Monday, January 7, 2013

Reinventing the ecological wheel – why do we do it?

Are those who do not learn from (ecological) history are doomed to repeat it?

A pervasive view within ecology is that discovery tends to be inefficient and that ideas reappear as vogue pursuits again and again. For example, the ecological implications of niche partitioning re-emerges as an important topic in ecology every decade or so. Niche partitioning was well represented in ecological literature of the 1960s and 1970s, which focused theoretical and experimental attention on how communities were structured through resource partitioning. It would be fair to say that the evolutionary causes and the ecological consequences of communities structured by niche differences were one of the most important concepts in community ecology during that time. Fast-forward 30 years, and biodiversity and ecosystem functioning (BEF) research slowly  has come to the conclusion that niche partitioning to explains the apparent relationship between species diversity and ecosystem functioning. Some of the findings in the BEF literature could be criticized as simply being rediscoveries of classical theory and experimental evidence already in existence. How does one interpret these cycles? Are they a failure of ecological progress or evidence of the constancy of ecological mechanisms?

Ecology is such a young science that this process of rediscovery seems particularly surprising. Most of the fundamental theory in ecology arose during this early period: from the 1920s (Lotka, Volterra), 1930s (Gause) to 1960s (Wilson, MacArthur, May, Lawton, etc). There are several reasons why this was the foundational period for ecological theory – the science was undeveloped, so there was a void that needed filling. Ecologists in those years were often been trained in other disciplines that emphasized mathematical and scientific rigor, so the theory that developed was in the best scientific tradition, with analytically resolved equations meant to describe the behaviour of populations and communities. Most of the paradigms we operate in today owe much to this period, including an inordinate focus on predator-prey, competitive interactions, and plant communities, and the use of Lotka-Volterra and consumer-resource models. So when ecologists reinvent the wheel, is this foundation of knowledge to blame, is it flawed or incomplete? Or does ecology fail in education and practice in maintaining contact with the knowledge base that already exists? (Spoiler alert – the answer is going to be both).

Modern ecologists face the unenviable task of prioritizing and decoding an exponentially growing body of literature. Ecologists in the 1960s could realistically read all the literature pertaining to community ecology during their PhD studies –something that is impossible today with an exponentially growing literature. Classic papers can be harder to access than new ones: old papers are less likely to be accessible online, and when they are, the quality of the documents is often poor. The style and accessibility of some of these papers is also difficult for readers used to the succinct and direct writing more common today. The cumulative effect of all of this is that we read very little older literature and instead find papers that are cited by our peers.

True, some fields may have grown or started apart from a base of theory that would have been useful during their development. But it would also be unfair to ignore the fact that ecology’s foundation is full of cracks. Certain interactions are much better explored than others. Models of two species interactions fill in for complex ecosystems. Lotka-Volterra and related consumer-resource models make a number of potentially unrealistic assumptions, and parameter space has often been incompletely explored. We seem to lack a hierarchical framework or synthesis of what we do know (although a few people have tried (Vellend 2010)). When models are explored in-depth, as Peter Abrams has done in many papers, we discover the complexity and possible futility of ecological research: anything can result from complex dynamics. The cynic then, would argue that models can predict anything (or worse, nothing). This is unfair, since most modelling papers test hypotheses by manipulating a single parameter associated with a likely mechanism, but it hints at the limits that current theory exhibits.

So the bleakest view of would be this: the body of knowledge that makes up ecology is inadequate and poorly structured. There is little in the way of synthesis, and though we know many, many mechanisms that can occur, we have less understanding of those that are likely to occur. Developing areas of ecology often have a tenuous connection to the existing body of knowledge, and if they eventually connect with and contribute to the central body, it is through an inefficient, repetitive process. For example a number of papers have remarked that invasion biology has dissociated itself from mainstream ecology, reinventing basic mechanisms. The most optimistic view, is that when we discover similar mechanisms multiple times, we gain increasing evidence for their importance. Further, each cycle of rediscovery reinforces that there are a finite number of mechanisms that structure ecological communities (maybe just a handful). When we use the same sets of mechanisms to explain new patterns or processes, in some ways it is a relief to realize that new findings fit logically with existing knowledge. For example niche partitioning has long been used to explain co-occurrence, but with a new focus on ecosystem functioning, it has leant itself as an efficacious explanation. But the question remains, how much of what we do is inefficient and repetitive, and how much is advancing our basic understanding of the world?

By Caroline Tucker & Marc Cadotte


Jeremy Fox said...

