Adults, under ideal conditions (i.e., given unlimited time), evaluate which of two sets is "most" by counting up the objects in each and determining which has the higher count. So, does this mean that one must be able to count in order to understand "most"? Halberda et al 2008 set out to answer this question, and found that the answer is "no": some kids that can't count and even some that can evaluate "most" using the Approximate Number System. What does this imply about how meanings link up with other systems of mind, and how such linkings change over the course of development? Lidz et al 2011 advance the idea that the meaning of a sentence is a "canonical specification of truth conditions" or "specification of a default verification procedure", is this right? We will look at the results of several experiments from the UMD-JHU group to try to unpack the question of the relationship between linguistic meaning and extralinguistic verification.