Deconstructing niche construction: A TVOL conversation between Gordon Burghardt and Kevin Laland
by Gordon M Burghardt & Kevin N Laland
26 June 2017
The concept of niche construction stresses a dialectical relationship between organisms and their environments, rather than one being passively shaped by the other. It has deep roots in evolutionary thought but only now is resulting in a systematic research program as part of the Extended Evolutionary Synthesis (EES). Join Gordon Burghardt and Kevin Laland as they take a deep dive into the subject.
Kevin co-directs the extended evolutionary synthesis research program funded by the John Templeton Foundation. Gordon is Alumni Distinguished Service Professor of Ecology and Evolutionary Biology at the University of Tennessee in Knoxville, with a longstanding engagement in the subject.
Gordon: The extended evolutionary synthesis research program, supported by the John Templeton Foundation, is a great opportunity to move the field of evolutionary biology along in multiple directions. One of the 22 projects, led by Kevin Laland, is entitled “Adaptive trends and parallel evolution generated by niche construction”. Kevin, I think we should start with some background explanation for our readers based on the following questions. What is niche construction? How can it be recognized? What are a few examples? And, finally, why is it an important aspect of the extended evolutionary synthesis?
Kevin: As the term implies, ‘niche construction’ refers to the process by which organisms modify their own and each other’s evolutionary niches. Examples of niche construction include the building of nests and burrows and modification of physical and chemical conditions by animals, and the alteration of soil chemistry and nutrient cycling by plants. If such activities alter natural selection, then evolution by niche construction can occur. The niche construction perspective was brought to prominence through the writings of Harvard evolutionary biologist Richard Lewontin. Lewontin pointed out that organisms are not just passive victims of selection, but actively construct and modify environmental conditions in nonrandom ways, in the process biasing their own evolution.
My interest in the extended evolutionary synthesis began with the recognition that similar emphases on constructivism and bias are features of other biological subfields. For instance, there is a constructivist tradition in developmental biology which, rather than seeing form as encoded in instructions in the genome, emphasizes how organisms shape their own developmental trajectory by constantly responding to, and altering, internal and external states. Likewise, current interest in extra-genetic forms of inheritance has led to the recognition that, in addition to passing on genes to the next generation, parental organisms actively construct developmental environments for their descendants. In addition, partly in response to Mary Jane West Eberhard’s influential writings and her emphasis on “genes as followers rather than leaders in evolution”, research into developmental plasticity has explored the active role of the phenotype in directing evolutionary events. Also relevant here is evolutionary developmental biology (evo devo), which has drawn attention to the fact that phenotypic variation can be biased by the processes of development, with some forms more probable than others. This bias again resonates with niche construction theory, which emphasizes organism-derived biases in selective environments.
As I see it, the extended evolutionary synthesis merely highlights the common ways of thinking about the processes of development (constructive development rather than programmed development) and the active role that the phenotype plays in evolution that independently pre-existed in all these subfields. It draws attention to the fact that there is another way to think about evolution than the standard line. That is not to suggest for a second that traditional evolutionary perspectives are wrong or unproductive, only that it might sometimes be useful to think about problems in a different way. Niche construction is no more central to the extended synthesis than any of these other fields with which it shares a common mindset.
Gordon: Thanks for this compact description, which we will explore more thoroughly later. At the moment, however, I want to discuss niche construction in a historical context. Exploring the history of ideas and findings is something I have found valuable, and published on, in many of the research topics I have explored. You mention that Lewontin “brought the niche construction perspective to prominence.” In what sense do you mean this? Did he verbally make the case that evolutionary biologists should take the phenomenon seriously, perform empirical research, demonstrations, or tests, or did he model niche construction in a pioneering way (like Hinton and Nowlan did in their paper on the “Baldwin effect”, which so impressed Maynard-Smith that he quickly drew attention to it in Nature in 1987)?
