tl;dr: No, my paper’s not having much impact. But that might be because there’s no potential impact for it to have, because the IDH is already a “ghost”.
Let me back up. For those of you who are just joining us, the intermediate disturbance hypothesis (IDH) says that species diversity is, or should be expected to be, a humped function of disturbance frequency or intensity. It’s a widely-cited, highly-influential idea that features prominently in various ecology textbooks, but it’s been strongly criticized on empirical and theoretical grounds (e.g., Chesson & Huntly 1997 (great paper), Mackey & Currie 2001, Miller et al. 2011). I don’t think those criticisms have been noticed and taken on board nearly as much as they should’ve been. So I’ve written a bunch of posts highlighting those criticisms and referring to the IDH as a “zombie idea”—an idea that should be dead, but isn’t. It lives on not because the people who buy into are bad scientists—they’re not—but because of the basic conservatism of the scientific process. The IDH was a good idea when it was first proposed, and it quickly took root. And once any idea takes root, it’s very hard to uproot (see also). But that doesn’t mean we shouldn’t try. Criticism of ideas is part of science.
But criticism of ideas on a blog only gets you so far, for various reasons. For instance, the majority of ecologists don’t read this blog.* So in 2013, I published an opinion piece in Trends in Ecology and Evolution that grew out of my blog posts, arguing that the IDH should be abandoned. It’s now been a bit more than two years since that paper was published, enough time for any impact it might be having to start showing up in the peer-reviewed literature. So I decided to look at how Fox (2013) is being cited. Did I slay the IDH? Or did my paper have the same effect as Clarence Darrow’s closing speech in the Scopes Trial?** Or somewhere in between?
As of mid-August, Web of Science listed 42 papers citing Fox (2013).*** That’s a lot for an ecology paper that’s only two years old! So between that, the fact that TREE is a prominent venue, and the fact that Fox (2013) was the most-downloaded TREE paper the year it was published, Fox (2013) certainly hasn’t gone unnoticed or been ignored. But just for context, Connell (1978), the classic paper that coined the term IDH, has been cited 581 times since 2013. Now, some of those citations are from papers having nothing to do with the IDH—but as you’ll see in a sec, the same is true for papers citing Fox (2013). So only a small minority of IDH-related papers published since 2013 cite Fox (2013).
It’s one thing for people to read or even cite a paper, another for the paper to affect what sort of science gets done and how that science gets reported and discussed. So I skimmed those 42 papers to see how they cited Fox (2013). Below, I summarize the results, because the big picture is what matters. I don’t want to single out particular authors. Note that one or two papers cited me in more than one way and so are counted twice.
No idea how they cited me because I couldn’t access the full text: 10 papers. Even though I’m at a large research university that subscribes to lots of journals. I think this has interesting implications, which I’ll return to.
Cited in passing: 15 papers. The most common category. These are citations for statements in passing that the IDH has “recently been criticized”, or “remains controversial”, or is the subject of “ongoing discussion”, etc. They usually occurred early in the introduction.
Cited as a primary reference on competition-colonization trade-offs: 3 papers cited me as a reference on how competition-colonization trade-offs work. Even though Fox (2013) doesn’t actually explain competition-colonization trade-offs—it just cites papers on the topic.
Cited as a primary reference on the storage effect: 1 paper that isn’t about disturbance and diversity cited me as a source for a statement about how the storage effect works. Which I can sort of see, since Fox (2013) has a textbox explaining the storage effect.
Cited as part of a reply to my paper: Sheil & Burslem 2013 (see also).
Cited because the authors totally agree with me: 4 papers cited me to say that the IDH, as traditionally defined, hasn’t worked out. Three of these went on to report research on diversity-disturbance relationships based on modern coexistence theory. The authors of these three papers are now my science crushes. 🙂 I’m especially pleased to see that one of these papers was co-authored by someone who reads and comments on Dynamic Ecology. :-)
Miscited: 2 papers miscited me. That’s annoying, but it’s not a big deal so I won’t say more. Neither is the worst I’ve ever been miscited.
Cited by mistake: 1 citation was for a statement about nutrient retention in lakes, in a paper that has nothing to do with disturbance or diversity. I assume the authors must’ve clicked the wrong “Fox 2013” in their reference management software.
Cited in some other way: 4 papers.
