Notes

1. 'A model which took account of all the variegation of reality would be of no more use than a map at the scale of one to one' (Robinson, 1962, with a reference to Lewis Carroll, 'Sylvie and Bruno').

2. Sanjaya Lall has shown in a series of important papers that, from 1979 onwards Indian technological capabilities have accumulated to the point where India is an exporter of technologies. (See, for example, Lall, 1984a.)

3. Dosi, Pavitt and Soete (1°90) suggest that Marshallian competition might be a limiting cave appropriate to situations in which innovation teas ceased.

4. Economic theory teas often lagged behind changes in economic systems, or changes in policy. There are many examples, of which the Prebisch justification for import substitution is especially well known. For speculation along these lines about the Prebisch-Singer thesis, see Cooper and Fitzgerald (1989, p. 12). More appositely, it is worth noting that changes in perceptions of the innovative process are quite frequency ascribed to historical changes affecting industry structures. See, for example, Freeman (op cit., p. 10) on Schumpeter's change in view on sources of innovation, in particular his recognition of the role of intrafirm research and development.

5. There are useful surveys in Dosi (1988) and Dosi, Pavitt and Soete (1990).

6. The concept of 'quasi-monopoly' as applied to the temporary advantage of an innovating firm was first used by Vernon (1966).

7. The specifications which follow are drawn in part from the discussion in Dosi, Pavitt and Soete (1990, Chapter 4 on 'The Innovative Process'), which in turn draws on a large body of theoretical and empirical literature.

8. Frances Stewart and Jeffrey James (1982, pp. 9 ff.) apply the idea to the context of developing economies. Elsewhere Stewart (1979) used the localization of technological change in the Atkinson and Stiglitz style to explain why technology from industrialized countries is inappropriate for developing countries.

9. David makes much wider inferences, well beyond the idea of localization: 'Choices of technique become the link through which prevailing economic conditions may influence the future dimensions of technological knowledge' (op. cit., p. 4). There is a useful discussion of these ideas in Rosenberg (1982, p. 166 et seq.).

10. Rosenberg (1982, pp. 120 ff.) broadens the concept to cover 'reaming by using'. The automaticity of the learning process as formulated by Arrow (op. cit.) has been strongly criticized by a number of authors, including Bell et al. (1982, 1984).

11. Or what Dosi describes as 'cumulation ... in the acquisition of problem solving capacities' (Dosi, 1988, p. 1128).

12. In his first formulation of the economics of innovation, Schumpeter was much concerned to explain monopolistic rents from innovations as a return to entrepreneurship (Schumpeter, op. cit.). In a modem setting, the role of these rents is much simpler to understand: they are perceived by businessmen as the 'return' to the (often massive) allocation of resources (like R&D resources) to the production of innovations.

13. The history of industrialization (and of industrial technology) is in a sense an account of sectoral patterns of innovation which change over time: innovations ID the United Kingdom during the Industrial Revolution were centred in the textiles vector itself. Later, as the capital goods sector emerged, innovations in textiles machinery and many other types of equipment came more and more from machine-making sectors. Further in response to bottlenecks in textile inputs (bleaching agents and dyestuffs in the first instance), the chemicals industry-the first 'science-based' vector- emerged. The story continues: it is hardly surprising that there should be intersectoral differences in innovation in modern industry which is a product of the historic process.

14. Discussions of these chapters of Capital can be found in Rosenberg (1982, Chapter 2, an essay reprinted from the Monthly Review, Vol. 28, July-August 1975). See also Cooper (1971). Marshall also reflects on the emergence of new forms of industrial organization and their relationship to the application of science to production (Marshall, 1899, Book IV, Chapters VIII and IX).

15. Freeman (1989) points to the relationship between innovation and product differentiation. He describes 'defensive' innovation (which is highly R&D intensive), as follows: 'Defensive R and D is probably typical of most oligopolistic markets and is closely linked to product differentiation' (p. 176).

16. Cooper et al. (1974) attempted to link the Sylos-Labini framework as developed by Modigliani (1958) to explain barriers to entry and intra-industry patterns of competition.

17. See Mansfield et al. (1981): 'It has long been recognised that the costs of imitating new products have an important effect on the incentives for innovation.... If ... firms can imitate an innovation ... substantially below the cost to the innovator of developing the innovation, there may be little or no incentive for the innovator to carry out the innovation'.

18. On the importance of incremental improvements to production processes see Enos (1962).

19. In fact, Freeman distinguishes six typos. His sixth category, 'Opportunist strategy', is not necessary for our discussion.

20. Freeman may understate the significance of innovation in the traditional sectors, or at least the technological pressure under which such firms have come, especially in recent times. Traditional firms are generally 'supplier dominated', and may from time to time be faced with a critical need to adopt new technologies as a matter of survival (or as an interim measure to drive costs down as far as possible by real wage reductions). This is evidently happening in a number of vectors because of the incidence of new microelectronic control systems, which widen the range for mechanization (see Hoffman and Rush, 1988, for a discussion of the garments industry).

21. An early attempt to specify 'Channels and Mechanisms of Technology Transfer', which draws this distinction, is given by Cooper and Sercovich (1971). There is a survey of literature on this question in Stewart (1981).

22. There is some evidence that European technology-supplier firms to Indian industry are increasingly sensitive to the creation of potential compositors, especially in view of increasing composition in their home markets from the newly industrialized countries (see Cooper, 1988). But Bell and Scott-Kemmis (1988) express doubt whether this is two of British firms.

23. 'It is obvious that there is no room in economics for long chins of deductive reasoning' (Marshall, 1899, edition 1966, p. 644). It is a little unclear whether Marshall intended this as a statement of fact or as a normative judgement. As 9 statement of fact, it has not boon borne out by recent developments in economics, many of which bear witness to a willingness to pursue long chains of deduction, to the almost total exclusion of embarrassing facts. Since Marshall was well aware of the methodological weaknesses of economists, it seems to me more reasonable to regard this as a normative statement-a stern warning in fact.

24. An important qualification to this is implicit in Freeman's further remarks that 'Unless imitators enjoy significant market protection or privilege they must rely on lower unit costs of production', and, in relation to innovative strategy in developing countries, 'even a successful imitative strategy, although it may lead to industrial development, will reach a point where export competitiveness in labour costs may increasingly conflict with the goal of higher per capita incomes' (op. cit., pp. 170 and 184). These arguments as they apply to trade patterns will be developed more fully below.

25. See, for example, Dahlmann and Westphal (1982), and Bell d al. (1982) for a review of some of these. See also Dahlmann et al. (1987), and Pack and Westphal (1986).

26. Subsequently, arguments about tendencies to sub-optimal investments as a justification for interventions by a social welfare state were largely brushed aside by the crude empiricism that 'bad markets are better than bad governments'. This of course is just as impossible of proof as was the crude delinking argument of the descriptive Dependency school.

27. The best known attempt at superordination is Chenery, 1961.

28. For a detailed history of Indian policy and experience see Nayar (1983). A critical accounts of Indian policy is in Desai (1988).