is the post:
"Hmm... I wonder. Various languages categorize nouns in various ways - by shape, by gender or animacy, by their utility to humans, by whether they are considered to be inherently possessed, and all sorts of other odds and ends. What languages, if any, categorize their verbs in a similar manner? And on what semantic basis? And where is the categorization marked?"

I would consider the Slavic motion verbs vs. other verbs such a distinction; these use the various prefixes (generally derived from prepositions) in a much more consistent way than do other verbs, and they have an added dimension in their aspectual system (altho' many grammars will state that this added distinction is not an aspect, I think it is close enough for this post to say that they have two axes of aspect whereas other verbs just have one).

I wouldn't be surprised, however, if for nouns, the categorization generally is fixed, and for verbs it's more generally derivative? So that, say, the closest analogy to 'by utility for humans', which I'd for no particular reason say is 'beneficialness for humans', the classification gets more tricky for verbs. Right, knives per se are very utilizeable by humans, but cutting ... can go both ways? Of course, a knife can be a bad thing as well in the wrong hand, ... do languages that classify nouns along utility have these classes as statical classes or do they permit changing stuff around?

As for classification along possession, shape, etc. The Slavic verbs of movement sort of go into the 'shape' slot, on some semantic grounds; verbs of movement have a specific sort of shape that other verbs might lack (altho' some verbs that indicate movement, iirc, don't qualify in some Slavlangs, so it's a rather exclusive club of verbs). Possession would make sense for verbs that mark actions that are culturally limited to various classes of people; maybe a verb that can only be performed by one gender would completely lack gender marking? Or have a different morpheme instead of the usual gender marking? Or whatever, but there's possibilities.
(C.f. the Russian verbs for marrying, where the verb for females actually is a prepositional phrase, za muzh / za muzhem, literally 'to behind the man' / 'behind the man', and for males, zhenitsya, 'to wife oneself [someone]' or whatever we'll literally translate it as - that is, to take oneself someone for wife' or such. That's an interesting classificational detail, altho' again, it's not really an entire classificational system, but nevertheless...)

This ties neatly in with one design idea I have for dairwueh, viz. the idea that the case of the object (accusative or some oblique case or partitive, not sure yet) marks whether the action is considered beneficial for the speaker (and listener, depending on in what standing they are) or not. Of course, this is neither derivational, nor a classification. It is rather just extra information added on.

Not a very coherent post, just some stuff that doesn't really add up to anything yet.