I'd actually just translate πάρεστιν with a dative as "can" with a personal subject in English. Expressing modal constructions impersonally this way is really not at all unusual -- you see it a lot in other Indo-European languages as well, not just Greek, so I wouldn't really consider it a special metaphor unless you mean the whole group of related dative usages (possession, ethical, etc). We can do it in English, too ("it is possible for him" vs. "he can"), it's just not the preferred structure.
I have no problem with any of the above. My attempt to explain conceptual metaphors was pretty lame. We use them all time without being aware of them. Lakoff et al make a distinction between expressions like "pretty lame" obviously a metaphor in the traditional sense and spatial conceptual metaphors which are constructed as an extension of a container metaphor related to our bodies. The latter are built into our language/culture and are not generally a matter of language user choice. In the last sentence the expression "language/culture" is treated like a container which holds things thus we say "built into" which is a spatial metaphor. Lakoff considers this an extension of body language, i.e., language extended from how we conceptualize our bodies as containers. I am not a disciple of Lakoff, he is just another linguist, hadn't heard of him until Michael Aubrey started talking about him.
I was spinning off the πάρεστιν LSJ A 2 "to be by or near one, c. dat." Thinking that the ability of someone to do something which you gloss as "can" may be conceptualized spatially as something (an ability) metaphorically being with some person or agent. Again, this analysis does not imply any sort of conscious choice on the part of the language user and thus not related to rhetorical strategies. The conceptual metaphors are part of the architecture of the language at a very foundational level.
I think it has to do with argument structure ("arguments" being here not rhetorical strategies but bits of a sentence which are ordered by the verb -- i.e., subjects, direct objects etc). When we think about participants in a sentence we can talk about their roles a couple of ways. First they have a syntactic function (grammatical subject or object). But they also have a semantic function ("thematic role"): agent, experiencer, recipient, instrument, etc. There are various theories about how thematic roles "map" into syntactic functions, but basically different languages do this different ways. I think English has a strong preference for filling the subject position of a sentence with an "agent" or "experiencer"-type noun. "The dog bit the cat." "I like pie." In Classical Greek (and other languages) this preference isn't nearly so pronounced. You're more likely to get sentences like "Pie is pleasing to me."
I agree that πάρεστιν with the dative relative pronoun has to do with argument structure. Putting the agent into a dative relative pronoun is probably a form of backgrounding if viewed from the English language perspective. That is one of hazards of text linguistics, we always lean toward an analysis which makes sense in our native language. It may be that an agent in Classical Greek can appear in the dative without reduced salience. I don't recall seeing anyone arguing that point however.
Notice the spatial metaphor "lean toward an analysis" which is probably an extension of what Lakoff calls conceptual metaphor.
C. Stirling Bartholomew
C. Stirling Bartholomew