28 SOCIAL SCIENTIST 1 The author would particularly like to thank Tom Asimakopulos, Frank Hahn, Geoff Harcourt and Elman Wolfstetter for their helpful comments. 2 For sophisticated expositions of neoclassical theory see, for example. Arrow and Hahn (1971) or Debreu (1959). For a less sophisticated exposition see Ferguson (1969) or, indeed, almost any conventional economics textbook. 8 Debreu, 1959, pp 78-80. 4 cf Bachelard, especially chapter 1. 5 Karl Marx, Theories of Surplus Value, Vol 3, P 503. • For the classic statement of this view see Hicks, 1963, especially chapter XI. ? The major work of this school is undoubtedly Srafia (1960). A number of important articles by Garegnani (1970), Pasinetti (1969), Robinson (1953-54) and others are reprinted in Harcourt and Laing (1971) or Hunt and Schwartz (1971). 8 The exchange value of any bundle of commodities depends on two factors : its physical composition and the price of each individual commodity in terms of the numeraire (money). In any comparison between two states of an economy, therefore, differences in the exchange value of capital or output may reflect either or both of of these factors. Let us assume that an equal amount of labour is performed in the two states and that each state is, in the conventional technological sense, efficient. By comparing one state with the other, we can define the marginal product of capital as the difference in output divided by the difference in constant capital, where all magnitudes are reduced to a common standard by expressing them in current prices i. e, in terms of their exchange value relative to the numeraire commodity. It is clear that this marginal product will reflect, not only physical differences between the two states, but also differences in relative prices. The exchange value of capital, for example may be higher in one state than another, not because its physical components are different, but because they have different prices. It is this shift in relative prices which accounts for the possible non-equality of the rate of interest and the marginal product, even when appropriate neoclassical assumptions are made. To get around this difficulty, neoclassical economists use a different definition of marginal product, and proceed as though relative prices in the two states were identical. As a device for micro, partial, or efficiency analysis, this approach is acceptable. In the study of income distribution, however, it is not, and differences in relative prices between the two states must be taken into account. The role of relative prices in this context was first recognised by Wicksell (1934). For good discussions of the subject see Metzler (1951) ^and Malinvaud (1953). 9 Karl Marx, Selected Works, p 227. 10 Both Garegnani (1970) and Pasinetti (1972) base their rejection of supply and demand theories, including general equilibrium theory, on the fact that aggregate capital* may not be a continuous function of the rate of profit. It is not clear from their arguments, however, why supply and demand theories must necessarily depend upon such continuity. It is true that traditional aggregate theory required this kind of continuity, but no such assumption is made by general equlibrium theory, which allows for the possibility of aggregate capital being a discontinuous function of the rate of profit. Whatever its other failings and they are many, general equilibrium theory is not open to this particular objection. 11 cf Dobb, 1970. 12 Karl Marx, Selected Works, p 182. 18 Karl Marx, Capital, Vol 1, 184-185. 14 Bharadwaj, 1963. " Karl Marx, Capital, Vol I, p 356, 16 Marx used the term 'merchant capital1 in two different senses : sometimes to denote commercial capital in general, even when it operated under competitive conditions, and other times to denote capital which made a profit by cheating or mono-ply in the sphere of exchange. In the text the term is used in this second and narrower sense.