My empirical investigation of research groups shows that community- focused approaches to the division of epistemic labor, such as Kitcher’s, do not sit well with research group collaboration. Rather, distinct group- focused approaches are needed because the division of labor at the group level raises epistemological issues that community-focused approaches typically sideline. To conclude this chapter, let me outline four epistemological aspects that group-focused approaches need to account for.
First of all, my comparative study of research group collaboration reinforces the argument that not only non-epistemic but also epistemic factors trigger the division of epistemic labor. For the groups observed, the prime trigger for the division of labor among group members is the distribution of expertise, an uneven distribution that either stems from the research group’s interdisciplinary composition or from the fact that the group features a strong gradient in seniority. In the planetary science group, division of labor capitalizes upon the uneven distribution of expertise, along with the material research infrastructure, among group members. In fact, the planetary science group exists as a group because it makes an interdisciplinary pool of resources available that group members otherwise, within mono-disciplinary research contexts, would not be able to access. The division of labor in the molecular biology laboratory, in contrast, takes into account the expertise that group members have, or lack, a lack that is to be mended by dividing the labor so as to create learning opportunities.
Second, the issue of learning draws attention to a peculiar aspect in which group-concerned accounts have to diverge palpably from community-concerned accounts of the division of labor. While community-concerned approaches such as Kitcher’s tend to cast the division of labor in the light of differentiation, the case of the molecular biology laboratory particularly shows that the division of labor can be a means of convergence through learning, and need not be a means of differentiation.
Third, group-concerned approaches need to shift attention from competition to collaboration. The division of labor in groups can, certainly, involve elements of competition between group members. But as my case comparative study illustrates, and as other literature has shown (e.g., Poulsen, 2001), research groups divide labor among their members by way of collaboration—and it is their close collaboration, not competition, that makes research groups a powerful instrument of scientific knowledge creation. Through collaboration, research groups effectively combine the resources that members can bring to the group, and the risk of scientific failure is borne collectively. Like scientific communities, groups need to spread the risk of failure, as well as the rewards of success, onto many shoulders. In research groups, collaborating individual scientists typically engage themselves in several projects or experiments at once, and they co-author many more publications than they would be able to publish alone. What group collaboration, thus, takes into account is not only the collective need to distribute epistemic risks, but the individual need to do so as well.
Lastly, group-focused approaches to the division of labor concern knowledge in-the-making, that is, knowledge claims that have not yet passed formal peer review and cannot be taken for “established" knowledge. Such knowledge in-the-making is difficult to rely upon. Therefore, epistemic dependence emerges as a pressing issue in group collaboration, an issue that community-concerned approaches, such as Kitcher’s, typically background. In Kitcher’s account, competing individual scientists pursue alternative lines of research independently. Only when one line of research leads to evident scientific success (or failure) do scientists either disseminate their findings or abandon their line of work and refer to others’ research. Epistemic dependence is not a predominant issue here because the scientific community at large has the means, the peer- review mechanisms and the critical mass to decide what counts as evident scientific success, as “established" trustworthy knowledge, and what does not. Research groups, however, do not possess these means; and since the division of labor in groups is collaborative rather than competitive, members have to rely on one another early on in the process of scientific knowledge creation—an epistemic challenge upon which the following chapters will shed more light.