This little exercise changed how I think about the IDH. Here are my thoughts:
- I’m not surprised that the most common response to Fox (2013) in the literature has been to just note its existence in passing and move on. It’s very rare for any one paper, even a high-profile paper, to have much effect on anything. Especially a paper that tells people to stop doing something, but doesn’t spell out in any detail what they ought to do instead. Well, I did say people should apply modern coexistence theory to diversity-disturbance relationships. But I didn’t say how to do that. Fortunately, several groups are now doing it.
- Following on from the previous bullet: I of course can’t tell if any papers weren’t published because of Fox (2013), either because I changed people’s minds about what work to do, or because I changed reviewers’ and editors’ minds about what work to accept. But honestly, I highly doubt that Fox (2013) prevented publication of any papers that would otherwise have been published. That’s just not how science operates in my experience. (UPDATE: Let me just emphasize that I really did not expect Fox (2013), or my various blog posts, to make much difference to how the field as whole treats the IDH. These first two bullets are mostly aimed at any readers who are under the mistaken impression that a single critical paper or a few critical blog posts can have a big influence on an entire field.)
- Here’s the thing that struck me most: not only did these 42 papers mostly only cite Fox (2013) in passing, they mostly only cited the IDH in passing. Even the ones that were about diversity-disturbance relationships. Not that they didn’t mention the IDH at all–many began with a one-paragraph review of the IDH. But they hardly ever returned to the IDH later in the paper. And with very few exceptions, they weren’t framed as tests of the IDH, often focusing on system-specific biology instead. They mentioned the IDH, but went on to focus on the effects of disturbance in their own study systems. Without much or even any attempt at generalizing the results, and little or no suggestion of any implications for the IDH. Honestly, I’m not sure why the large majority of these 42 papers bothered citing the IDH, much less Fox (2013). Their citations of the IDH didn’t do any intellectual work–didn’t provide a strong motivation for doing the study, didn’t provide a testable hypothesis… Which isn’t a criticism of these papers, by the way. I’m not suggesting that they all should’ve been framed as tests of the IDH, or that they should’ve made more effort to generalize their results. I’m just describing how most of these papers cited the IDH, and Fox (2013)–as boilerplate or throat-clearing. Basically just nodding towards the IDH and what’s been said about it recently, before getting down to the specifics of the paper. Ironically, the papers that made the strongest use of their citations of the IDH, at least to my eyes, were the ones that agreed with Fox (2013). They used the lack of success of the IDH to motivate the need for an alternative theory of how disturbance affects diversity, and to motivate the need for applications or tests of that alternative theory.****
- Here’s the other thing that struck me: many of the citations of Fox (2013) were from papers in specialized journals (Geobotanica, Annals of Forest Science, Cold Regions Science and Technology…). Some of these journals are so specialized that not only hadn’t I head of them, but the big research university that employs me doesn’t even subscribe to them. Most of the remaining citations were from unselective journals like Plos One. And a couple of the exceptions were papers that weren’t even about the IDH. All this actually mirrors the citation pattern for Connell (1978). It’s a classic paper that’s massively cited—but most of its citations these days are in specialized, low-impact journals. Of those 581 citations it’s had since 2013, only 13% come from a long list of leading selective journals in general ecology/general biology/general science.***** And some fraction of that 13% are papers that aren’t actually about the IDH.
- The previous two bullets together suggest to me that maybe the IDH isn’t a zombie idea after all. Maybe the IDH is a ghost idea. An idea that, while not dead, isn’t sufficiently alive to have much effect any more. Ecologists writing for a broad audience about big, general ideas and reporting what they feel are major novel results mostly aren’t trying to test or further develop the IDH (there are exceptions). And ecologists writing for a specialized audience about disturbance-diversity relationships in particular systems mostly aren’t trying to test or further develop the IDH either. The IDH is just the big, general idea to which they give a passing nod before they turn their attention to the system-specific details of interest. So the IDH isn’t dead—but as best I can tell, it might as well be.
A while back, I used citation data to argue that the IDH remains a popular subject of active research—that it’s alive and well, even if I think it shouldn’t be. But now that I’ve seen how the IDH, and my critique of it, are cited, I’ve changed my mind on that. I think the IDH is still widely taught to undergrads, and I question whether it should be (depends on how it’s taught, I think). But for the reasons discussed above, and for other reasons, it now looks to me like a ghost idea as far as current research is concerned.
As always, looking forward to your comments. I’ve learned a lot from past comment threads on this topic.
*When I give talks on my blogging, and ask the audience who reads Dynamic Ecology, less than half the hands go up.
