Posted by: Jeremy Fox | May 23, 2012

What makes for productive scientific debates?

Science is full of debates. Some are productive, some aren’t. What makes for a productive debate?

First, a few remarks about what I mean by a “productive” debate. I don’t mean a debate that leads to agreement on all or even any points, either among the main participants or among non-participants. For instance, consider a debate on some matter of empirical fact. If round-earthers and flat-earthers debate the shape of the Earth, and eventually agree that it’s flat (or “compromise” and agree the Earth is hemispherical), does that make their debate productive? I’m not saying that one side or the other is always right, or that compromise positions are never right, merely that resolution of any sort is not a marker of a productive debate. Now, it’s always unproductive if participants can’t agree on what questions they’re debating.* But there are lots of reasons why a debate might fail to settle on an agreed resolution, and being “unproductive” is only one of those reasons.

What I mean by a productive debate is a debate in which the participants engage with one another, meaning that they pay close attention to and understand the other side, and respond to the other side’s evidence and arguments rather than strategically ignoring them. A productive debate also is one in which all relevant evidence and argument gets fully aired, and arguments are pursued to their logical conclusions and their full implications considered. A productive debate also is one that doesn’t get sidetracked by misunderstandings. This means that the participants need to choose their words carefully and precisely, be clear and explicit, and expect the same from other participants. A productive debate also is one in which no one engages in personal attacks, and in which no one takes criticism of anyone’s views (no matter how strong) to be a personal attack.** Productive debates may not reach an agreed resolution—but if they don’t, they at least make the issues crystal-clear. That clarification of the issues is both a very useful outcome (especially to students and others learning for the first time about the subject of the debate), and the most that can be expected.

By that standard, there are lots of productive debates in ecology. Recent debates over MaxEnt, aired in part in Oikos, are an excellent example. Debates over “sampling effects” in biodiversity-ecosystem function research, sparked in part by an Oikos paper (Aarssen 1997), led to productive development of new experimental designs and statistical techniques that resolved the issue (Loreau et al. 2001). The big debate over ratio-dependent functional responses eventually led to agreement on some issues and “agreement to disagree” on others (Abrams and Ginzburg 2000).***

Actually, probably most “tit for tat” exchanges of comments in the literature are productive by my standard. After all, that’s what “tit for tat” means—you raise a point, and I respond to it rather than ignoring it or dodging it. Indeed, there’s a productive debate in ecology every time reviewers do a good job reviewing a ms and the authors respond. Of course, in such cases there’s an editor involved, who effectively can force both sides to debate productively, on pain of having their comment go unpublished, their review ignored, or their ms rejected. 😉

Notably, in the examples I listed where the protagonists eventually came to partial or complete agreement, this was only after a lengthy period of vociferous disagreement. If you are the sort of person who wants to see agreement at the end of a debate (like my fellow editor Dustin), well, I think you can only get that, if at all, by first letting the debate run its natural, debate-y course.

Which isn’t to say all debates in ecology and evolution are productive. The recent spat over inclusive fitness in evolutionary biology hasn’t, as far as I can tell, raised any issues that weren’t already familiar, and seems if anything to have muddied the waters rather than clarifying them. Further back, while I don’t know the punctuated equilibrium literature well, my impression is that that debate had both productive and unproductive elements. Productive in that it sparked interest in some important issues and prompted some new empirical research. Unproductive in that it proved difficult for the protagonists to agree on the questions at issue, and on what would count as an answer. IIRC, the punctuationists were rather shifty and difficult to pin down on what exactly they were claiming, and the same data were infamously interpreted as evidence for and against punctuated equilibrium by the opposing sides.

Anyway, that’s what I would argue. Who wants to debate me? 😉

*It’s fine, and often necessary, to debate the choice and framing of the question, and what counts as evidence for a given answer. But at some point, all sides do need to come to an agreement as to what questions are at issue and what would count as an answer if debate is to be productive.

**Which means, among other things, that “politeness” can be the enemy of productive debate, if by “politeness” you mean “not saying exactly what you mean, in order to avoid possibly offending someone.” That’s not what I mean by “politeness”. In my view, empirical evidence and logical argument are never impolite, and if anyone else takes them that way, that’s their problem. This doesn’t mean you can’t try to phrase what you have to say in a polite way, but you can’t do so at the expense of clarity and explicitness if you really want to have a productive debate.

