An Ethical Argument for Philosophy Co-Authorship; on Friendship and Disagreement

This piece was co-written and co-published with Eric Schliesser.

The most dazzling example of co-authorship is Paul Erdős, who co-wrote more than 1400 papers in mathematics with 485 collaborators. (What is your Erdős number?) To do this, he became functionally homeless: “His modus operandi was to show up on the doorstep of a fellow mathematician, declare, “My brain is open,” work with his host for a day or two, until he was bored or his host was run down, and then move on to another home.”

In the sciences, co-authorship is normal. In the humanities, it is uncommon. In philosophy, it is almost non-existent. (See chart.)[1] Yet philosophy is not without famous co-authors (e.g., Marx and Engels). What’s more, some monographs ought properly be considered co-authored, like John Stuart Mill’s collaborations with Harriet Taylor Mill: “when two persons have their thoughts and speculations completely in common it is of little consequence, in respect of the question of originality, which of them holds the pen.”

There is little scholarship on co-authorship in philosophy, but a steady trickle of blog posts suggests that there is interest and anxiety on this point. (Helen de Cruz, Mark Zelcer, Robert Paul Wolff)

1.      What is co-authorship generally?

Co-authorship in the sciences is ideally ruled by two rules rooted in a particular sociology of labs and research groups. First, co-authors contribute to the scholarly endeavor for a piece of publishable scholarship by planning, executing, or analyzing the results of some sort of research. Second, co-authors compose the written portion of the research, either collectively or through some division of labor.

Both of these ideals are violated, of course—there are plenty of massive multi-authored articles where scholars receive token authorship (gift authorship) or someone who made substantial contributions is not credited as an author (ghost authorship.) In that way, these are “endorsed” norms, not the “enforced” ones: violations abound and are even legitimated as common practice in some “big science” research areas. But this remains the practical ideal.

2.      Why is co-authorship deprecated in philosophy?

Professional philosophers collaborate, usually through disputatious conversation, but usually not in a way that counts multiple thinkers as the author of a single paper. We are also much less likely to cite our peers than the agenda-setting papers in our sub-fields, especially as a part of a generic literature review. (See Kieran Healey’s data.) It is more common for close collaborators to co-edit than to co-author: for instance, Amartya Sen and Martha Nussbaum famously created the “capabilities approach” together without co-authoring any of their foundational papers. They did, however, co-edit the seminal volume, The Quality of Life in 1993.

So while we do not imagine ourselves to be lone geniuses communing with the ancient canon, it is the case that many philosophical papers are projected as manifestos of a single principled thinker. Indeed, philosophers are particularly prone to a kind of “subauthorial collaboration” that is formalized in lavish acknowledgements in the text or an early footnote. (Cronin, Shaw, La Barre, 2003)

David Lewis, who understood philosophical politics as well as anybody (recall),  comments on this kind of “ghost” authorship in his paper, “Causation as Influence:”:

“This paper mostly presents the latest lessons I’ve learned from my students. Under the customs of the natural sciences, it should have been a joint paper, the coauthors being (in alphabetical order) John Collins, Ned Hall, myself [David Lewis], L.A. Paul, and Jonathan Schaffer. But under the customs of philosophy, a paper is expected to be not only a report of discoveries but a manifesto; and, happily, the five of us have by no means agreed upon a common party line.”[2]

Yet analytic philosophers often imagine themselves by analogy to the natural and social sciences, and American and continental philosophers increasingly emphasize the collaborative nature of our enterprise. Why, then, do we not use co-authorship more?

One obvious reason is the relative paucity of grants and research support compared to the natural and social sciences. Without an incentive to adopt the lab or research group model, most working philosophers are not members of a funded research unit with reporting requirements. Rigorous research is understood to support ‘research-led teaching,’ and so the single-author model is based primarily on the single-teacher model. We should thus expect the introduction of larger private grant-making institutions, like the Templeton Foundation and the Berggruen Institute, and public European grant agencies, to usher in an era of increasing research group size in philosophy, and the more frequent co-authorship.

We expect that  new collaborative technologies will decrease the costs and difficulties of collaboration to the extent that philosophers will more often overcome them, even without any increased funding or benefits. Google Docs, Dropbox Paper, and simple Microsoft Office sharing, versioning, and commenting will all tend to ease the speed with which scholars of all varieties—including philosophers—create collaborative documents. Social media, chat programs, and email ease collaboration over distance to replicate the proximity by which labs and research groups co-compose.

