POC: Alternative way of determining parameters

[teunbrand]

This PR explores to fix #1516.

It follows the approach proposed in https://github.com/tidyverse/ggplot2/issues/1516#issuecomment-1405478064. I've numbered the comments in the code below so it'd be easier to follow.

This PR is currently only implemented for 1 stat and 2 geoms, complying with 'a minimal version' of the approach. It is for this reason still a POC, and we still have to decide if we like this approach.