the rationale for the positive-normative distinction remains unchanged. Because its hypotheses are susceptible> to rational discussion by empirical evaluation, positive economics is considered to be scientific. On the other hand, rational discussion about normative economics is not possible because it is concerned with value judgements which are, by their very nature, a matter of subjective choice.
There are some significant theoretical developments that undermine the justification for the positive-normative distinction. Rational discussion about value judgements has been shown to be possible. The best example is the so-called impossibility theorem of Arrow (1951), that it is impossible to have a social welfare function which satisfies a set of five seemingly innocuous assumptions about individual and social rationality.Another example is the analysis of the structure and nature of value judgements. Sen (1970) has shown that many value judgements are held conditionally; they may be given up if factual circumstances change.
Economists are beginning to appreciate that the 'choice between competing empirical hypotheses on the basis of testability, refutability or any other criterion is not a simple matter and involves methodological judgements which like value judgements are not empirical (Boland, 1982).
Simple concepts, fundamental to economics, such as price and quantity do not correspond to observable entitles and are, in fact, theoretical (Wong, 1978).
In giving up the justification for the positive-normative distinction, there is no further commitment to abandon the distinction altogether. The distinction between facts and values is important. It is not, however, the only methodological prescription that should be followed. Rational discussion of economics would also be significantly enhanced by following the dictum of Gunner Myrdal (1932) and adopted by Joan Robinson (1962) that values, whether ethical, political or methodological, which impinge on the selection of problems considered by economics and the theoretical frameworks chosen to address them, should be explicitly acknowledged and open to debate.