**”[P]recisely the same as if he’d bawled it up a rainspout in the interior of Afghanistan”, according to H. L. Mencken.
***I went with WoS because I didn’t want to have to sort through all the conference proceedings and other flotsam that Google Scholar indexes. You get the background research you pay for on this blog.
****Let me emphasize that I have no idea if that’s how the authors of these papers intended their citations of the IDH to come off. That’s just how they came off to me. That perhaps deserves a post of its own at some point: what is the introduction section of a paper for? I suspect there’s a range of views on this.
*****Science, Nature, Nature Communications, PNAS, Plos Biology, Proceedings B, Phil Trans B, Eco Letts, Ecology, Ecol Monogr, Ecol Appl, Am Nat, JAE, J Ecol, J Appl Ecol, Funct Ecol, Conserv Biol, Global Ecol Biogeogr, Ecography, Oikos, Oecologia, Ecosystems, Global Change Biol, AREES, TREE. I’m sure you could quibble with the composition of this admittedly-arbitrary list, but it wouldn’t change the results much. And before anyone asks, yes, this does seem to be a change in how Connell (1978) is cited. 40% of the 632 citations of Connell (1978) that occurred before 1990 were from the journals on my list. Even though several of those journals didn’t exist in 1978, or even before 1990.
Related posts
I previously did a similar exercise with a paper critiquing a popular approach in phylogenetic community ecology. It wasn’t having much impact either.
Dynamic Ecology 
It would be interesting to know how many people teach the IDH in their ecology courses. Is that one indication of whether something is a zombie or a ghost?
I’d be curious to know that too. For context, I think you’d want to ask about various other classic ideas of similar vintage. And more recent ones too, like biodiversity and ecosystem function.
The folks who teach intro ecology at Calgary have stopped teaching the IDH, thanks in part to my posts.
Interesting analysis. Only comment I have is that the paper also has an Altmetric score of 71, which is none too shabby for such a paper and given its age, including a citation on Wikipedia.
I confess I don’t pay any attention to Altmetric scores. Seems like it’s adding apples and oranges. I do like that the paper was widely tweeted when it came out, and widely shared on social media. I just don’t see any point in trying to combine various metrics into one number.
Yes, agreed, the Altmetric score is an odd one. But it’s at least transparent and you can see how it’s calculated. Unlike ResearchGate….
I think the problem is that your refutation is based on proving the original, linear model cannot generate the predictions it claims, if appropriate controls are accounted for. However, in the paper you then go on to show how many nonlinear models can generate IDH. So, the claim that IDH should be abandoned seems to rest on a subtle distinction between the “idea” behind IDH, and the mathematical model of IDH. Abandoning IDH seems like a stretch if the core argument against it is that the original formulations of the models which simulate it cannot generate its predictions. By making this claim you demonstrate that the idea and the model are separate. You then discuss the type of mathematical formulation required to generate the predictions of IDH. I think most people are fine with the need to invoke non-linear dynamics to generate IDH, and therefore do not see a need to abandon IDH as a useful conceptual framework.
Hope that doesn’t come off as combative! Its good to discuss these things, and I appreciate all your work on this!
Thanks for your comments, they’re not at all combative.
“However, in the paper you then go on to show how many nonlinear models can generate IDH.”
That’s not quite my view; let me clarify. I discuss how nonlinear models can generate coexistence in a disturbed or fluctuating system that wouldn’t otherwise occur. But those nonlinear models often (usually?) *don’t* produce coexistence of *more* species at intermediate frequencies or intensities of disturbance than at low or high disturbance frequency or intensity. See, e.g., Miller et al. 2011.
As to whether the IDH remains a useful “conceptual framework”, I guess I’d ask what intellectual work that framework is doing, if we grant that:
-we need to look at nonlinear models to understand coexistence in disturbed or fluctuating systems
-those nonlinear models embody coexistence mechanisms that were not contemplated by the originators of the IDH, who mostly focused on invalid mechanisms that can’t actually work (I admit this one is a bit debatable; you can read Connell 1978 as being based on competition-colonization trade-offs)
-those nonlinear models often don’t predict a humped diversity-disturbance relationship
-we don’t often observe humped diversity-disturbance relationships, even in controlled experiments
In light of that, what do we *gain* by continuing to talk about the IDH? What’s the point?
