New economics of science, economics of scientific knowledge and sociology of science: the case of Paul David

Abstract

For a little more than twenty years, the terminology used in the economics of science has changed significantly with the development of expressions such as ‘new economics of science’ (NES) and ‘economics of scientific knowledge’ (ESK). This article seeks to shed light on the use of these different terminologies by studying the work of the economist of science Paul David. We aim to use his work as a case study in order to argue for a difference between NES and ESK and to show, in a concrete way, the sociological ambiguities now going on in the economics of science.

Keywords:

• economics of science
• new economics of science
• economics of scientific knowledge
• sociology of science
• sociology of scientific knowledge
• Paul David

Acknowledgements

This paper was presented at the IX Conference of the International Network for Economic Methodology (2011). For very helpful comments and suggestions, I would like to thank Gildas Appere, Yves Gingras, Franck Jovanovic, D. Wade Hands, Philippe Le Gall, Muriel Travers, Zamora Bonilla and two anonymous referees. I also thank Gemma Davies for her assistance in the editing process. The usual disclaimer applies.

Notes

 1. Although the literature on ESK most often uses the term ‘content’ to deal with epistemological questions, we instead use the term ‘epistemic’ which is more appropriate (I thank Yves Gingras for this remark).

 2. We do not mean that there was no economics study of science before this period. Indeed, one of the precursors is the American pragmatist philosopher C.S. Peirce with his ‘Note on the Theory of the Economy of Research’ (1879) (see Rescher 1976; Wible 1994a, 2008).

 3. A first distinction between NES and ESK appears: contrary to ESK, NES gets an article that institutionalizes and lays the groundwork for its research program. Some institutional elements are important to understand why the term ‘new economics of science’ is more widespread than ‘economics of scientific knowledge’ among the economic profession. We can take the example of the ‘Conference on the Need for a New Economics of Science’, hold at the University of Notre Dame in 1997 and note the ambiguous assimilation between NES and ESK in its call for papers (http://www.mail-archive.com/[email protected]/msg11948.html – emphasis added):

The different perspectives on (quasi-)economics of science and/or scientific knowledge can be organized in terms of old and new economics of science. Old economics of science consists of an institutional approach to science, an argument that science is a market, a unity-of science approach, and a clear definition of the organizational framework of scientific research. New economics of science consists of a contextual approach to science, an argument that science cannot be commodified, a disunity-of science approach, and a questioning of the units of organization in science.

 4. The literature on ESK most often only cites SSK. However, critical developments by STS authors should also be considered.

 5. This working definition does not say anything about the characteristics of scientists themselves. This is so because our goal is to define the scope of ESK in the most consistent way, without dealing, for example, with the kind of rationality attributed to scientists by economists. This latter issue is, however, major and should be dealt with in future research (see Downes 2001).

 6. This concept is an old one. Indeed, it was used by Robert Boyle in the seventeenth century and was more recently popularized by Price (1963). For a discussion of this concept, see Collins (1974) and Crane (1969). For David, rational agents maximizing utility are interested in joining these colleges for two main reasons. The first is the ‘exchange value’ of information: because of the cognitive division of labor between researchers, participating in networks allows an individual researcher to be able to resolve its potential problems much more effectively than in isolated conditions. The second reason why to join these networks is that they confer a status and are a sign of recognition from the scientific community (ibid., p. 127).

 7. Markov random field theory can be defined as ‘a branch of probability theory that has lent itself to application in the study of social networks’ (David 1998, p. 147).

 8. Formally, a process cannot be said to be path dependent if it has not at least two absorbing states.

 9. Percolation theory ‘conjures up the image of droplets of water moving under the deterministic pull of gravity through a dis-ordered, random medium – such as filtration tank filled with sand and pebbles of different sizes. When the water, entering at some source sites, eventually finds its way into enough open channels to pass throughout, wetting the entirety of the interior surfaces, complete percolation is said to have taken place’ (David 1998, p. 151). Here, this means that the network has to get some specific features for a cognitive function to successfully ‘percolate’ throughout the network.

10. See also Ylikoski (2003) for an interesting discussion of Dasgupta's and David's use of the Mertonian framework.

11. These are the two main points treated by the authors, but they also cite, for example, the concept of ‘multiple discoveries’ (Dasgupta and David 1994, p. 506).

12. David 1998, p. 119.

13. Indeed: ‘the following parallels material developed in Dasgupta and David (1994)’ (ibid., 129 f. n. 22).

14. We will not deal here with the subtleties of Latour's definition of a ‘network’.

15. We do not mean that these issues are exclusive.

16. We can also find passages in David's paper where he links macro level of analysis to institutional questions. For example: ‘One wants at the end of the day to know how the workings of these micro-level networks are connected to the properties of the macro-structure of scientific institutions’ (David 1998, p. 123).

17. See the critique of Giraud and Weintraub (2009) and Tyfield's (2009) reply.