This sort of scenario was so undesirable that societies must have created money to facilitate trade, argues Smith. Aristotle had similar ideas, and they’re by now a fixture in just about every introductory economics textbook. “In simple, early economies, people engaged in barter,” reads one. (“The American Indian with a pony to dispose of had to wait until he met another Indian who wanted a pony and at the same time was able and willing to give for it a blanket or other commodity that he himself desired,” read an earlier one.)

The same thing holds true of all other possessions; for barter, in general, had its original beginning in nature, some men having a surplus, others too little of what was necessary for them: hence it is evident, that the selling provisions for money is not according to the natural use of things; for they were obliged to use barter for those things which they wanted; but it is plain that barter could have no place in the first, that is to say, in family society; but must have begun when the number of those who composed the community was enlarged: for the first of these had all things in common; but when they came to be separated they were obliged to exchange with each other many different things which both parties wanted.
On paper, this sounds a bit like delayed barter, but it bears some significant differences. For one thing, it’s much more efficient than Smith’s idea of a barter system, since it doesn’t depend on each person simultaneously having what the other wants. It’s also not tit for tat: No one ever assigns a specific value to the meat or cake or house-building labor, meaning debts can’t be transferred.