The fundamental point to be grasped here is a point that has been made.  The absolute resides not on the level of sensible presentations but in the field of abstract propositions and invariant expressions. (CWL 3, 166/189-90)

Science does not advance by deducing new conclusions from old premises.  Deduction is an operation that occurs only in the field of concepts and propositions.  But the advance of science, as we have seen, is a circuit, from data to inquiry, from inquiry to insight, from insight to the formulation of premises and the deduction of their implications, from such formulation to material operations, which yield fresh data and, in the limit,  generate the new set of insights named a higher viewpoint.  A basic revision, then, is a leap.  At a stroke, it is a grasp of the insufficiency both of the old laws and of the old standards. At a stroke, it generates both the new laws and the new standards.  Finally, by the same verification, it establishes that both the new laws and the new standards satisfy the data. (CWL 3, 166/190)

What holds for standards, also holds for their use.  It is necessary to define as accurately as possible the precise type of magnitude that is to be measured.  It is necessary to define the precise procedure that leads from the measurable magnitude and the standard unit to the determination of the number named a measurement.  At each stage in the development of a science, these definitions will be formed in the light of acquired or presumed knowledge.  But at every subsequent stage, there is the possibility of further acquisitions and of new presumptions and so of the revision of the definitions.  Such a revision involves, not the deduction of new conclusions from old premises, but a leap to fresh premises.  Such is the generic notion of measurement. Clearly, it contains within itself the possibility of successive differentiations that result from revisions that occur in the abstract field of definitions, principles, and laws.  We have now to turn our attention to the revision involved in the notions of spatial and temporal measurements by the Special Theory of Relativity. (CWL 3, 166-67/190-91)