Heading to a discussion group on Fox (2013) vs. Huston (2014, Ecology) in a half hour…
I wonder whether many readers simply have great difficulty in fully understanding the arguments in Fox (2013). By your own admission, the theoretical issues are quite subtle. And many readers won’t take the time to (or simply can’t) follow all the mathematical details. One overarching argument – that disturbance can promote coexistence, but not for the reasons originally proposed – is a tricky one given that the verbal models do mention so much stuff. That’s one potential problem with verbal models, but it also means that identifying their most critical core components (to reject or to keep using) is not so easy. To me, one of the most interesting questions (mentioned by Huston) arising from all this is just how much overlap there is (or should be) between the study of patterns in species diversity and the study of mechanisms of long-term stable species coexistence. One and the same? Totally different? If somewhere in between, where exactly?
“Heading to a discussion group on Fox (2013) vs. Huston (2014, Ecology) in a half hour…”
So I’d better reply immediately so that I can contribute to the discussion from afar! 🙂
“I wonder whether many readers simply have great difficulty in fully understanding the arguments in Fox (2013).”
Quite possibly. All I can say is I did my best to both explain the issues, and explain why the theoretical subtleties are important to grasp if you’re going to work on or teach this topic. But at some level, if the topic’s hard, it’s hard. Ecology isn’t easy.
“One overarching argument – that disturbance can promote coexistence, but not for the reasons originally proposed – is a tricky one given that the verbal models do mention so much stuff. That’s one potential problem with verbal models, but it also means that identifying their most critical core components (to reject or to keep using) is not so easy.”
Agreed. For me, the key thing is that we all quit teaching and using ideas that don’t work. Ultimately, I don’t care about what Connell (1978) or whoever “really” meant. Sheil & Burslem argue that we don’t need to throw out the IDH, we just need to redefine it and narrow it so that it only includes valid ideas about how species coexistence can work. I’d be perfectly happy with that if it was possible. But in practice, I think it’s hard to keep the name for something while radically redefining it. I can think of cases where that’s been done–think of redefining “group selection” so that it no longer includes naive “good of the species” verbal models. But I can’t think of many. Further, I suspect that, if you do want to get everyone to redefine the IDH as Sheil & Burslem suggest, the best way to do it is to argue for its abandonment. The argument for abandonment hopefully forces people who want to keep the IDH to start paying attention to theoretical subtleties and decide which bits they want to keep.
“To me, one of the most interesting questions (mentioned by Huston) arising from all this is just how much overlap there is (or should be) between the study of patterns in species diversity and the study of mechanisms of long-term stable species coexistence. ”
That’s a very astute comment. A lot of what’s going on under the surface here is that *lots* of ecologists don’t actually care about mechanisms of coexistence per se. What they really want, ultimately, is to explain and predict species diversity. Which is totally fine, I definitely don’t think the study of coexistence mechanisms should *replace* studies of species diversity. But I would say that, if you care about explaining patterns in species diversity that are maintained over the long term, then at a minimum your explanations had better not contradict what we know about mechanisms of species coexistence.
I think one important direction for future work on coexistence theory is to try to turn it into a theory of species diversity. Under what conditions (combinations of model parameters) can more or fewer species coexist? Are there any general reasons to expect certain coexistence mechanisms (or *all* coexistence mechanisms) to be weaker or stronger in many-species communities? Is is easy to get lots of species to coexist just by throwing lots of different coexistence mechanisms in together? (the answer to that last one seems to be “no” based on the few models we have so far…). Note that it’s quite possible that we won’t find any general patterns if we go down this road. Miller et al. 2011 suggests that we shouldn’t actually expect any general relationship between disturbance and diversity. Peter Abrams has an old review paper arguing that there’s no reason to expect a general relationship between local-scale productivity and local species diversity, based on the mathematical theory available at the time. But we’ll see. FWIW, I think you’re going to start seeing more groups predicting patterns of species diversity using mathematical models of coexistence mechanisms. Kat Shea’s group at Penn State is doing this.
I think part of the uphill battle is that the IDH theory isn’t named “coexistence-driven diversity peaks at intermediate levels of disturbance”. It is named “intermediate disturbance hypothesis”. So the name itself doesn’t invoke coexistence mechanisms even if the first proposers do. Thus I think in many people’s minds IDH is inherently an empirical pattern (its not exactly a claim of correlation since its hump-shaped but basically its not much more than a claim of correlation between two variables). While the empirical evidence for this claim is surprisingly mixed, it is certainly not decisively against IDH. And there are plenty of non-coexistence theories that can produce IDH (e.g. ones that incorporate successional dynamics).
But I do agree with you that if that is the version of IDH people are promulgating, then they need to be very explicit in rejecting the original coexistence derived version. Its kind of having your cake and eating it too to continue to bask in this aura of having a mechanism but then claiming you are not tied to the mechanism when people talk about it being wrong.
