We often think of quantifiers like “each” and “every” as expressing a relation between two sets. On this view, the meaning of “each/every circle is green” is something like “the circles are a subset of the green things”. This correctly captures the relevant truth-conditions, but does it accurately describe how these quantifiers represented in speakers’ minds? In this talk, I’ll present data from adults and kids to argue that “each” and “every” have as their meanings different representations. “Each” expresses first-order quantification over individuals, with no reference to sets or groups as such. Meanwhile, the representation for “every” is second-order, in making reference to sets, but it is not a relation between two sets. It is a one-place quantifier over its ‘scope’ relativized to a group expressed by its ‘restriction.’ If this story about “every” generalizes to other quantifiers, it might offer a way to account for the well-studied semantic universal that all quantifiers are conservative.