Kevin: Lewontin’s case was a verbal one. He wrote a series of very influential essays that laid out the basic arguments. He stressed how organisms, through their activities and choices, determine what is relevant in their environments, and hence influence how natural selection acts on them. However, while these essays were inspirational to a large number of people, what they didn’t do was spell out how evolutionary biologists could cash in on this way of thinking through experiments or formal theory. As a result, Lewontin’s articles were generally read as an aspect of his wider critique of adaptationism. The same, incidentally, holds for Conrad Waddington, who made similar points more than a decade before Lewontin. Waddington had argued that adaptive evolution depended on four factors, the familiar genetic and natural selective systems, but also what he called the ‘epigenetic system’ (what nowadays we would refer to as a role for developmental plasticity in evolution) and the ‘exploitive system’, which equates to niche construction. These were influential ideas that left lasting legacies, but nonetheless were generally interpreted as part of Waddington’s critique of traditional neo-Darwinism. My collaborator, John Odling-Smee at Oxford University, was the first person to attempt to turn Lewontin’s ideas into a program of research.
Gordon: Yes, this view of Lewontin is my recollection also. When the University of Tennessee hosted the Society for the Study of Evolution meetings in Knoxville in 2001, Dick was the invited featured speaker and gave a public talk about the importance of niche construction. Let me tell an anecdote from that meeting as a move into a larger point. At the end of the meetings we had a large evening BBQ gathering at the Knoxville Zoo. I was walking with Lewontin through the zoo and, in light of his lecture, asked him who wrote this: “Animals construct nature for themselves according to their special needs.” Without breaking his stride, Dick said it was probably Karl Marx. In reality it was written by the politically conservative Jakob von Uexküll in his seminal book, Umwelt and Innenwelt der Tiere (Environment and Inner World of Animals) in 19091. Uexküll had an enormous influence on the early ethologists including Lorenz, Tinbergen, Baerends, and others. So the idea that organisms construct their worlds both perceptually and behaviorally is actually quite old, and beautifully illustrated in the book von Uexküll published with Kriszat in 1934. In a recent new translation of this and other writings with commentaries by Dorian Sagan and Geoffrey Winthrop-Young, the final paragraph claims that von Uexküll can be viewed as “a precursor of post-Darwinian evolutionary theory.”
I also mentioned earlier the “Baldwin effect,” proposed almost simultaneously in the late 1890s by James Mark Baldwin, Lloyd Morgan, and Henry Fairfield Osborn; then called organic selection. This was even earlier than von Uexküll published. Baldwin was particularly interested in behavioral development, social evolution, and how plasticity and variability may more rapidly fashion new behaviors in novel environments than ordinary natural selection and yet did not invoke Lamarckian processes. Waddington’s genetic assimilation (1950s) is a related process and West Eberhard (2003) covered these topics, though not von Uexküll. What is your opinion of these early contributors? Why were their ideas so ignored in mainline evolutionary biology? Is the major recent advance with niche construction theory the addition of a program of research incorporating selection to the mix, new data and experiments, or models that go beyond the now thriving Baldwin effect enterprise? I know you have published on some models yourself.
Kevin: Thank you. I didn’t know that von Uexküll had written on this topic, but I am not surprised. The fact that niche construction occurs is, of course, immediately apparent, and most people recognize this. What is less obvious is that niche construction has important evolutionary consequences, but here too there have been pioneering thinkers that have explored this issue. You can think of the so-called ‘Baldwin effect’ (I say so called because Baldwin was neither the sole, nor the first, person to propose it) and genetic assimilation as special cases of evolution by niche construction. Niche construction theory is much broader than these ideas: for instance, organisms can modify selection through means other than learning, and generate selective feedback on loci other than those underlying the niche-constructing activity. This leaves it potentially more important.
The issue as to why these ideas have not yet become mainstream is complex and multifaceted. In part, it relates to a failure of those championing them to translate them into an empirical program, but there are some fundamental conceptual barriers at play too. What plasticity-led evolution, niche construction theory, and these more historical ideas have in common is that they run up against long-established ideas about how biological causation works (for instance, Mayr’s separation of proximate and ultimate causation and his corollary that proximate causes are irrelevant to evolution). They equally challenge the widespread but erroneous conception that there is a ‘genetic program’ that contains ‘instructions’ for building phenotypes, which implies that phenotypic outputs have been prescreened by earlier natural selection, and hence cannot be novel in any meaningful sense. My hope is that the extended evolutionary synthesis will prove of value by drawing attention to the fact that these issues are all connected.