And on the topic of the distinction between species richness and coexistence theory, I think this is fundamental and important. It is not at all obvious to me (indeed it is counterintuitive to me but this parenthetical claim is subjective) that coexistence theory will ever (or should ever) be our best approach to developing theory around species diversity. I think we do know at this point that non-closed systems (dispersal) and evolution and environmental gradients play central roles in diversity (at different scales) and if I was looking for a place to bolt those on, coexistence theory wouldn’t be my first starting point I don’t think.
“Sheil & Burslem argue that we don’t need to throw out the IDH, we just need to redefine it and narrow it so that it only includes valid ideas about how species coexistence can work.”
Thanks for the reference … a pedantic point, but I think it matters: Rather than “redefine” I would say we should simply return to the original successional “trade-off” model that Connell explained and illustrated. The problem is not with this original idea — we all seem to agree it is correct. It was what was first called the “IDH”.
Problems arise with ideas that were subsequently mixed in under the IDH label — those are the “misuses” and “errors” that need to be corrected. Misuse of the label does not make the original idea false (or ghost or zombie) — it remains correct. It just needs to be correctly applied.
@Douglas:
Fair point and not pedantic at all; my phrasing here is sloppy and puts words into others’ mouths. I should’ve just said “restrict” or “limit” or some such. Apologies.
I continue to respectfully disagree about what Connell originally meant–I don’t think it’s entirely clear what he originally meant. As for misuse of labels vs. correctness of original ideas, I also respectfully disagree. I think we have to critique the IDH we have, not the IDH we wish we had (or, arguably, had in 1978).
But in any case, I now think the point is largely moot. As I said in the post, I don’t see much evidence that the general mechanisms underpinning diversity-disturbance relationships are of much interest to a critical mass of current researchers. And insofar as that’s likely to change, I suspect change will come not from a revisiting of old ideas, but as as part of a broader burgeoning of interest in modern coexistence theory. A burgeoning that I think is underway already: see https://dynamicecology.wordpress.com/2015/08/15/esa100-impressions/.
But these are all matters on which reasonable disagreement is possible.
“at a minimum your explanations had better not contradict what we know about mechanisms of species coexistence” – totally agreed.
“one important direction for future work on coexistence theory is to try to turn it into a theory of species diversity” – I look forward to seeing that! Although I agree with Brian that there’s a whole lot more to patterns of species diversity than coexistence mechanisms, so the question remains open as to whether coexistence theory is the key piece, or just one of many things to make sure is not contradicted.
I also agree that many see the IDH as a “do you see what I see?” theory. Grime saw a pattern, had a compelling explanation for it, and subsequent tests often just asked whether the pattern was repeated somewhere else – important to know, but without providing a ton of insight into what generates the pattern to begin with. Same goes for the hump-shaped productivity-diversity relationship.
@Brian and Mark:
I agree that many people just see the IDH as a purely empirical claim, or at least don’t care all that much about the underlying mechanisms. I’d say that the IDH’s track record as a purely empirical claim is lousy–too lousy for it to be considered a general pattern, and based on too much data to expect its track record to improve if only we collect some more data. My sense is that’s pretty widely acknowledged at this point.
Maybe more importantly, I’d question whether “do you see what I see” pattern hunting is an effective research approach in ecology. There are times when “black box”, purely phenomenological research approaches can work. Medical research is probably the best example–you just put drugs into people, along with placebo controls, and record what happens. Without any (or at least very much) theorizing about why you get any differences between the effects of drug and placebo. I think medical research is less like that these days, but for decades that’s pretty much how it was. That approach works if you’re trying to discover, say, if penicillin fights bacterial infections. But I think it’s ill-suited to any situation in which there’s lots of among-unit variability relative to the effect size you’re seeking (as is just now currently being recognized in psychological research, although psychologists seem to have often fooled themselves into thinking their results have stronger theory behind them than they actually do.) Basically, if you’re doing pattern hunting, and you don’t find a really clear-cut pattern, a strong signal that jumps out from the noise, you’re stuck. You’ve hit a dead end. Well, you can try seeing if a pattern emerges if you control for some covariates or something. But in practice in ecology, you generally only have a vague idea what covariates might be relevant, and even those that you think are relevant often turn out not to be. Lots of examples that could be given. Don’t get me wrong, I think at an early stage in a research program, pattern hunting can be a useful thing to do. Who knows, maybe you’ll stumble across a strong pattern, and then theory has a good target to shoot at. That’s more or less what happened with research on biodiversity and ecosystem function. But at some point, all concerned should be prepared to admit that there’s just no pattern there (at least not one strong enough to be worth caring about), give up, and move on.
