before you have a solid ontological understanding of a concept, lacking powerful descriptive axioms we're forced to rely on a list of characteristics e.g. intrinsic properties, how it interacts with other objects, etc. if we restrict ourselves to things we already have a vigorous understanding of how will we consider new hypothesis?
by establishing the concept of scope and discriminatory criteria for inclusion or exclusion. Look up 'bongard problems' and appreciate that while some are based around topological or temporal progression, others require cataloging random arrangements and establishing membership criteria.
