Measurement expressions in Uzbek acquire different meanings when the same numeral and unit enter different grammatical constructions. This article examines how those differences arise and how they affect the presentation of quantity. The study uses qualitative constructional analysis of examples assembled in an Uzbek research manuscript, supplemented by published research on classifiers and independently checked textual instances. Schematic comparisons are distinguished from attested usage; neither is treated as evidence of corpus frequency. The analysis separates counting expressions from standard measures, bodily portions and container measures. It then traces the contribution of the suffixes -lik, -cha and -lab, case marking, predicates and evaluative modifiers. The results show that -lik presents a measured value as a property of a referent, while -cha marks an approximate value. Interpretations of -lab depend on whether the construction describes repeated portions, extensive quantity or figurative scale. Case-marked phrases identify limits, intervals, differences and attained values through their interaction with predicates and accompanying words. Percentage expressions make this dependence especially clear because a change and its endpoint can contain the same numeral but describe different relations. A constructional description therefore needs to record the dimension, numerical value, morphology, syntactic role and contextual interpretation together. The proposed distinctions support more accurate translation and provide categories for subsequent corpus annotation without implying that their distribution has already been measured.