Re: coexistence theory as a theory of species richness, I’d just like to see people who want to explain species richness (as opposed to merely describing it or statistically predicting it) to derive their hypotheses from mathematical models. So that we all know what exactly the hypothesis is and what it predicts. I’m sure that, for many species richness patterns of interest, the relevant models will incorporate environmental gradients and dispersal–indeed, we already have various models that do that. (Aside: you can perfectly well run such models through a Chessonian analysis and fit them into the framework of modern coexistence theory. Chesson’s stuff doesn’t assume a closed system or a spatially-homogeneous system). I wouldn’t be at all surprised if, empirically, there’s no particular relationship between the strength of some Chessonian class of coexistence mechanisms (the storage effect, relative nonlinearity, fluctuation-independent mechanisms…) and the number of locally-coexisting species.
Brian: not sure what you mean by “non coexistence theories” of succession that can produce an IDH-type pattern. Successional models of the sort with which I’m familiar certainly do fit within modern coexistence theory. I’m thinking for instance of Pacala and Rees’ “successional niche” model (Am Nat 1998 I think? going by memory; it’s cited in Fox 2013). The long-term coexistence in such models of course happens at the landscape level rather than within any particular patch. Which just goes to show that coexistence in such models depends on spatial or spatio-temporal variation–which fits within modern coexistence theory.
@Brian and Mark:
Another issue that plagues a lot of pure pattern-hunting research in ecology is people disagreeing on exactly what pattern they’re looking for and how to measure it. See recent debates over the hump-backed model for a prime example, but it crops up with the IDH too. What’s “diversity”, exactly? Species richness? Some diversity index? (take your pick, you’re spoiled for choice!) Or what? What’s “productivity”? Is it total biomass? If so, should that include leaf litter? Or what? Are we looking for a humped relationship involving mean diversity? Or maximum diversity? Or what?
The trouble is, these debates are mostly unresolvable. For instance, appeals to the statistically-desirable properties of this or that diversity index are hardly ever decisive (see https://dynamicecology.wordpress.com/2012/05/02/advice-on-choosing-among-different-indices-of-the-same-thing/ and https://dynamicecology.wordpress.com/2015/06/25/the-most-common-way-to-fish-for-statistical-significance-in-ecology/).
” I think we have to critique the IDH we have, not the IDH we wish we had (or, arguably, had in 1978)”
Thanks for the thoughts. This point nicely captures the essence of our main disagreement.
Let me try an analogy that may help persuade. Taxonomists try and standardize taxon names. When one organism ends up with multiple names or one name has been applied to more than one distinct taxon the earliest published description that provides sufficient characters to provide an operational definition is generally considered valid. This implied validity holds even if the resulting taxonomic key is challenging to apply and results in mistakes. We would never discard a name just because people make mistakes when trying to apply it (though there might be efforts to improve the key).
It seems to me that we should do this with our named-hypothesis too Should we rather accept that fundamental definitions meander about in a popularity contest.
Surely the need for clear fixed definitions is something most of us could agree on? Then we could then discuss what constitutes an adequate definition.
@Douglas,
Hmm, interesting thought. But I don’t think the cases are analogous. Scientific hypotheses aren’t sufficiently like species. I don’t think we can, or should, name scientific hypotheses the way we name species. And I don’t think trying to do so would add clarity to substantive scientific discussions.
Mark, it would be fun if you would give a few highlights from the discussion. I know I would be interested to hear anyway.
Thanks to Jeremy for those thoughtful responses. In response to Jim, key points/outcomes in our discussion, based on my biased recollection of a free-flowing exchange among ~15 people:
(1) It takes a fair bit of knowledge of ecological theory/models to really grasp some of the arguments in Fox (2013). Most grad students need more training in theory to fully understand statements like this: “trade-offs generate disturbance-mediated coexistence via nonlinearities and nonadditivities, not because disturbances simply reduce species’ densities”.
(2) Neither of the two approaches outlined by Huston – “Empiricism” and “Logic” – can be thought of as “better”, but rather a feedback between them is optimal. That said, this is not so easily accomplished. If a verbal model includes a very long list of possible underlying processes, it will almost always be possible to build a simulation that reproduces a qualitative empirical pattern, rejection of the theory might be nearly impossible, and you end up with not much insight into what key processes created the initial observed pattern (point for Fox). But if a model is simple enough that we can fully understand all of its ramifications, it is likely too simple to capture what’s happening in nature (point for Huston – his example is a model of disturbance that doesn’t simultaneously deal with productivity).