Gordon: Thanks for making these points. Before going on I should just mention that an early pioneer of animal behavior, Douglas Spalding, actually came up with the essence of the “Baldwin effect” regarding the transmutation of a learned behavior to an innate instinct in an article published in 1873! JBS Haldane had this paper republished in an early volume of what is now the journal Animal Behaviour in 1952; in his introduction he made explicit comparisons with Waddington and pointed out the excellence of Spalding’s proposed experimental design. But back to niche construction theory; my major professor, Eckhard Hess, often stated in his classes in the early 1960s (when I took them) that humans were the only self-domesticated species. We modified the environments in which we lived as well as the operation of our own bodies through such devices as clothes, controlled fire, eyeglasses, medical science, machinery, etc., that altered survival prospects through a type of artificial selection. These had effects well beyond learning, including our physical features, that now feedback to environmental plasticity and culture.
Today we realize that other species, such as social insects, do this also. In a forthcoming paper2 you and your co-authors explicitly place niche construction theory as an intermediate process between natural selection and domestication. Could you explain how you came to this connection and how and why it is part of the extended synthesis? Also, if artificial selection was something recognized by all evolutionists back to Darwin, why was it and niche construction theory not recognized as a distinctive mode of evolution that also occurred in the natural world along with natural selection? Was it due to the focus on the human species as the agent of selection? Is niche construction theory really a form of self-domestication in which organisms are active players?
Kevin: Let me deal with the last question first. If you are asking whether humans have created circumstances that generated natural selection favoring prosocial tendencies, I am persuaded by the evidence for gene-culture coevolution shaping many aspects of the human mind, including cooperation. However, I am not convinced that “self-domestication” is the right term here. The same gene-culture coevolutionary dynamic has favored war and aggression in other circumstances, so the parallels with domesticated animals seem somewhat superficial.
The issue as to whether niche construction in some respects resembles artificial selection is, to my mind, more interesting, and connects much more clearly with our research project. During artificial selection, breeders and experimentalists deliberately select for particular characteristics, such as high yields or attractive plumage. The breeder/experimentalist is imposing a direction on the selection that ensues through consistent and sustained activities that determine and control the fitness of individuals in the selected population. The predictability and generality of artificial selection can be contrasted with the frequent unpredictability and local contingency of natural selection in natural populations. It is probably a reasonable generalization to suggest that responses to natural selection in the wild are typically weaker, less consistent, and less predictable than responses to artificial selection.
In our forthcoming article2, John Odling-Smee, John Endler and I suggest that niche construction occupies the middle ground between artificial and natural selection. Like artificial selection, the selection that niche construction generates will often be reliable, directed, orderly and highly consistent across diverse organisms performing similar niche-constructing activities. This ought to mean that selection arising as a result of niche construction will be more predictable than responses to environmental elements with little or no niche construction.
The motivating idea behind this (and indeed, behind niche construction theory more generally) is that niche construction imposes a bias on evolution through systematically modifying environments in a directional manner. If that is correct, constructed features of the environment should be qualitatively different from non-constructed features, and responses to selection stemming from the former should be more predictable than those stemming from the latter. Our project puts this to the test.
Gordon: I understand the comparison with artificial selection where humans domesticated species in what seems to be a goal directed manner. The development of quantitative genetics was spurred by the goal of faster growth, meat quality, disease resistance, and many other targets. By self-domestication I was not just referring to prosocial behavior but all the kinds of changes, genetic and cultural, that humans have engaged in to change many aspects of our species, or subpopulations, anatomy, physiology, behavior, and cognitive capacities through modifying the environment or engaging with it differently than before. Of course, unplanned and unexpected things happen, as in the increased prevalence of myopia, obesity, and allergies in children, but niche construction does seems to be a process whereby organisms take some control over their future evolution. What you also are saying is that niche construction “will often be reliable, directed, orderly and highly consistent across diverse organisms performing similar niche-constructing activities.” I think that this is an exciting claim; how are you planning to test this in your project?