***Note to youngsters: yes, there once was a massive debate in ecology over a particular functional response model.


  1. Are you assuming that there’s a different answer to your question if it’s a scientific debate vs any other type of debate? Because I don’t see any difference and would argue that the most important criterion is that one has to be personally detached (un-invested if you will) from the various possible conclusions that could be reached. One part of this means that you’re willing to explore concepts/possibilities that could lead to conclusions that you don’t particularly like, but which you have to admit if being fully honest, need to be explored.

    I liked the term “strategic avoidance”. I’d never thought of it that way but I think you pegged it exactly. It’s really a form of dishonesty I’d say.

    • Yes, I think my suggestions as to what makes for productive debate apply more broadly. Whether that means that debate participants need to be uninvested in any particular conclusion is an interesting question. I think it’s possible (but perhaps difficult) to be at least somewhat invested in a particular conclusion, but also debate productively (and thus, possibly be led to change one’s mind). Investment in a particular conclusion on the one hand, and intellectual honesty on the other hand, need not be mutually exclusive.

      • Correction, you actually used the term “strategically avoiding”–my mistake.

        I agree that the two need not be mutually exclusive, but I think to the degree that one is invested in a particular conclusion/result, that for most people that makes perfect objectivity and honesty correspondingly difficult. I think this is an underlying tenet of, for example, Buddhist philosophy, has to take a good look at one’s attachments, and try to be free from them, or at least be aware of them, before making assessments (on anything). I have an interesting book titled “Buddhism and Ecology” but I don’t remember how much it discusses this topic. It’s probably been discussed elsewhere I’d imagine.

      • I agree.

      • I think I just won the “response eliciting the shortest response by Jeremy ever” award, didn’t I?

      • No.

      • aiyiyi, you used “strategically ignoring” not “avoiding”. I can’t even make corrections correctly.

      • However, when it comes to corrections of corrections, I’m not too bad.

  2. Back to the topic, I seem to remember, vaguely, that a lot of the Clementsian vs Gleasonian debates involved some first class talking past one another, often involving different understood meanings of terms, though I can’t give any examples off hand.

    • A debate with Clements getting hung up on terminology? Can’t imagine why…

  3. When reading this, I was going to respond almost like Jim Bouldin in his first post, as I do not see how debate in the ecological literature is any different than other types of debate. Although, I wonder if publicized is more efficacious or productive than non-publicized debate?

    I do not fully agree, however, with needing to be detached from conclusions. We would certainly be lying to ourselves if, when engaged in any debate, we did not have a preferred outcome from any given debate. (I feel like this is something that most scientists I know actually deny.) This is probably just a semantic difference from Bouldin, but I think that attachment is good, as being “attached” to seeking the truth is how we make progress in science. I think that some attachment is based on (i) seeking the truth and others are (ii) for their particular outcome. I’d further argue that the latter attachment is a bias that grows as a function of time in a particular area, which is why we see late-career scientists most often making these arguments. The fields that are newer or groups with high turnover (e.g., Bourbaki) fall more under the former category because they do not succumb to spending long amounts of time in an area, where competition (-,- for the pairs of individuals and – for the group as a while) ensues.

    One of the most fruitless debates in biology, in my opinion, is group selection. To my knowledge, group selection was putatively accepted dating back to Darwin and was not even an issue until George Williams ruined Wynne-Edwards. Based on little evidence, Williams almost single-handedly destroyed an idea and a person because of what was largely his beliefs. No, look where we are. Three decades went by where even peeping the forbidden words of “group selection” had one burned at the stake like a witch! I guess the idea ran its course, as it is back on the table and we can continue studying it. I was just wondering if its temporary banishment was overall more productive and better for the whole of knowledge.

