I went with 'identities' because a fact is essentially an identification of something, while anything you might believe about that something prior to identifying it conclusively, is conceptual framework.
Take water, for instance. We know a number of facts about the substance known as water. We can identify the substance by observing the facts we have.
Water boils at around 100°c. Any liquid that doesn't boil at that temperature does not share that identifying trait, and does not have the same identity.
All our knowledge is based on assumptions. The key is in reducing the number of assumptions to what you really need, and remaining aware of them to be able to test and update them, but not strip yourself of important features and structures.
Yes, and it is also based on facts. But I have been wondering about this method. The ideal of having as few assumptions as possible stems from a time where people thought it was actually possible to have no assumptions at all. But if you can never really eliminate assumptions completely, is there a point in minimizing them?
Wouldn't it be better then to have more assumptions, and even assumptions that contradict each other? That would decrease the risk of your assumptions being mistaken for facts, and if you are using your assumption to shape the questions you ask, having contradictory assumptions may help against being misled by getting good answers to the wrong questions.