Concepts and circularity
Concepts (categories of objects) are simplifications of an underlying, continuous/fluid reality, for the purpose of procedure (regular actions in response to regular perceptual cues that yield regular outcomes).
They consist of the visible-observed (that is, of surrogates used to identify categorical belonging in the first place), and also those unseen properties which can be deduced with some reliability from the categorical belonging decided, properties whose pragmatic value underlies (“pays for”) the creation and upkeep of the concept in the first place.
Circularity emerges when the properties used to determine membership are the same properties deduced from belonging. To define a set as possessing X property, and then remark (as if insightfully) that all members possess X property, is to say nothing at all. Such investigations are solely lexicographic.
“Is” the concept the inferred properties depended upon for practical activity, or the surrogates used to identify belonging? Neither/both.
Sometimes we get confused what we care about: Are we speaking about “class” as the composite of surrogates which signal and communicate class, or as the abstract processes and individual realities which give rise to the expression of these surrogates?
Wrinkle (for the author or readers to solve): Cases where the category exists for coordinative purposes, and the surrogate is arguably “the point” (to ensure synergistic behavior and the ability to communicate about the thing itself).