Von Davier, Naemi, and Roberts (this issue) present a nice summary of the statistical ambiguity often encountered in making distinctions between qualitative and quantitative constructs. In this commentary, the author begins with two broad points. The first is that the mixture/factor arguments are most intriguing when firmly embedded in a theoretical or substantive context where the answer really matters. Problems in measurement often yield good examples. The first main point is that debates about whether a factor or a mixture model "really" applies should address: (1) Why does it matter?; and (2) What data will best address the mixture versus factor structure? The second broad point is that categorical constructs might map onto "real" phenomena that are "in the world," as it were, or they could be useful constructs that people "perceive" and "employ." People frequently perceive categories in which information is continuous--take color perception or even statistical inference, where "significant" and "not significant" interpretations too often sharply bisect a continuum of statistical information. It is not just the act of carrying out a hypothesis test that categorizes; the language of categorization carries forward into evidential reasoning. The second point is that people frequently (and productively) perceive and reason categorically. Reference to the "perspective" and "purpose" of the observer can be useful in understanding the role of quantitative and qualitative constructs. Von Davier et al. provide a very good "statistical" summary of the modeling challenges in trying to distinguish between mixture and factor models. They discuss the similarities among models that posit mixture or factor distributions, and how several model pairs make overlapping distributional predictions. In terms of content, the author thinks it would have been nice to see more attention to why the distinctions matter, and in what kinds of data.