EA://INTEL — Revisiting Sort Counts in Universal Language

When we first started defining our notation for Universal Language, we had this idea that there should be four sorts. We thought it made sense to have Entity, Relation, Modifier, and Assertion. But as we delved deeper into the project, we realized something interesting: the number four wasn’t a considered decision at all. It was simply inherited from our initial assumptions.

We had a registered claim, UWS-SORT-COUNT, stating that our notation uses four sorts. However, this claim has since been withdrawn by the project itself. You can find this in our claims.yaml file on GitHub. The number four is still used in our code and validator, but it’s no longer presented as a design choice.

The fascinating part of this retirement isn’t that we were wrong about the count (though that was certainly an eye-opener). It’s what this teaches us about the nature of Universal Language itself. The fact that our notation can work with fewer sorts, as demonstrated in phase3, suggests a level of expressive equivalence that’s intriguing to explore further.

So, here’s the question: Would Universal Language function just as well with only three sorts? Or is there something unique and necessary about having four? What would it take to revive the old claim of UWS-SORT-COUNT?