Lastly, we might think of philosophical research as long-shot bets (Langhe and Schliesser, 2017) that explore concepts, methods, or techniques that may be generally applicable. As such, it will be difficult to trace any particular result in the world back to a paper or monograph, and it is preferable to have many such long-shot bets rather than focus the efforts and attention of a highly skilled research team on sequential elaboration of just one of these gambles. It will generally be better to encourage idiosyncratic work done by individuals and small teams in hopes that one of those bets pays off. In the age of increasingly easy collaboration some enduring, virtual small research teams will emerge without, hopefully, fully displacing lone wolfs and disputatious research groups.

3.     What norms of co-authorship might philosophers embrace?

If philosophers do adopt co-authorship norms, which ones should we adopt?

We should start with the ideal of recognition of actual contribution to the research and knowing participation in the composition. Co-authorship should never be merely “honorary” inclusion for a member of a team or department who contributed to neither the research work nor the writing endeavor. However, the division of labor in some disciplines has the result of producing co-authored pieces where some credited contributors don’t even understand the paper they have helped to produce. Should we extend co-authorship to a student assistant who writes the literature review for a piece of research without understanding the rest of the paper? Should we accept co-authorship for a paper whose conclusions we do not endorse, if we supplied the formal modeling or computer code that underwrites it?

Our current professional ideal is consensus. On this ideal all authors must give credence to every jot and tittle of the work, even if their confidence varies slightly. Co-authorship in philosophy is stricter because philosophers can and should endorse all the claims, arguments, and conclusions of the paper. Papers can be thus be co-authored in philosophy when two or more researchers find a common interest, discuss it at great length, and truly co-compose the entire paper, contesting each argumentative move and turn of phrase until agreement is reached. As such, it should be very unlikely to see large groups of co-authors writing together, given the difficulty of producing such an exacting meeting of the minds, and we even look askance at philosophy papers co-authored by three or four scholars.

Another possibility is that philosophers working on closely-related research might try to divvy up the tasks in a field of research around some question such that only one or a few of the authors of a work understand the whole thing, while others are credited for their contributions without being “first” or “lead” author status. This requires trust in each other to handle sections of the paper that address relevant issues from their sub-fields; perhaps a paper that is 12,000 words long is written in two halves, with only the introduction and conclusion truly a joint project. On this ideal, co-authors retain to right to veto truly abhorrent claims made outside of their assigned sections, but only by threatening to dissolve the partnership. Otherwise, they can only register objections and hope to be heard.

A third possibility would be to follow the norm in law courts, where empaneled and en banc judges issue both dissenting opinions or in some cases join the majority in parts of their decision but not others, perhaps affirming the result but not all of the methods used to arrive at it. In such cases, a minority might dissent vehemently, while the court is understood to have rendered the opposite verdict: that minority dissent is clearly a separate research project in philosophical terms. They are not understood as co-authors of the majority opinion but as co-authors of a distinct opinion which did not win a majority of support. The more instructive question is what to do when judges affirm parts of the main decision but not all of it, and by analogy, situations where co-authors affirm parts of a research project but not all of it. Can it be possible for one of the co-authors to “sign on” to parts 1, 2, and 4 of a paper, while dissenting from arguments found in part 3 and in the conclusion? For example, in the past one of us has used a footnote to signal an alternative position from the one arrived at in the body of the paper.

A fourth possibility is majority voting. Bright, Dang, and Heesen (2017) argue that scientific work should aggregate researchers’ judgments. Claims and propositions should be made in a paper that receive the assent of the majority of the authors. One might well find oneself outvoted in some cases, and this would be fine so long as there was agreement that voting had not produced a contradictory or logically incoherent set of claims.

A fifth possibility would be a kind of deliberative dictatorship: a lead author could write a paper, assign sections for others to compose, and bounce ideas off of possible co-authors. At the conclusion, all participants who agreed with the final product could sign on as co-authors, while dissenters could produce their own dissenting papers to be published alongside. This is related to proposals that might allow peer reviewers to receive more recognition for their work as initial gatekeepers.

Could all of these modes of co-authorship flourish in philosophy? Are some of them inimical to the discipline?

