On scientific co-authorship (1): Why do scientists publish together?

The allocation of merit to individual scientists is one of the crucial aspects of how scientific systems work. Publication of ‘papers’ in important journals, and, still most significantly, citation of those papers in the works of colleagues, is perhaps (with all its shortcomings) the most determinant mechanism for recognising the value and capacity of each scientist, and hence the amount in which he or she deserves a share on the different types of resources that scientific institutions manage to collect.

However, multiple co-authorship is an extended and growing phenomenon within most of the academic disciplines (though much more in the natural sciences and technology than in the humanities, for example), in spite of clearly contributing to making less transparent the allocation of merit to individual people. Some evaluation processes take this into account by just valuing less co-authored papers than single-authored papers, or by applying some type of discount depending on the number of authors, or on the order a person occupies within the list of authors of a paper.

But one might wonder why in science writing could not be established some mechanisms like ‘movie credits’: why not just to include at the end of each co-author paper an indication of what has been the contribution by whom? Even it could be done quantitatively instead of qualitatively: the authors might agree on a percentage distribution of the ‘credit’, for example. Notice that I am not talking here about collaboration, but about authorship. The research process might be carried out with all the collaboration we might want, but after that, each author might identify a piece of the final work as the one on which he or she gets an ‘intellectual right’, or the co-authors could divide the ‘value’ of those rights according to some formula. Of course, there are multiple circumstances in which the former might be difficult or impossible (though this is not the case in the latter case), but even in cases in which it is rather clear what the contribution of each co-author has been, it seems that most often scientists prefer to share the whole authorship of a paper instead of letting other people know about the individual contributions.

Let’s examine some of the reasons that a group of researchers may have to present the conclusions of a piece of research as a collective claim, instead of as an aggregate of individual but separate contributions (for a general account of co-authorship and collaboration in science, see Wray, 2006). Of course, they might just be lying, in the sense that one or more of the co-authors have not contributed in any relevant way to the paper’s content, but are included for other reasons: exchange of favours, academic promotion of a friend, or, more frequently, sheer abuse of power. These types of co-authorship are rightly condemned by all scientific codes of conduct, but their commission is easily explained when the probability of the misconduct being undetected is high. A particular form of cheating would consist in a mere exchange of papers: you write one paper by your own, I do the same with another paper, but we signed both papers as if they had been the result of collaborative research between us. This is obviously another example of scientific misbehaviour, but I doubt it is as frequent as the former cases, not because it is more likely to be detected, but because it is probably not as profitable for authors at it might seem at first sight. If the merit for an individual from co-authoring two papers among two people is equal to the merit from single-authoring one paper, then you don’t gain anything from cheating in this way; and if the merit of the former is less than that of the later, or if we take into account the chances of being discovered and punished, then this strategy becomes clearly inefficient. It would only be ‘rational’ if getting two co-authored papers gave substantially more recognition to an individual than getting one single-author paper, and so, the fact that this kind of behaviour is not too common seems to point to the conclusion that scientific merit decreases more than linearly with the number of co-authors.

Legitimate co-authorship corresponds, obviously, to the cases in which there is a real collaboration between the co-authors. But we can divide this possibility into two different cases. First, a group of collaborating scientists may write a single collective paper simply because there is no clear enough way of separating the contributions of each researcher to the argument supporting the paper’s conclusions. Since ideas are discussed amongst all the members of the group, continuously examining them and proposing new ones on the face of the arising problems, it can certainly be the case that all the elements of the paper contain a significant contribution from every individual member of the team. After all, a single co-authored paper that in its final version is the paradigm of clarity and precision can be the result of a messy process of trial and error and of dozens of muddy drafts, a process the author is not able of reconstructing after the work is done. I assume that many co-authored papers fall into this category, in particular, those that are written by teams (usually, couples) that work together for many years and that produce a large amount of collective research (one can think, for example, in the psychologists Amos Tversky and Daniel Kahneman). Co-authorship in such cases is really unavoidable, and there would be nothing left to explain about its rationality.

Second, there is the possibility that a group of researchers collaborate and write a coherent article, but that in principle they could factor out the contributions made by each member. This is particularly the case when collaboration is mainly justified by the division of labour amongst people with different competences or from different areas, but also amongst colleagues in the same field that simply concentrate on a different part of the paper though they discuss together the general idea of the work at the beginning, and its final versions at the end of the process. What justifies in this case that the paper is signed by several co-authors, instead of each one reclaiming merit only for the part to which she has mostly contributed? After all, what some of them are doing is using the conclusions reached by the others as premises in a (shorter) argument. Assuming that the goal of the authors of those intermediate conclusions is to get cited, why not write a separate paper for each of those conclusions, one that will be cited by the colleagues doing the rest of the argument? Stated in terms of the economics of the firm: why are not the fragments of the argument ‘externalised’ as far as possible, giving to each of those fragments the possibility of being autonomously recognised for them? This case, which I will call ‘optional co-authorship’, is the one that I shall try to explain in the next entries.

Reference:

Wray, K. B. (2006), “Scientific authorship in the age of collaborative research”, Studies in History and Philosophy of Science, 37, 505-514.