    • Sorry, going to disagree with you on group selection. The group selection debate has had its unproductive moments (Nowak et al. is one, I think), but overall it’s been a rich and important debate, to which G. C. Williams made a very important contribution. Williams actually rooted out a lot of naive, unexamined, good-of-the-group intuitions (which mostly weren’t shared by Darwin, by the way, who actually had about as sophisticated and explicit an understanding of the issues as he possibly could’ve had). Williams didn’t ruin any ideas that didn’t deserve to be ruined. And his beliefs weren’t just his arbitrary personal preferences, like a preference for Pepsi over Coke or something. They were backed up with quite explicit arguments which anyone could read and decide about for themselves. It’s not as if he just browbeat people into adopting his own unargued-for personal preferences. If you think highly of current ideas about multi-level selection, then you should thank G. C. Williams (and others) for forcing proponents of group selection to “raise their game”.

      • Hi Jeremy,
        I was arguing that the debate was fruitless in the sense that we just picked up after nearly four decades on hiatus. I know that the arguments were valid, it was just that it had no been adequately experimentally and mathematically verified. That is why I prefer to not debate topics that have no evidence. What I think was particularly bolstering in the group selection debate was the ferocious arguing over the topic with little-to-no hard scientific evidence.
        I do pay respect to the biologists that laid the foundation for our current knowledge, but I am not about to type that the means by which we arrived at this end were the most efficient. I think that is the crux of this post–what does productive debate look like in biology? Surely as we gain knowledge we more closely approach truths, but under what circumstances is this expedited? Do you think that Lamarkian evolutionary processes being dismissed for over a century was good or bad for evolutionary biology?
        Thanks for the comments!

      • You’re right that the group selection debate has been marked by a relative paucity of empirical evidence on many issues, although not on all. Whether more evidence would’ve caused the debate to proceed differently, I’m not sure. Many of the issues are explicitly conceptual, meaning they’re to do with the proper interpretation of agreed evidence. It’s not clear how more evidence would settle such debates.

        Not sure what you mean by your remark about Lamarckianism at the end there. Lamarckianism in the sense of “what Lamarck actually said” was and continues to be dismissed (“Lamarckianism” in this sense is much more than “inheritance of acquired characters”). And yes, that is good for evolutionary biology.

    • Being “attached” to discovering the truth is a very good thing indeed, and should be the objective of every scholar. Being attached to some particular outcome or result, apart from wherever the evidence best leads, is the problem. It’s certainly fine, indeed necessary, to defend a conclusion or result that originates from good scientific practice. Hope that clarifies a bit.

      • Darwin’s writings are an interesting case study here. Darwin certainly didn’t lack attachment to his own theory of evolution by natural selection, but he was also a cautious guy who always wanted to have as much data as possible before settling on a conclusion, and who had a strong attachment to intellectual honesty and getting the right answer even if it wasn’t his preferred answer. He famously gave more weight to evolutionary mechanisms besides natural selection in later editions of the Origin, in response to criticisms which he felt unable to adequately address. As opposed to, say, trying to “sell” his idea by emphasizing only supportive evidence and arguments, and leaving it to his critics to offer contrary evidence and arguments.

        Having said that, it’s important to recognize, I think, that science *as a whole* can function just fine even if *individual scientists* are attached to their pet ideas. You just need different scientists to be attached to different ideas (and for no one to be so attached to their pet ideas as to willing to do unethical things like make up data). This is an “adversarial” model, if you like, somewhat analogous to the adversarial model in the US court system. David Hull is one of many folks who’ve analyzed the practice of science as a whole in these sorts of terms. Indeed, many scientific practices, not least peer review, basically function as ways of tempering the biases of individual scientists so as to ensure that science as a whole is less biased and ideally unbiased.

        I’ve been thinking about this recently because I’ve read some unfortunate stuff recently from some sociologists of science who seem to infer serious bias in science as a whole from bias on the part of individual scientists. Which is exactly like arguing that the US court system must, say, be systematically biased in favor of defendants, because some lawyers are defense lawyers.

      • Interesting; I didn’t know that Darwin gave more weight to other mechanisms besides natural selection in later editions, but it certainly doesn’t surprise me, given how meticulous, curious and wide minded he was. I don’t have the impression that self promotion meant much to the man.

        I haven’t read anything by David Hull but will put him on the list of things that will never get read 🙂 Based on your synopsis, I would have to disagree with him, for a number of reasons. First, I would side with seed dispersal in that this will often be an inefficient way to come to correct conclusions on a topic, analogous in my mind to the way that error in a regression model necessitates adding more data points to increase the confidence in the true signal–it would be much better (i.e. more efficient) if there were no error at all and you therefore only needed two data points to define your (linear) relationship.

