This paper suggests a new way of thinking about public goods. By making assumptions about the utility functions of consumers along the lines suggested by the work of Lancaster and Muth on the technology of consumption we obtain a new formulation of the public goods problem which appears to have some interesting implications. In one sense it can be seen as a special case of the original formulation by Samuelson, so that it is not inconsistent with his analysis. On the other hand it makes it possible to give a more precise analytical representation of properties that have for long been associated with public goods, and to derive Samuelson's "polar case" as the limiting case of a more comprehensive definition. It also suggests a more meaningful notion of the benefits derived from public goods supply, thereby providing some theoretical underpinnings for the extensive empirical work on that problem. Moreover, the formulation can easily be extended to cover congestion externalities, which are closely related to public goods supply, but which have so far not been satisfactorily integrated with the basic theory. Throughout the paper we are concerned with collective or public consumers' goods. The case of collective producers' goods, which has some important similarities with the present analysis, has been discussed in another paper.