It is a curiosity that anthropological attempts to describe non-monetary transactions may well end up concentrating on their monetary aspects. This can happen in specifying how persons measure the relative worth of the things they transact. It has become conventional to admit a range of ‘social’ factors in the computation of relative worths, and the longstanding debate about the embeddedness of transactions in relationships is now taken for granted. Nevertheless, the principles by which comparability is established between the things themselves invariably remains the analytical focus, whether these things are material or metaphysical in nature, and whether the transactions are balanced or unequal. The procedure incorporates what one could call a barter model of value.

Barter has always been a somewhat mixed analytical concept. Anthropological theories of exchange systems have tended to take barter, as indeed they have taken trade, as a self-evident activity. For barter appeared to ‘provide the imagined preconditions for the emergence of money’ (see Introduction and Humphrey, 1985: 48). It was regarded as a strategy through which people obtained things that they needed, that is, doing in the absence of money what people with money also do. In short, it was understood in ‘monetary’ terms. The discovery of the gift, however, was the discovery of people exchanging things which they did not need, and unpacking that paradox has dominated anthropological theorising on exchange ever since.