Great question. Certainly there are specific ideas that have been rediscovered many times. Apparent competition is one, as Bob Holt himself has always been careful to point out.

More broadly, I suspect "reinvention" is often in the eye of the beholder. The stereotype of old scientists is that they always think that everything the youngsters are doing these days was known long ago. While the stereotype of the youngsters is that this just shows that the oldsters are out of touch with the details of modern work. Probably the "truth" (if there is one) is a mix of both. There probably aren't any *really* fundamentally new ideas left to be discovered in ecology--no scope for a future ecological Darwin. But that doesn't mean the work of the science is done, any more than evolutionary biology was finished by Darwin.

So in suggesting that ecologists often just put old wine in new bottles, you two are writing like much older people! ;-)

Hans Castorp said...

I'm afraid that it is not a problem of familiarity with literature, but of familiarity with nature. McArthur, for instance, was not only a mathematician, but also a good ornithologist and a careful observer. Every now and then we get out of the office and turn to nature, and "rediscover" a few basic facts that are known as the "foundations" of ecology. But of course the latter approach is methodologically flawed and logically inconsistent.

Caroline Tucker said...

Hi Hans - I think that you're getting at a question we have blogged about before here, which is the tension between empirical (hands-on) and theoretical research. And there is definitely something to that, but even if people are just going out and "rediscovering" facts about nature, they should be aware that multiple people have already discovered and written about those facts. And if "nature" were so simple that we can go and discern the important aspects of it, there would be no need for ecology, in all its complexity. But yes! we should all go outside more often :)

Jeremy - I would be the old, pessimistic one. Marc is still willfully optimistic about the future of ecology ;) Even if we've discovered everything we need to know in ecology (hopefully not?), we haven't put it together yet into a useful framework. Given limitations on funding, it seems like rediscovering or restating old ideas is still a poor use of our time.

Hans Castorp said...

Literature is not theory. Empirical work is not the same as familiarity with nature. And the three laws of mechanics are extraordinarily simple and yet are the foundations of all physicis (quantum mechanics too). Of course from basic facts to every ramification of theory there is a long way to go. But consider for instance Chesson; he did not the so much repeated (and rather boring) complain about the scarce realism of Volterra model; he went beyond, and proposed the storage model 8and others) but the storage model is build on the foundations of Volterra's

Anonymous said...

I'm a young researcher who does read and cite the "classic" literature on a topic. However, recently I have been surprised to receive several reviewer comments along the lines of "There are many more recent papers that could be used to support this point instead of citing outdated literature." Seems to me that there is a subset of ecology that doesn't see any value in foundational papers.

Marc Cadotte said...

I agree with this, and it is a little puzzling in some ways. Just because a paper was published more recently should not mean that it has greater value, rather citations should be ranked according to: 1) appropriateness; and 2) precedence -giving priority to older papers that first addressed a particular question. But I would guess that the focus on journal 'importance' (heavily weighted to impact factors that are calculated using citations from the previous 2 years) has resulted in shift towards valuing the newest and latest.

Tim Poisot said...

Hi Caroline, Marc,

Thanks for the great ideas in this post. I've proposed an alternative point of view here: My main point is that perhaps, the increasing rhythm of publications, and the way we structure our introductions, make it appear as if we were always trying to re-invent the wheel. I'd love to read your ideas on this point as well!

Anonymous said...

This is a great discussion - I'm glad I found this post!
As Anonymous has pointed out, when I did my undergrad degree, it was constantly impressed on us to cite 'recent' literature (i.e. last 10-20 years), unless it was some seminal work. I still come across many who think this.

I don't agree this is always true - mostly because one of the most thrilling parts of research is tracing things back to their original source! Scientists are ultimately detectives, after all! :)

Caroline's 2 questions (foundation lit being incomplete or education/practice being the problem) are valid. We can't do anything about the first one, but we can do something about the second. I've written quite a few posts about how I think the education system affects science - I think it is something that really needs to be addressed (in some countries more than others!)

I just read an article about how 'doing research' means 'googling' for much of the younger generation today - anyone who teaches or marks assignments can tell you that! Maybe there is an opportunity for curricula to spend just as much time on the foundation literature and history of current theories and concepts (as compulsory, not elective courses) as on practical skills and new developments? They are complementary bodies of knowledge, not mutually exclusive.

And I agree with Hans, multidisciplinary knowledge is also valuable - most of the 'great' scientists were 'multi-professionals', and you could argue that the reason they had such an impact on history and science today, is because they had such a breadth of knowledge and skills! Good research doesn't happen in a knowledge vacuum :)