Kevin: That’s right. We plan to test it in two ways. First, living organisms control aspects of their environment, often pushing environmental factors into states that they would not otherwise occupy in an entirely abiotic world. In this manner, they impose a structure on some of the selection that they experience, and should reliably trigger adaptive responses, or buffer such responses, depending on circumstances. In principle, this predictability should be detectable in responses to selection. We expect that niche construction will typically generate consistent selection, both across generations (i.e. sustained over time) and in space (i.e. the same, or related, species should construct consistent niches over most of their range), which will be manifest as reduced temporal and spatial variance in selection differentials and gradients relative to non-constructed environments. Other predictions concern evolutionary rates. For instance, we expect that environment buffering (counteractive) niche construction will typically reduce the rate of response to selection, as manifest in smaller linear selection gradients (a way of measuring selection that computes the effect of a trait on fitness holding the effects of other traits constant). Conversely, novel (inceptive) acts of niche construction will on average generate unusually strong selection, at least initially. We are currently compiling datasets that measure evolutionary responses to constructed and non-constructed sources of selection, and will be testing our predictions using meta-analytical tools.
The second approach deploys comparative phylogenetic statistical methods to explore trait coevolution. For illustration, when an animal builds a nest, it immediately creates or modifies selection for the nest to be defended, maintained, regulated, and improved upon in design, and for other animals to steal the nest material, destroy it, squat in it, or dump eggs in it. These are robust selective responses that can be anticipated irrespective of whether the builder is a bird, a fish, a wasp or a cockroach. We anticipate that it should be possible to predict sequences of trait evolution and trait coevolution across multiple traits in instances where these result from niche construction, and that this predictability will be enhanced relative to autonomous features of the environment. We also expect that niche construction will generate parallel patterns in selective responses amongst independent lineages. These predictions can be tested using comparative phylogenetic methods applied to animal artefact construction and associated behavior. Again the first stage is data compilation – we are currently building detailed databases about birds’ nests, spiders’ webs, and so forth. We will then go on to make predictions about the evolution of specific sequences of traits in response to nest or web building (i.e. spider web building led to the evolution of refined or larger web structure, which in turn favored subsociality, etc.). We expect that researchers will be able to predict more extended evolutionary sequences that follow niche-constructing evolutionary events compared to other evolutionary events, and that similar niche construction will explain some cases of parallel evolution (for instance, burrowing in both spiders and mammals should trigger the evolution of resistance to fungal diseases).
Gordon: Thanks for this set of fascinating examples and associated predictions and methodological approaches. In discussing niche construction with colleagues I find that some of the most enthusiastic scientists are those who work with plants. Joe Bailey and his lab, for example, work on trees and these can respond differently, over their geographical range, to different nutrients, such as nitrogen, in the soil, with diverse genotype – environment interactions involved, including with other plants. Such interactions also alter the soils and the survival and growth of both different genotypes and other species. His lab has documented, through soil exchange experiments, some fascinating results. Are plant-plant and plant-soil systems also good model systems and might they even provide more amenable systems than animals? Are you also including plants as well as animals and artefact construction and behavior in your studies on niche construction?
Kevin: That’s an excellent point. Both John Odling-Smee and I have backgrounds in the study of animal behavior, so in the past we have tended to lean perhaps overly heavily on animal artefact examples. However, niche construction is a general property of all living organisms. Plants are orienting to light, creating shade, influencing wind speed, altering hydrological and nutrient cycles, producing root exudates that change soils, and modifying their environments in a myriad of other ways. Recently, Sonia Sultan has written a terrific book (Organism & Environment: Ecological Development, Niche Construction and Adaptation) which covers these many botanic examples in wonderful detail, and shows how plant niche construction is both ecologically and evolutionarily important. You are right too to suggest that there are often practical advantages to working with plants and soil systems. The analyses that we are doing are completely reliant on the available data, and because many of us are fascinated by animal behavior, there is good data available on birds’ nests and spiders’ webs, so it makes sense to exploit this in our comparative phylogenetic work. However, our analyses of selection gradients include quite a lot of plant examples.
Gordon: Well, I think we have covered lots of ground in how niche construction can be an important factor in evolutionary change. Perhaps we could end by addressing some of the criticisms of niche construction separate from the controversies about the extended evolutionary synthesis in general. Some critics have claimed that there is nothing in niche construction that cannot be incorporated into the traditional neo-Darwinian approach. I do not agree; it has allowed us to look at, and study, phenomena that the traditional view ignored. I am reminded of the criticisms of cognitive ethology by behaviorists and current arguments pitting association learning and cognitive reasoning explanations of problem solving and other complex abilities in animals against each other. These abilities would have never even been considered interesting or important scientific fodder by traditional behaviorists; and while sometimes the dueling explanations are unresolvable with our current knowledge, taking them seriously does in itself advance science. Can you speak to some of the specific criticisms that have and are appearing in the literature on niche construction that you think are most relevant and if, how, and when they will be resolved? Or is it more that the research approach and questions asked by niche construction researchers will be the most enduring legacy? Please expand as much as you want and have the final word!