        Second, I think it’s a slippery slope indeed to narrow one’s focus down to a particular viewpoint, easily leading to arguing, entrenchment of positions, misinterpretation of opposing positions, and hostility, in a worsening negative feedback loop. This benefits nobody and often requires a third party to come in with a fresh and wider perspective to clean up the mess. I mean, what’s the point of taking a particular stance if you know that there is legitimate evidence that argues that you might have missed something important or have a limited view? I see no redeeming value in that. Or is it rather, that people really don’t see any value in others’ arguments, in which case, maybe we scientists ain’t quite as sharp as we should be?

        As to the sociologists, I have to say that anybody who equates bias in certain individuals with bias in the science as a whole has some screws loose. You will find exactly that viewpoint in quite a number of those who deny human effects on climate change–I see them all the time. They’re over-generalizing from particulars. I think this is common among those who look at a given issue from “the outside”.

      • Yes, as someone (can’t recall who) aptly remarked, the 6th edition of the Origin (the final one published in Darwin’s lifetime) could’ve been titled “On the Origin of Species By Means of Natural Selection and a Whole Bunch of Other Things”. It’s for this reason that, in the seminar I teach on the Origin, I have students read the first edition. It’s shorter, it’s clearer, it’s better argued, it’s actually closer to getting things right (Darwin’s critics were mostly off-base, so accommodating their criticisms turned out to introduce errors), and probably most importantly, is the edition that actually started the revolution and changed the world. In contrast, the 6th edition (which, unfortunately, is probably what you’re getting if you just randomly pick a copy off a bookstore or library shelf) is mostly of historical interest as a record of Darwin’s final views.

        Don’t let my very brief synopsis shape your view of David Hull. I’ll add that much of Hull’s work on this is descriptive rather than prescriptive. He’s arguing about how science actually operates rather than arguing for an idealized vision of how it ought to operate. Well, that, and then explaining why the actual way that science operates can nevertheless work despite what seem at first glance to be some highly suboptimal features. Science As A Process is his big book on this. He argues that scientific progress is analogous to evolution by natural selection (Hull was a very good and influential philosopher of evolutionary biology, so if you’re inclined to dismiss this analogy, don’t–Hull certainly knows whereof he speaks here). The book uses the pheneticist vs. cladist debate in the 1960s as a case study, a good choice in some respects although perhaps a questionable choice in others (the debate was perhaps an atypical scientific debate in some important ways).

        I don’t necessarily disagree with the issues you and seeddispersal raise with the “adversarial” model of scientific debate. I just think that what’s key is, as you briefly suggest, the reaction of “third parties” to the debate. As long entrenched camps air all the issues, engage with one another rather than, e.g., strategically ignoring one another’s points, etc., then I think their debate will be productive in the sense that it will be useful to third parties, who will have all the evidence and argument they need to come to a reasoned decision on the issues. Although I will grant that such debates between entrenched camps will tend not to hit on resolutions like “hey, it turns out we’re both wrong!” or “hey, turns out our argument is pointless because we’ve both mis-framed the whole issue!”. Debates between less-entrenched camps, where all sides are more focused on getting things right and thinking things through than on defending their own current position, are probably more likely to hit on such “outside the box” resolutions.

      • Perhaps worth noting another productive example of the “adversarial” model of debate–debating societies. The Munk Debates provide a prominent modern example.

      • Although then again, I’m not sure how productive formal debates among US Presidential candidates are. Maybe that just goes to show that how well the “adversarial” model works depends on the formal and informal “rules” that govern the conduct of the debate, and on whether those rules are obeyed. I’m not sure.

      • Well, contrary to my previous comment, I’ve actually pulled Hull’s book off the shelf and begun reading, even if it may only be the first chapter. This definitely appears to be a book worth reading.

      • Reading stuff I recommend, as well as stuff I write, could start eating up a lot of your time if you’re not careful. 😉

  4. Thanks Jeremy.

    “…debates between entrenched camps will tend not to hit on resolutions like ….“hey, turns out our argument is pointless because we’ve both mis-framed the whole issue!”.”

    Something about an ice cube’s survival chances somewhere springs to mind 🙂

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s


%d bloggers like this: