On standard accounts, comparatives like those in (1) all represent measurement of the same thing: coffee. How measurement proceeds is determined in (1a) by the meaning of "much", in (1b) by the meaning of "hot", and in (1c) by the meaning of "many". The same goes, mutatis mutandis, for the comparatives in (2), where running events are measured. This accords with intuition.
(1a) Al has as much coffee as Bill does.
(1b) Al has as hot coffee as Bill does.
(1c) Al has as many coffees as Bill does.
(2a) Al ran in the park as much as Bill did.
(2b) Al ran in the park as quickly as Bill did.
(2c) Al ran to the park as much as Bill did.
I review the semantic evidence for the standard picture, and provide several reasons to doubt it. Consequently, I argue for a radical reanalysis in which different things are measured in (1) and (2): the (a) examples represent measurement of masses/processes, the (b) examples represent measurement of states, and the (c) examples represent measurement of pluralities. "much" determines how measurement proceeds in each case.
This analysis has some novel consequences. It provides a uniform account of when comparatives are felicitous, regardless of the syntactic category of the expression targeted for comparison; it predicts widespread grammatical effects on the available dimensions for measurement, which are observed; and it characterizes speakers' implicit understanding of "degree constructions" as explicitly measurement-theoretic.