The report of the Economic Science is not always that of a discipline in search of recognition. The economic analysis has been useful to people who were thinking about how science and scientific research work or should be organized. They also could be used to rethink more traditional issues, such as the justification of scientific knowledge. That's what it will be discussed here.

During the second half of the 20th century, philosophers of science in connection with economists have become interested in a possible analogy between science and economy. Scientists may be regarded as agents that produce a certain type of property, skill, and trying to sell to buyers''''potential, their colleagues

that they are trying to convince. However, the analogy with the market is not perfect, because the knowledge produced is not a private good in the classic sense. The''currency''of the scientific community is not the usual money: scientists are not directly rewarded for their discoveries, or to publish an article. It is fair to consider that they receive the consideration from their peers. The currency is then the recognition or credit they may have. It is therefore quite tempting to try to understand the behavior of scientists in their research activities, assuming they are rational actors seeking to maximize their credit.

Of course, motivation more natural and noble for scientists, in a sense, is the pursuit of truth. It is the assumption usually made. But we must not forget that Scientists are still primarily men. Philosophers of science with some sociologists have pointed out forcefully in the 1970s: Woolgar, Latour example. After all, perhaps, that science was a''social construction''as another. The question of truth in science, and its allegedly objective was well rested.

The assumption that scientists are motivated by credit and provides an interesting look, even if it is a simplification. This event will allow the question of reconciling the goal of selfish agents (maximizing their credit), and the overall goal epistemic (that is to say an object of knowledge) of the entire community. In any case, is the goal that the company would give them, since it pays them to work.

One can distinguish at least two problems: the truth of knowledge and of work organization. For the former, the motivation of scientists is most directly the truth too, can we continue to argue that scientific knowledge remains an objective? On the other hand, one wonders if the selfish motivation will not be detrimental to overall progress: the self-organization of selfish agents, it allows a good efficiency in the production of knowledge?

Many authors have attempted to answer these two questions in the affirmative. For the first problem, that of truth, a way to respond was based on the actual organization of the research scientists. Verification systems, and incentives should allow scientists to ensure the objectivity of the result. For example, there is science tradition of trying to replicate the experiments or the results to be published. Thus, philosophers have sought to show he was not in the interest of agents to try to cheat. The concept of rationality among agents on the scene, as it does in economics.

For the second problem, the organization of work, the question of''invisible hand''- similar to that of Adam Smith in economics - has been invoked explicitly or implicitly (Hull, Goldman, Kitcher). In short, the idea is that individual agents do not intend to produce a certain result, however, that arises as a byproduct of their activity selfish it is not necessary that officers are aware of this phenomenon that occurs, the result seems to have been created intentionally.

Apart from the question of''invisible hand'', many models have been proposed to understand the production of scientific knowledge or the division of work among researchers. What is the optimal strategy to adopt towards a panel of various competing theories, and between which to choose?

We can not complete the idea of an economic analysis of scientific research, without raising the question of policy research, including an evaluation. Many authors have noted the difficulty of such an approach, including the risks of an overly simplistic approach. The models we have discussed here can afford the very simplifying assumptions, because they have the advantage of studying some salient features: for example, influence the motivation of credit, not the truth. But when it comes to truly evaluate the scientific production, and the researchers themselves, indicators should be chosen with great care, otherwise they may well not try to optimize something that has no interest (eg an article by Yves Gingras on this subject).