(3) Some dismay from a senior ecologist that we’re still talking about this so much, given that 25 years ago it seemed that Huston’s DEM model was more or less accepted and we should have moved on to other unsolved problems.
(4) Some uncertainty/disagreement among people about whether Fox and Huston were (1) viewing the same problem from different angles, and so would actually agree given time to discuss, (2) viewing the same problem but fundamentally disagreeing on the processes that link diversity and disturbance, or (3) not even attacking the same problem (coexistence in one case, diversity patterns in the other).
@Mark:
Re: your #4, let me know when you figure out the answer! Between Huston’s criticisms of Adler et al., his criticisms of Fox 2013 (neither of which is about the DEB “model”), and his defenses of the DEB model, I too am unclear on what the root disagreements are here (there’s probably more than one “root”).
I doubt we’d agree if given enough time, though.
Jeremy said ” I don’t think we can, or should, name scientific hypotheses the way we name species. And I don’t think trying to do so would add clarity to substantive scientific discussions.”
Thanks for the reaction but I’m not sure I follow the reasoning. We do name scientific hypothesis and we do want that to represent something clear and meaningful dont we? For example, your article and many blogs refers to the IDH and what you and I understand by that term is a source of disagreement.
What are the alternatives? If you don’t want names then how should we refer to them?
I don’t see any convenient options myself, but perhaps others see this very differently so I am interested in pinning this down.
thanks again
@Douglas
Just to explore this analogy further: when we name new species we assign a type specimen. What would be the type specimen for a hypothesis? Would it be the paper itself? Or the person/people who proposed the hypothesis? Or the hypothesis? In each of these cases there are difficulties with archiving/curating the type in the same manner as we would for type species. It’s an interesting idea but I think I agree with Jeremy that putting a fixed name to a hypothesis is more problematic that for a species (which in itself is not at all straightforward!)
@Jeff
thanks for the interest.
What I was hoping for from Jeremy was some agreement that clear names and definitions are desirable. Then indeed we can get into the details of just what an acceptable “type” definition would consist of. I think a clear written definition of some type will work (so yes the publication) but that is separate to just asking whether a clear accepted definition is desirable … it is desirable isn’t it? (I don’t see the argument that says it isn’t desirable just because it is difficult.) Any help with pinning this down is welcome …!!!
Let me suggest the need for a future blog piece on just how ecological theory should be defined and labelled. It seems an important topic, but is maybe too big for a comment thread alone. Answers thus far have only suggested what it shouldn’t be … I’d be interested to hear ideas on what it should be. This goes beyond arguments on the IDH (but may also clarify views on that). Just an idea as 1. I would be keen to hear and 2. I think it would be constructive. Thanks.
Thanks for the suggestion Douglas, but afraid I can’t promise anything. I’m still struggling to “get” why it’s so important to you that we have some sort of agreed, taxonomy-style system for naming scientific ideas. Not only do I have no idea how such a system could work (sorry), I still struggle to understand why one would want one in the first place.
I do think it’s interesting to look at how changes in scientific terms map onto underlying changes in scientific ideas. For instance, I remember reading an old paper by historian and philosopher David Hull, suggesting that scientists who win intellectual debates tend to be those who are able to change their meaning so as to rebut or defuse criticisms, *without* changing terminology (or even admitting that they’ve changed their meaning). As I recall, he used punctuated equilibrium as an example.
” I’m still struggling to “get” why it’s so important to you that we have some sort of agreed, taxonomy-style system for naming scientific ideas. Not only do I have no idea how such a system could work (sorry), I still struggle to understand why one would want one in the first place.”
Let me try and persuade you using the IDH as the illustration (though the point was meant more broadly).
It struck me that we agree on the valid versus invalid concepts and mechanisms. That’s good. But we don’t agree on the label and what it implies. That is a problem (whole papers and debates are now built on our differing use of labels rather than on ecological mechanisms and relationships).
What is the “IDH”? I think I know and so do you, but it appears we don’t agree. There is no generally accepted definition (I admit I thought there was, but let me set that aside here). Our dispute may mislead and confuse others that the successional trade-off ideas have been invalidated. We agree they have not been invalidated. But does everyone reads carefully enough to note that caveat? So perhaps we may be doing more harm than good with the focus we have.
