I dunno. I think the differences are more quantitative than qualitative, but I make an exception for EQs that model particular EQs if they get it right. For example, I really liked using the original Pultecs in the studio, because they had very gentle curves. I was glad when modeled versions came out that emulated that kind of response (and didn't have the hum!!). Granted, I could probably get the same curve with other EQs that have nothing to do with the Pultec, but if the interface is done right my muscle memory knows where to go to get the sounds I want.
Analog EQs can have
extremely complex responses that are a jambalaya of phase shift, ripple, ringing if inductors are involved, interaction among filter skirts, Miller effect with tube gear, etc. etc. At that point the EQ becomes more like a processor and most definitely, some processors "like" some sounds better than others. And i suppose that there would be some merit to cataloguing these preferences for future use...or maybe not. Maybe it's better to try a happy accident of using the EQ that sounded good on bass with snare. Who knows.
If my voice needs a +6 dB shelf starting at 2 kHz, then pretty much any EQ is going to give the quality I want to my voice. Spending time trying out different shelves might yield useful results, but spending that time backing up my data will likely yield more useful results in the long run.
I think the emotional impact of any given track is 85% the person playing it, 10% the sound of the instrument itself, and 5% the processing. That's not at all to diminish the importance of that 5%, but I'm just trying to put things in perspective.