A concept of utility which is based on the idea of a preference ordering, or ranking, rather than on the Marshallian concept of measurable utility. It is assumed that a consumer is capable of comparing any two alternative 'bundles' of goods and deciding whether he prefers one to the other or is indifferent between them. Several assumptions are made about the nature of this preference ordering, which lead to the conclusion that all possible bundles of goods can be grouped into sets in such a way that the consumer is indifferent between all the bundles in one set and is not indifferent between sets. These 'indifference sets' can be arranged in increasing order of preference. It is often convenient, though not necessary, to assign numbers to these sets, adopting the convention that the higher a set is in the order of preference, the higher its number should be. Any increasing sequence of numbers could perform this function. For example, we could indicate successive indifference sets by the sequence 1, 2, 3, 4 .. , by the sequence 1, 10, 2,000, 2,001 ... or by any other rising sequence. It has become the practice to call such numbers 'uttlity numbers', and, since they simply describe the ordering of indifference sets, they are said to represent 'ordinal utility'. This usage is rather unfortunate, since the word 'utility' bears no relation to the Marshallian notion of the intrinsic satisfactions induced by the consumption of a good. Indeed the theory of ordinal utility has tried to purge itself entirely of this notion.
|Reference: The Penguin Dictionary og Economics, 3rd edt.|