natural kind: a standalone construct

John Stuart Mill's 1843 A System of Logic introduced the idea of a real Kind: a class of things, such as a chemical element or a biological species, whose members share an open-ended, inexhaustible cluster of properties, as opposed to a class fixed by human convention such as all objects weighing one pound. W. V. Quine's 1969 essay Natural Kinds made the term central to modern philosophy of science by asking why the groupings that support successful induction, that is, projecting a finding from a few cases to the whole class, seem to track the world's own divisions rather than our labels. Saul Kripke and Hilary Putnam later argued that terms for natural kinds, like water or gold, refer by way of an underlying nature fixed by the world (H2O, atomic number 79) rather than by a description held in the speaker's head, and Richard Boyd's 1991 account recast many such kinds as homeostatic property clusters, families of properties held together by shared underlying mechanisms. The construct is distinctive as philosophy of science's standard tool for asking which classifications are discovered rather than invented, the ones whose membership carries explanatory and predictive weight across a discipline. Because the same properties recur reliably among its members, a natural kind is the category that licenses generalising from samples to the whole class.[1][2][3][4]

Unit

A class of objects unified by a shared, open-ended cluster of properties and picked out by a scientific classification

Recovery means

Not applicable

Measured by

Identified by whether membership supports reliable inductive inference and figures in explanatory laws, often traced to a shared underlying nature or mechanism