Structuring a Paper

Here’s a whine about a growing, relatively new practice I see in papers I referee. The structure of the paper involves a nice intro, a quick presentation of the view and argument for it, followed by a section titled “Objections” or “Objections and Replies”. There may be occasions on which this kind of structure is appropriate, but most of the time it isn’t, I think.

The reason is this. It makes the last section of the paper look like a compendium of what I’ll call “conversations with Alexes” (named after my valued colleague, Alex Pruss, who always has interesting things to say about any topic in philosophy). As such, it has no internal structure, merely bouncing from one issue to another, leaving the reader in despair for wont of a natural stopping point.

I prefer the more obvious strategy: state a thesis and then provide an argument for it that includes nothing not necessary to defend the thesis and everything necessary, really necessary, to show that it is true. If the pre- “Objections” parts do that, it’s not clear why we need the “Objections” section. After all, there is nothing one can do to block the possibility of an objection, so I prefer a strategy that focuses on what papers really need to do instead of on things one would like to say to Alex.

OK, whine over.


Structuring a Paper — 16 Comments

  1. I don’t like the structure much, but I confess I use it pretty often. It’s partly a response to people doing a bad job refereeing papers. Many referees don’t know the difference between an error and an objection, and they reject papers on the grounds that they can think up an objection to it. Sometimes, the objection itself is either irrelevant or involves a misunderstanding of the paper’s argument, and sometimes it doesn’t.

    Either way, this puts pressure on authors to brainstorm all the lines of objection they might be able to think of, and it include answers to them in the paper. They hope that by doing so, and especially by including whatever objections the referees raised at the last journal where the paper was submitted, they’ll avoid having the paper rejected on the grounds of an objection raised by a referee.

  2. I feel your pain, Jon. In some cases, though, that sort of structure may be excusable to some extent. Many referees tend to raise objections that are, to put it nicely, not particularly central to the paper’s main theses and arguments. To get a paper past such referees, we often need to anticipate their not-so-central objections. The proper place for discussing those sorts of objections is often in a second part of the paper, a part that comes after the main arguments. And so we put those objections and replies there, not because we think of them as necessary parts of our arguments, but because we need to placate referees who will raise those objections anyway.

  3. Following up on the above comments, it would be nice if we could attach those sections of the papers as an addendum and if the refs agree that there’s nothing to these objections the section can be dropped.

  4. I’m not sure anticipating dumb objections is, by itself, a good justification for a reply to objections section–especially, if the section considers a number of completely unrelated objections.

    As a rule of thumb, I don’t address referee objections that I consider dumb, irrelevant, or not quite important enough. Following this rule of thumb helps prevent clutter. One exception is when I have heard two or more people giving the same sort of objection. Another possible exception would be if the referee gave an R&R and said that the paper had to consider the objection to be publishable. But, in either of these exceptional cases, I simply address the objection when it arises in the text, often with a footnote.

    I’m sure there are cases in which a separate objection section is needed, but I’m guessing that this isn’t particularly often.

  5. Chase, I sorta expect that most of my papers will be rejected for stupid reasons at least once. I think it is better to live with this reality than to try to prevent it by adding a section to deal with mostly unimportant objections. Don’t make the paper worse by trying to appease bad referees. Make the paper the best you can so that you can appease good referees.

  6. Sounds reasonable to me — sounds like someone has just internalized Ralph Johnson’s definitions of “manifest rationality.”

    “Johnson wants to invoke “manifest rationality” as a criterion for argumentative products at two different levels. At the illative level, rationality is manifest in the structure of reasons and conclusion, while at the dialectical level, rationality is manifest as a response to objections and alternatives. These are two different standards of rationality that correspond to distinct, but equally important, aspects of argumentation.”

    See Johnson’s book by that name.

    David Beard, UM Duluth.

  7. I agree with Chris. There is a real danger of letting the worst reviewers determine what most papers in a discipline look like. If left unchecked, this tendency (Jon says the practice is growing) might turn the literature into a series of unimportant objections to arguments that are intended only to occasion them. Look back through the history of many disciplines and I suspect you find, perhaps not whole disciplines, but sub-specialities, degenerating into precisely this sort of thing. It is a misguided attempt to “keep the conversation going” by appeasing (i.e., by seeming interested in the arguments of) the most mediocre contributors.

  8. Hey Chris,
    You wrote
    “I think it is better to live with this reality than to try to prevent it by adding a section to deal with mostly unimportant objections.”

    Better how? It’s clear that _the paper_ would be better, but that’s just one consideration in trying to figure out what’s best to do. If you think about the time it takes to get a piece of work through one round of reviews (if a paper gets rejected, you’re lucky if you can get a second opinion on that paper within 6-9 months which means that it’s unlikely that you’ll get two opinions on any paper within one job hunt cycle) and the tendency of many referees to try to find _some_ reason to reject any paper they see, I think it’s irrational for someone who needs publications for tenure, for landing a job, etc… to not try to preempt bad objections that could prevent a good paper from being accepted by a year or more. (That’s why I like the idea of attaching the replies to objections (bad objections) as an addendum that can be dropped if no one thinks it needs to be kept in.)

  9. Hey Clayton,

    First, I like your addendum idea to some degree, provided that editors and referees are willing to consider papers with an optional addendum or a separate document altogether. My remarks were directed primarily against this idea: the extra section should be added to satisfy bad refs, where the extra section is treated as a section not as an addendum. I’ll continue to respond to that idea here.

