Quantum weak values are not unique; what do they actually measure?

Precise definitions of "weak [quantum] measurements" and "weak value" [of a quantum observable] are offered, which seem to capture the meaning of the often vague ways that these terms are used in the literature. Simple finite dimensional examples are given showing that weak values of an observable are not unique, and in fact arbitrary weak values can be obtained by appropriate weak measurements. This implies that a "weak value" of an observable A, *by itself*, can furnish no unambiguous information about A; any information in a weak value is inextricably connected with the particular measurement procedure used to obtain that weak value. Moreover, arbitrary weak values can be obtained using a "meter space" of dimension as small as 2. A "Remarks" section questions the utility of "weak measurement".