4.      Philosophers should co-author more of our work

Given the fact that some forms of collaborative recognition do exist, why suggest co-authorship as an alternative model for philosophy? Let’s start by dividing the reasons for co-authorship into roughly epistemic and roughly ethical categories, even if this is a division that is easily collapsed. Co-authored papers may simply be better for having multiple composers, readers, and researchers attached to them. The division of epistemic labor will often lead to better-written, more carefully crafted, or simply more copious publications: many minds make light work.

At the same time, co-authorship is partly about recognizing the contributions of our peers. In that sense, it is ethical. This is a weaker defense of co-authorship, since there are alternative methods for providing recognition. If a paper issues from a conversation with a colleague or a good objection raised at a conference or blog post, we philosophers would normally expect to mention that in a footnote, not to grant the colleague or objector co-authorship status. Philosophy papers are sometimes imagined to be the record of the thoughts or analysis of a single agent, and group agency seems much more difficult in these cases because we are so rarely in anything resembling agreement. We are rarely of a single mind, ourselves, so this is no big impediment, but this also ignores the fact that one can commit to a written product while having varying confidence in its disparate elements.

But there’s a significant ethical claim that might recommend co-authorship: the ideal of scholarly friendship. Co-authorship can be a way to channel professional philosophical relationships in productive ways, a norm for guiding conversations and arguments towards shared, potentially overlapping projects. There is independent reason to believe that shared projects are an intrinsic good tied closely to well-being. (Korsgaard, 1992) Thus we should, if possible, prefer to share the tasks associated with philosophical research with others, not just after publication but throughout the scholarly endeavor. Philosophical co-authorship is desirable just because philosophical friendship is desirable.

Shared projects are possible both between equals and between mentors and students. As such, co-authorship is a way to encourage productive collaboration within departments and with undergraduate and graduate students. In the pedagogy-first model of much philosophical research, departmental colleagues at most small schools should not co-author their research because this leads to overlapping areas of interest and knowledge. A department with only a handful of philosophers should instead hope that its faculty have as little in common as possible, even if they must share governance of their department and spend their careers working side-by-side. But if co-authorship underwrites philosophical friendship, then even a maximally pluralistic department should seek opportunities to co-create research, actively seeking agreement and shared methods, research areas, and conceptual terrain.

Like friends, co-authors need not agree on everything.[3] Finding some method for adjudicating those disagreements is important, but philosophical writing can encompass these minor dissents or majoritarian procedures, just as our departments do. The key is that the commitment to co-author—like the commitments of friendship—is a commitment to resolve disagreements using whatever methods are available. Friends do not obsess over decision-procedures, though we adopt them to ease tensions for the sake of shared projects. The same should go for co-authors.

Like friends, co-authors need not be equals. We see in the sciences that co-authorship allows a kind of scholarly mentorship, and in philosophy graduate students experience intense collaboration for the first time while writing their dissertation with a senior scholar. We even acknowledge that this is akin to co-authorship by treating dissertation advising as something close to co-authorship for some professional purposes.

More of this sort of mentorship should be encouraged. The practice of learning from another scholar does not end when a philosopher receives a PhD, and probably we shouldn’t pretend that it does. Perhaps newly-minted PhDs aren’t yet ready for the full burdens of a research program, or perhaps they would benefit from mentorship when they move on to a new research program. Or perhaps not: perhaps this would end with more domination by senior scholars, as the division of labor creates permanent hierarchies. But it’s not as if our current, academic political economy is hierarchy-free.

A final reason is merely accuracy: our authorial norms give a false idea of our practices to the rest of the academy. We should consider revising them to align ourselves with our fellow academics. Let’s not pretend that scholarly productivity metrics are irrelevant or that Deans do not look askance at our publication records compared to other disciplines.

[1] Despite the best efforts of some of the authors of this post!

[2] The quote and suggestion comes from Christopher Hitchcock, who offered it as a comment to an older post on philosophy coauthorship at New APPS.

[3] We are indebted to Andrew Corsa’s and Eric Schliesser’s unpublished research on Margaret Fuller’s ideas on friendship and magnanimity.

Apologies to Eric Schliesser

In my last post, I noted that Jason Brennan’s published work strongly opposed disenfranchisement in the ordinary sense, and I claimed that Eric Schiesser had misrepresented his words in order to derive that conclusion. Today, Eric Schliesser supplied an unpublished paper in which Brennan offers an argument for experimentation with competency tests to disenfranchise incompetent voters.

