Abstract
It is not always clear to what extent the logic and vocabulary of measurement as used in different scientific disciplines are mutually coherent, nor how measurement can be demarcated from, say, opinion. In recent decades there have been a number of attempts to provide necessary and/or sufficient sets of conditions for when measurement is achieved, usually in terms either of inputs (e.g., whether an evaluated property is a quantity), or outputs (e.g., whether a procedure assigns numbers according to a rule). We argue instead that the public trust attributed to measurement is best justified in terms of the structural features of the process rather than of its inputs or outputs.