Kevin: Thank you. I think there are both interesting and tedious reasons for the contention over niche construction. The latter largely come down to misunderstandings. For instance, people have attributed to niche construction theory the view that niche construction is a neglected topic within evolutionary biology, and then taken issue with that assertion. In fact, niche construction theory makes no such claim, as a careful reading of our work would make clear. Studies of niche construction go back at least to William King’s 17th century work on ecological succession, and include classic nineteenth century works on beavers by Lewis Henry Morgan and on earthworms by Charles Darwin. Aspects of niche construction are captured by much existing ecological and evolutionary theory, for instance, models of resource depletion, habitat selection and frequency dependent selection. We review this material in depth in our monograph. Our argument is that, whereas niche construction has been widely studied as a phenomenon, it has not been widely recognized as an evolutionary process (i.e. niche construction is not formally listed as a cause of evolution in virtually all evolutionary literature). Such misunderstandings are frustrating, but I won’t dwell on them here.
More interesting is contention that reveals differences amongst researchers in their approach to science. Niche construction theory, with its developmental perspective and its emphasis on organismal agency, represents another way to think about adaptation. I believe that pluralism of perspective is critical to science, and that alternative viewpoints can be of great value, provided they encourage the generating and testing of novel hypotheses, or open up new lines of enquiry. Our early work on niche construction theory was based in Richard Lewontin’s laboratory at Harvard, and maybe I picked up something of the dialectical method from him. Niche construction theory has promoted empirical investigations of niche construction, ecological inheritance, and eco-evolutionary dynamics, it has encouraged mathematical analyses of these phenomena, and been applied to understand a huge range of topics, from the oxygenation of the oceans, to the origins of language. For me, that is enough to justify our currently heterodox stance. I believe scientific theories should be judged on the basis of the research that they stimulate, and hundreds, perhaps thousands, of researchers in diverse disciplines are finding niche construction theory useful.
However, clearly many evolutionary biologists think differently. They see science as less to do with utility, and more to do with a search for the truth. They are convinced that their understanding of evolution is correct (or at least, more correct than Lewontin’s), and consequently they view niche construction theory not as a valuable alternative way of thinking, but as muddle, or distortion. What attracts people to niche construction theory is primarily Lewontin’s emphasis on the active agency of the organism, which resonates with their own assessments of the role that agents play in their focal science. However, for many traditionalists any such emphasis is not only deeply misguided, but also unnecessary as, in their view, the phenomenon of niche construction can be understood perfectly well when stripped of these connotations.
Perhaps the major source of contention, however, surrounds the claim that niche construction operates as an evolutionary bias. This argument goes back to the writings of Erwin Schrödinger, the Nobel Prize winning physicist. Schrödinger noted that living organisms are out of thermodynamic equilibrium, and that in order to preserve their out-of-equilibrium status, and create order in their bodies and in their immediate surrounds, organisms must do work on their environment. He specified that this work cannot be random; it must be directional. It follows that organisms must change environments in systematic and directional ways. That is why we believe niche construction must impose a directional bias on selection. However, this means that the case for niche construction being an evolutionary process currently rests on a line of reasoning, rather than clear experimental or mathematical proof. I think that is the main reason why it is controversial. Niche construction theory is likely to remain contentious within the evolutionary mainstream until such a time as it is clearly demonstrated to operate as a bias, through solid empirical or theoretical work. A theoretical demonstration is plausible, but for the moment, our focus is empirical. With luck this research project may provide some such evidence.
1. Partial translation including the quotation in Burghardt GM. Ed. (1985). Foundations of Comparative Ethology. Van Nostrand Reinhold: New York. Umwelt, after von Uexküll, now often used as the environment as perceived by the organism or ‘personal world.’
2. Laland KN, Odling-Smee FJ, & Endler J. (Forthcoming) Niche construction, sources of selection and trait coevolution. Interface Focus.
This post was originally published in This View of Life (TVOL), 19 June 2017.