What if coauthors disagree about what their ms should say? (UPDATED)

In a recent interview, Richard Lewontin talks about how he and the late Stephen Jay Gould came to write their famous polemic “The Spandrels of San Marco” (ht Small Pond Science). Basically, Lewontin says all the polemical bits were by Gould, and that he only wrote one non-polemical section. And he says Gould went too far in the polemical bits, taking unreasonably extreme positions. A few quotes from Lewontin, to give you the flavor:

Steve and I taught evolution together for years and in a sense we struggled in class constantly because Steve, in my view, was preoccupied with the desire to be considered a very original and great evolutionary theorist. So he would exaggerate and even caricature certain features, which are true but not the way you want to present them…He would fasten on a particular interesting aspect of the evolutionary process and then make it into a kind of rigid, almost vacuous rule…

Most of the Spandrels paper was written by Steve. There is a section in there, which one can easily pick out, where I discuss the various factors and forces of evolution…

This surprises me. Not for the gossip about Gould’s motivations–I’m not much interested in that–but because Lewontin is more or less admitting that he put his name on a paper that he didn’t entirely agree with. Which surprises me because my attitude is very different. I don’t let a paper go out with my name on it unless I agree with every word of it. I figure I’m an author of the whole paper, not just “my” bits of it.

To be clear, my concern here isn’t with the technical soundness of my coauthors’ work (which in some cases I couldn’t actually check even if I wanted to), or with different people writing different bits of an ms. It’s with whether my coauthors and I all agree on the interpretation and implications of our work, and what to do if we don’t.

I’ve been involved in collaborations in which we disagreed about interpretation, sometimes very seriously. But in the end every collaboration with which I’ve been involved has managed to write a paper everyone was happy with.

There are degrees of agreement and disagreement, of course. I’ve had collaborative papers that would’ve been slightly different if I’d been the sole author–there’d have been differences in emphasis, or some points would’ve been phrased differently. Perhaps that’s what’s going on in the case of “Spandrels”. Maybe Lewontin would’ve preferred different phrasing or more (i.e. any!) nuance, but he basically agreed with Gould’s main points so was happy to put his name on the paper.

One way to resolve disagreement among coauthors would be for them to lay out their disagreements in the ms. One occasionally sees papers like this, but only from “adversarial” collaborations between intellectual opponents. There’s no reason in principle why friendly collaborators who only partially disagree couldn’t do the same thing, but I’ve never seen it done. (UPDATE: They say the memory is the first thing to go. Andy Gonzalez comments to remind me that he and Andrew Hendry have a friendly disagreement about the prevalence of local adaptation. They wrote a dialectical paper about it. And see the comments for other examples of adversarial collaborations in which intellectual opponents wrote joint papers clarifying their areas of agreement and disagreement.)

The various meanings of “authorship”, and different standards for authorship, are relevant here (see this old post). If you think of an “author” just as “someone who made a substantial contribution to the work reported in the ms”, then maybe you don’t assume that every author necessarily agrees, or should agree, with everything in the ms. The author list is just a list of people who contributed in various ways to producing various bits of the ms. Not a list of people who agree with everything the ms says.

I’m guessing this is an issue on which folks have very different experiences and views. So here’s a little poll. Do you think coauthors should agree on everything their ms says?