Experiencers and agreement

[Šọ̈́gala]

So, the conlang that I'm working on has polypersonal agreement on verbs – specifically, transitive verbs mark person for two participants, one of which is considered more central and the other peripheral (the central participant is placed first, closer to the verb, to the extent that they can be disentangled). For local clauses (i.e. those in which a core participant is in the 1st or 2nd person), the hierarchy for marking is Agent, Patient (essentially accusative), but for non-local clauses the hierarchy is Patient, Agent (essentially ergative).

I have been thinking about how to handle give-receive and experience type verbs. I'd like opinions on whether my current idea seems stilted, unrealistic, implausible in light of real-world typology.

Basically, my inclination is to treat give-receive verbs as equivalent to agent-patient verbs, but for the theme (i.e. the experienced phenomenon) to be relegated to a lower place on the hierarchy. So, in practice, the hierarchy for local clauses would be:

1) Agent/Sender
2) Patient/Recipient/Experiencer
3) Theme;

and for non-local clauses:

1) Patient/Recipient/Experiencer
2) Agent/Sender
3) Theme

Thus, for verbs of experience, the effective hierarchy would be the same in local and in non-local clauses (central=Experiencer and peripheral=Theme), whereas, for other verbs, the hierarchy is flipped.

But is it unrealistic and weird to treat give-receive verbs differently from experience verbs? Do languages consistently treat the theme of an experience as equivalent to a sender?

[Sakir]

Although I don't feel qualified to plant a stamp of 'Naturalism Approved' on it, I think it'd be a neat grammatical wrinkle to add to your conlang. It seems that it'd increase the texture of the language, and make it feel more 'messy'/naturalistic.

[Šọ̈́gala]

If only there were a Naturalism Approved certification I could go for for aspects of my conlang!