    I wasn’t ignoring the time pressures; rather, I was operating on this assumption: if you don’t add the extra section, you will irritate bad refs AND if you DO add the section, you will irritate the good refs. Suppose, for the sake of argument, that Kvanvig is a good ref. It seems that he is less likely to accept a paper if it has the awkward, compendium-like, respond-to-dumb-objections section.

    Assume that the likelihood of getting a good ref rather than a bad one is .5. Then adding the awkward section is just as likely to irritate your refs as not. Given this assumption, then, it isn’t clear that adding the awkward section helps you get stuff published. So the time pressures seem irrelevant and it will usually be all things considered better to go with leaving the section out.

    Of course, there are two natural responses. First, you could argue that the optional sections won’t irritate good refs as much as they will irritate bad refs not to have it. Second, you could argue that it is more likely that you’ll get a bad ref than a good ref. I think both arguments would be at least somewhat plausible. But I’m not sure either argument (or both taken together) would USUALLY be strong enough to outweigh the general rule: write the best paper you can and hope you get refs that appreciate the goodness of the paper.

    And another thing to keep in mind: even if you decide it is all things considered better to respond to certain dumb objections, it doesn’t follow that you should consider these dumb objections at the end of the paper in a separate section. It could be that they should be considered as they arise in the paper, perhaps in footnotes.

  10. It is also worth pointing out that, in addition to the points Chris mentions, the goal here isn’t publication. It’s to make a contribution, one that won’t be ignored in the way most publications are ignored. Adding the objections section may get one past some referees, but has a cost after publication.

    Of course, many brilliant philosophers write badly, and have great reputations in spite of the writing. So one can always write to get past bad referees and hope for the best!

  11. Jon,

    I agree that the goal isn’t publication, but it’s a goal. Given some of the silly objections that figure in referee reports that stand in the way of publishing good papers, I stand by the claim that it’s unwise not to try to preempt. You’re right that this makes the paper less readable in some ways. That’s why I think you should be allowed (nay, encouraged!) to identify the response to objections section as something you’d be willing to drop if the referees don’t see anything in that section worth publishing.

    Chris is right that the footnotes are a good place to address these objections. Two worries. Some referees don’t read the footnotes (recently someone said I should have credited X for something on pp. n in spite of a footnote on pp. n that credits X for the very point at issue). Some journals don’t like footnotes to contain anything substantive.

  12. Jon,

    Why would addressing objections “cost” you after publication? If the argument is good, why would it detract from the paper’s value for the author to address some objections?

    You say that the paper would “[leave] the reader in despair for wont of a natural stopping point.” For someone who doesn’t like the objection-and-replies section, wouldn’t the natural stopping point be right before that section?

    I confess to preferring your strategy, and to sharing the goal of making genuine contributions. But I also agree with others that bad refereeing really complicates things.

  13. There is an interesting feature starting to appear regularly in online Computer Science conference referee systems which allows authors to write brief replies–150 words for each review. (This is unimaginable without an automated system: ECAI had over 700 submissions this year; each submission received at least 3 reviews, and the time from submission to decision was two months.)

    The latest versions ask authors to categorize their replies by one of three categories: (i) The referee made a technical error; (ii) the referee was too hard on the paper and here’s why; and (iii) the referee violated professional standards and the Program Chair should take a look.

    Some author’s reflexively complain over a negative decision, which ends in a wash, but not all do. Having author feedback, while not perfect, catches whopping errors and it seems to cut down on the content-free hostile reviews which are a real bore.

  14. This is not a comment on the utility or aesthetics of an “objections and responses” section in philosophy papers, but on Jon’s “thesis + one powerful positive argument” model for philosophy papers. If this model were universally adopted, we wouldn’t have Bill Lycan-style arguments (e.g., “I’ve counted seventeen objections to representationalism, and I refute them here”). This would be a loss. Especially if one powerful argument for representationalism were already published elsewhere.

  15. Jared, that’s hasty–nothing I said implies that papers can’t be a compendium of responses to objections current in the literature. I’m not a big fan of such papers, but that’s a different story. My concern is that it is easy to have a misplaced idea that one can anticipate bad objections that might be raised by bad referees, and forestall negative decisions by including an “objections” section. I think a better strategy, and a more powerful paper, results from working harder at teaching what one takes oneself to have learned in thinking hard about some problem. Focusing on the pedagogical task of paper-writing allows one a nice basis for deciding when comments by referees need to be addressed and when they don’t, just as when we teach in class, we recognize that some questions don’t show that our methods need improvement while other questions show that we didn’t do a very good job from a pedagogical point of view. Philosophy papers that try to teach what the author has learned are great fun to read, very fun to referee, and much easier to publish in good places. I highly recommend the practice.

  16. Hi Jon,

    I really share your qualms about this kind of structure. But I didn’t take your original complaint to be one about including all sorts of lame objections in a final “objections” section. I took it to be a structural complaint. Even if the objections are good ones, there is something problematic about the “let’s list them all in a single section and answer them one by one” approach. Anyway, I took this to be the complaint, and it sounds right on to me. Better, I would think, is to save that last section for answering the Single Big Worry that you’ve been saving up for the whole paper (if there is such a worry) and just answer objections to specific steps in the argument when they arise naturally in presenting the argument.

Leave a Reply

Your email address will not be published. Required fields are marked *