Eric, please accept my apology.

Here are some telling highlights from Brennan’s paper:

  • In this paper, I argue that the practice of unrestricted, universal suffrage is unjust.  Citizens have a right that any political power held over them should be exercised by competent people in a competent way.  In realistic circumstances, universal suffrage violates this right.  Since nearly all current democracies have universal suffrage, all current democracies are to that extent unjust.
  • Restricted suffrage is about as unjust as voting age laws.  It creates a ruling relationship between different classes of citizens based on a distinction that all reasonable people can accept in the abstract, but about which in practice there will be reasonable disagreement.  In contrast, universal suffrage is about as unjust as a policy of enforcing jury decisions not matter what, even when we have conclusive grounds for thinking the jurors were incompetent or made their decisions incompetently.  Thus, universal suffrage appears to be more intrinsically unjust than restricted suffrage.
  • We do not know for sure whether voter examination systems would produce better or worse results than democracy universal suffrage.  However, as I have argued, such systems are less intrinsically unjust than democracies with universal suffrage.  And there are good reasons to think they will produce better results than democracy with universal suffrage, though there are reasons to worry they will not.  Since we are unsure of the consequences, but have reason to expect them to be positive, we might experiment with voter examination systems on a relatively small scale at first.  For instance, perhaps it would be best if one state in the U.S. tried the system first.  (We would want to start with a relatively non-corrupt state, such as New Hampshire, rather than a corrupt state, such as Rhode Island.)  If the experiment succeeds, then the rules could be scaled up.  Similarly, consider that a few hundred years ago, we have little experience with democracy.  Some advocated democracy in part because they believed it would tend to produce better and more just outcomes than monarchy.  Others worried that democracies would be even more corrupt, or would collapse into chaos.  In light of their lack of experience, a democrat might reasonably have argued in favor of experimenting with democracy on a relatively small scale, and then scaling up if the experiment succeeds.

Unlike a call for abstention from voting, a call for restricted suffrage is certainly support for disenfranchisement. I intend to respond to this paper in depth when it is published, but for now I will say that Brennan ought not to look to Burkean conservatism for practical objections to such experiments, but rather to Hayekian liberalism or Arendtian republicanism (cf. Brennan’s comments on civic virtue.) As always, the difference between micro and macro, small experiments and institutional redesigns, should not be overlooked. (No doubt Brennan is exploring territory similar to the prediction markets I discussed with Robin Hanson last year here and here.)

Must we destroy the profession in order to save it?

Jason Brennan, The Ethics of Voting, 2011, page 5:

“The right to vote and the rightness of voting are different things. I do not argue that we should disenfranchise anyone. Though I think many voters are wrong to vote, I will not argue that anyone should prevent them from voting.” (Emphasis mine)

Eric Schliesser, New APPS, 1/3/12:

“The following tentative remarks were caused by reflection on the recent publication of books that (with qualification, of course) condone… disenfranchisement of ignorant voters.”

Jason Brennan, interviewed for a web article posted to The Daily Need, 4/15/2011:

“Since writing ‘The Ethics of Voting,’ I’ve actually become more sympathetic to the idea that maybe people should be formally excluded from voting,” Brennan said.

Of course, there are obvious dangers implicit in this view, as Brennan admits. Special interests, for example, might co-opt the voting process to exclude those who won’t support their agenda. Incumbents might bar voters who are likely to oust them from office. And literacy and comprehension tests have an ugly history dating back to the Jim Crow era, when they were used to disenfranchise African-Americans. That led to the passage of the Voting Rights Act of 1965. (Emphasis mine; the second paragraph is from The Daily Need article, but it is partly a gloss of Brennan’s book, page 108)

Eric Schliesser, New APPS, 1/4/12:

“When I wrote my original post, I had no idea that Brennan had already slipped down the slope of my not so ‘idle concern.’ As a community, we are politely witnessing (refereeing, reviewing, etc) the philosophic ground-clearing being done for –and the accompanying public marketing of — ideas that will justify a certain kind of elite rule. […] (In the book he argues for a sub-set [of disenfranchisement], self-disenfranchisement.)” (Emphasis mine)

In his post, Schliesser argues that we ought to consider a professional code of conduct that bars sympathetic consideration of topics like not-voting or torture when that research threatens to supply “a rhetorical fig-leaf to let politicians and generals morally off-the-hook for atrocious deeds,” or research that assists the state in “annihilat[ing] enormous number[s] of innocent people deemed enemy by the government.”

