If you can never assume that any 2 concepts are distinct from each other, how do you model your database? I'd be curious how you would model the example you stated as a relational database that mimics how humans think