This commentary examines the role of the unit from the perspective of the definition of measurement in physics as the ratio of two magnitudes, one of which is defined as the unit; it is an important and timely contribution to measurement in the social sciences. There are many different points that could be commented upon, but the author will emphasize just four. First, the point that every analysis of a set of responses using the Rasch model has its own unit, and is a property of the set of responses, is important to make explicit, not only in words as it has been done elsewhere, but also algebraically in the model as has been done in this article. Second, it is instructive to see the relationship between the unit and discrimination made explicit and a new perspective on the relationship between the logistic measurement function of the Rasch model and the 2-parameter logistic (2PL) model. Third, Humphry stresses that measurement is not, as written by Stevens (1946) and copied by many others (though not all), just an assignment of numbers according to some rule. Fourth, and most distinctively, Humphry's formalization of empirical factors that might govern the units in any analysis is a major contribution.