I don't think I'm advocating an overexpressive form of complexity here but I feel I can see cases where it's difficult to avoid without having to jump through hoops uselessly. What if you have to setup a meeting with, let say, 7 individuals (7 invariants since they are all required to attend), what do you do? Do you create 7 actionable items, one per invariant or do you create one actionable item referencing 7 invariants. For simplicity's sake, I would have a tendency to want to create 1 actionable item with 7 invariants because it otherwise seems like overkill, to me, to create 7 action items if all you need to do is to ultimately create one email to 7 names to then have to check off 7 action items rather than 1.
Then, from any direction across the 7 invariants, I'm able to reconstruct knowledge about who else needs to be at the meeting if I need to remember who all needs to attend the meeting rather than having disjointed actionable items. I'm not saying there's no way to make this work but I am saying that adding expressive power can simplify task creation quite a bit while increasing the available wealth of information.
It's simply not realistic to have your "7 invariants" and to say "Hey.... hmmm... Bob's here, Mary's here, Dooku is here, Fred's here, Lakeisha is here, Algonquin is here, and Anastasia is here... let's check off all these "invariants," and see if there's anything I could be doing right now. HEY, WE'RE SUPPOSED TO HAVE A MEETING! PEOPLE, LET'S MEET!"