personally, i’m fine with the way the existing observation field filtering works now. observation fields are meant to be sort of a better-than-nothing way of recording extra stuff onto an observation. so trying to add too much functionality here is probably not a great path, in my opinion.
i’m not exactly sure why you want to query data this way, but you should know that GloBI aggregates properly licensed observations from iNaturalist and other sources of data, and it might provide a nicer way for you to query for various interactions between different taxa. GloBI’s search results come with references back to their source observations, too.
if you’re trying to stay within the iNat data ecosystem only, then the way i would handle your case is to get all observations that contain Nectar Plant as an observation field, and then do my own filtering for whatever taxa you’re interested in. this is probably most easily accomplished using some sort of scripting language. for example, the Python-based Juypter Notebook refernced in the original post in this thread could be adapted to handle that sort of thing: https://forum.inaturalist.org/t/whats-the-best-way-to-share-python-code-nowadays/48554.