Where does aggregate collide?

Leibniz used the term aggregate (Latin aggregatum) from 1686 in his correspondence with Antoine Arnauld to name a collection of things whose unity is supplied by a perceiving mind rather than belonging to the thing itself: a heap, an army, or a body treated as one is an aggregate, real in its parts but not a single genuine substance. Lodge's 2001 study Leibniz's notion of an aggregate traced how the idea requires both a perceived relation among the parts and a mental act of treating them as one. The referent here is a category in metaphysics, the standing of a mere collection set against a true unity, and it has nothing to do with a total quantity in an economy, a filler in concrete, or a computed summary of database rows.[1][2]

Unit

One collection considered as if it were a single thing

Recovery means

Not applicable

Measured by

Not measured; argued for by analysis of the parts, their relations, and whether a genuine unity is present

Keynes introduced aggregate demand as a central construct in The general theory of employment, interest and money in 1936, defining it as the total spending that all buyers in an economy plan at a given level of income, and arguing that this economy-wide total, not the sum of isolated markets, sets output and employment. Hicks's 1937 paper formalized the link between that demand and interest rates in a model that became standard teaching. The referent is a measured macroeconomic quantity, scaled in units of national currency per period, and it is distinct from a metaphysical collection, a material in concrete, or a database operation.[3][4]

Unit

Total planned spending in an economy, in currency per period

Recovery means

Not applicable

Measured by

Estimated from national accounts as the sum of consumption, investment, government spending, and net exports

Chamberlin and Boyce built the functions COUNT, SUM, AVG, MAX, and MIN into the query language SEQUEL in 1974, so that a single expression could fold a whole set of rows into one summary value; these are the aggregate functions of relational databases. Klug's 1982 paper gave them a formal footing by proving that relational algebra and relational calculus stay equivalent once aggregate functions are added. The referent is an operation in a query language that collapses many records into a scalar, and it shares only the name with a total in an economy, a collection in metaphysics, or a filler in concrete.[5][6]

Unit

One value computed from a set of rows

Recovery means

Not applicable

Measured by

Returned by a query that groups rows and applies a function such as COUNT, SUM, or AVG