Unfortunately, as stated, the proposition is self-refuting. Here’s why: Schliesser suggests we ought to consider restrictions on easily misunderstood work on moral taboos. His own post could easily be interpreted as a call for a ban on politically unpopular research. In fact, my friend Leigh Johnson interprets his question in this way: as an encouragement to ban “Forbidden Knowledge.” In the comments, Schliesser claims that this is a misreading, but Johnson is both a charitable and smart reader: if it’s possible for her to derive this interpretation of his words, it’s quite easy for them to be misused in this way by “politicians and generals” in search of a “rhetorical fig-leaf.” If Brennan’s book with its clear warning against disenfranchisement can be read as a call for disenfranchisement, then certainly Schliesser’s post can be read as a call for a ban!

In this sense, Schliesser’s post is precisely the sort of thing that would violate a code of professional ethics like the one he’s describing! The counterfactual is not: “What would the world look like if Brennan had not written his book and articles?” The relevant counterfactual is: “What would the world look like if there was a code of ethics that restricted research on moral taboos?” Such a world is more likely to involve illiberal restrictions than the ones that Schliesser seems to favor.

But I want to focus on the justifications he gives between two different kinds of restrictions: informal norms and codes of conduct. As he points out, there already are norms of collegiality, politeness, and prudence that govern our professional interactions, norms that Schliesser violates by reviewing and now attacking a book he has not read and using a likely-sensationalized web interview to justify his position after the fact.

What is added when these informal norms are codified? The clarity and transparency of codified norms at first seems promising, but wherever there is a code of conduct, there must also be a process for amending it. Since Schliesser advocates against elite rule, I assume he would want this process to be open to all members of the profession, as would I. The alternative is elite rule by the professional ethicists!

But what makes Schliesser think that our profession would adopt rules that outlawed research on torture or epistemic problems with democracy, given the fact that such research is popular and widely believed to be central to the profession’s mission? Here it’s important to note that while many professions have codes of conduct, there is little evidence that they actually bind members of the profession meaningfully because they are a isonomic: the professionals themselves make the rules. Finance professionals, lawyers, engineers, and research scientists are all bound by rules of conduct that are so laughably generic that these same groups are regularly caricatured as unprincipled and even evil, both by the public and by disillusioned members of the same professions.

Another reason to codify norms is because codification supplies an opportunity for an intervention: though informal norms will govern what sorts of rules are considered legitimate, it’s also possible that new rules will eventually be internalized as informal norms. This is obviously what Schliesser hopes: that by sneaking a politically radical provision into the rules banning research on moral taboos, we will finally internalize the informal norms that actually prevent such work. But this is certainly not a democratic hope: it is governance through trickery by a different set of elites.

Now, Schliesser is certainly right that the philosophy profession is driven by the demand to scrutinize all assumptions and follow arguments where they lead, and that some professional philosophers are so enamored of this ethos that they will adopt contrary positions and embrace iconoclasm over prudence.

However, these drives seem less dangerous than Schliesser supposes, especially because these drives are already limited by the existence of informal norms governing the profession, which dictate who is considered fit for jobs, tenure, and grants. I think it is likely that efforts to codify and enact punishments to fit these informal norms are more likely to harm than help, more likely to transform into a broad and toothless set of guidelines or into Johnson’s feared ban on “forbidden knowledge” than to successfully institute new moral taboos.

Perhaps, in part, research like that of Brennan and Alhoff is not justificatory but rather inspired by the fact that we already inhabit a world governed by “a certain kind of elite rule” and one where torture is condoned by the highest political authorities. Jason Brennan’s voting ethics is looking pretty good the day after the 2012 Iowa caucuses with democracy reduced to an adolescent joke: “Obama, Romney win Iowa caucuses; Santorum slips into #2 spot.” Someone has to formulate that argument, if only so that we can discover a winning refutation or design institutions that do not so easily suggest the argument’s conclusion to onlookers.

If we want to do research that combats the status quo of elite rule and torture, we must understand the arguments that currently justify it, or else fall victim to overuse of “Weak Man” argumentation and research that is irrelevant to contemporary needs. To ban discussion of the facts of our common existence is an absurd self-destruction of the profession, especially its most emancipatory traditions, in an attempt to save it.