How do we fix that (put the emphasis back on what we can agree on and on the ecological content not the package)? I was pondering that and think I see a plausible answer (clear system of definitions) but am open to others.
Surely noone is in favour of definitions that are flexible enough to permit whatever misapplications arise. Don’t we need something more robust?
My taxonomy argument was just for fun: to illustrate that the definitions problem is not unique. The formal precision of mathematics may be a better model than taxonomy for the kind of structure we might seek. Certainly open to ideas.
I agree your point about slippery terminology winning arguments. But that also seems to be a reason to seek clear and sharp definitions (avoidance of acknowledging that meanings changed is surely not a “good thing” [except maybe for the arguer?].
Lack of clarity and plastic definitions adds to the problem of doing good ecology too: if I am testing a different theory than you but call it the same name that cannot be idea.
Anyway, my thought was if we all understood each other, and the theories, better that has to be a good thing. (So we would focus on the ecological content and we could agree to abandon “that IDH” but not “this IDH”). That is my basic line of thought … hope it makes better sense!
@Douglas,
Thanks for elaborating. I agree with you 100% on the value of precision. I think you ask a very good and important question when you ask how to increases the precision of our reading, thinking, and writing in ecology. Unfortunately, I don’t think there’s any easy answer to that question. In part because there are sometimes good reasons (or seemingly-good reasons) to read and think superficially:
“I agree with you 100% on the value of precision. … Unfortunately, I don’t think there’s any easy answer to that question.”
Good … thanks, glad we have found agreement.
It seems to me that such imprecision (here of labels) is the source of our “IDH debate” (and to other similar debates in the literature). Yes there are challenges but it is up to us to confront that I think — how do we promote precision in the face of the challenges? As you say there may not be “easy answers” but I wanted to keep pushing as I suspect we can do more nonetheless. What are the options? As members of the research community (supervisors, teachers, examiners, writers, reviewers, editors …) what might we do differently?
@Douglas,
Apologies if I’ve linked to it already in this thread but forgotten: my old post on when (if ever) vagueness is a virtue in science: https://dynamicecology.wordpress.com/2015/06/21/when-is-vagueness-a-virtue-in-science/
“old post on when (if ever) vagueness is a virtue in science”
Thanks, yes, it is a good blog and I agree with the general points made (also in the comments).
The IDH is an interesting test case. We have a conflict between ‘my’ quite narrow IDH definition (that you agree does not need to be discarded at least as a concept) and ‘your’ broader IDH topic label that involves various people in various places quoting and misusing this original definition that we agree on (and mix in a few such as Huston who seem quite clear about their ideas being distinct). We agree many of these wider ideas and confusions are mistaken (and sure these should be abandoned, though I don’t see these as legitimate IDH ideas at all). So, if we follow some kind of simple taxonomic thinking, we have IDH (sensu stricto Connell 1978, who uses the Budongo forest succession as his type, as in Eggeling 1947 J.Ecol. ) versus IDH (sensu lato various secondary sources compiled by JFox 20013). We agree what they are, and on their validity, but repeating and repeating the simple messaging “abandon the IDH” and “the IDH is a zombie” seems too imprecise to me and undermines at least some of the clarification value that your article achieved otherwise.
We need clear sharply defined theories. I agree some vagueness can sometimes be useful or necessary, but not as part of an argument for apparently valid theories to be discarded with the duds — that requires precision.
Thanks for engaging. I’ll be quiet now!
On an only slightly related note: How should the concept of Zombie Idea in ecology be cited? The TREE paper dodges the term 🙂
You’ll have to cite my blog posts if you want to refer to the concept of zombie ideas. Yes, you can do that:
http://blogs.plos.org/mfenner/2011/07/21/how-to-formally-cite-a-blog-post/
http://blog.apastyle.org/apastyle/2012/02/how-to-cite-pseudonyms.html
Thanks. Of course it now also involves identifying Ghost Ideas, Zombie Ideas, and ideas that are Expected to Make the Transition from Living.
Don’t forget Buddy Holly ideas. 🙂
Those are really rare though. Much rarer than ghosts or zombies.
Pingback: Poll results: Which classic topics do ecologists care about? And which ones do they think ecologists collectively should care about? | Dynamic Ecology
Pingback: What are the most influential opinion/perspective papers in ecological history? | Dynamic Ecology
Pingback: Should publishers flag replication failures in the same way they flag retractions? | Dynamic Ecology