An ecological anthropic principle?

Ecological and evolutionary systems–populations, communities, ecosystems, species, food webs, whatever–only occupy certain regions of parameter space. That is, the systems we observe are only a small subset of those which could possibly exist. Individuals organisms may live for hours or years or even a century or more, but they don’t live for millenia. There are tens of millions of species, not dozens or trillions, and those species typically persist for millions of years, not a century or a million centuries. Food webs have connectances ranging from about 0.03-0.3, not 0.001 or 0.99. Most populations exhibit negative density dependence, but less than 30% of populations exhibit regular cycles. And so on.

So why do ecological and evolutionary systems occupy the regions of parameter space they do? This isn’t a question that often gets asked (but see Joe Felsenstein’s classic article on why there are so few kinds of species, compared to what we might have expected). For instance, in my own field of community ecology, analysis of theoretical models is typically concerned with identifying the regions of parameter space that allow stable coexistence. But why should nature occupy any particular region of parameter space? After all, nature may abhor a vacuum, but why should it abhor lack of stable coexistence, or a food web with connectance >0.3, or whatever?

I can think of several possible answers:

  1. No reason. It just so happens that nature occupies certain regions of parameter space. If there are any reasons why particular systems occupy particular regions of parameter space, those reasons are idiosyncratic and system-specific.
  2. Fundamental physical constraints. We only observe certain kinds of systems because those are the only ones consistent with the constraints imposed by the laws of fundamental physics.
  3. Historical constraints. We only observe certain kinds of systems because those are the only ones consistent with the constraints imposed by past historical events.
  4. Natural selection. We only observe certain kinds of systems because natural selection works; others would only exist if species were highly maladapted.
  5. The anthropic principle. Systems that occupied other regions of parameter space wouldn’t exist long enough to be observed, or else wouldn’t be consistent with human life (so that we wouldn’t be around to observe them).

#1 surely has something to it. After all, the regions of parameter space occupied by ecological systems clearly can change (e.g., due to things like climate change, anthropogenic impacts, and mass extinction events), so perhaps there’s nothing special about the regions of parameter space nature happens to occupy at the moment. But I doubt that’s the whole story. There’s never been, and never will be, a time when species lifetimes were measured in centuries, or when food webs had connectances of 0.99.

At the risk of offending any biomechanicists reading this, I suspect #2 is largely irrelevant. I don’t think very many features of ecological and evolutionary systems reflect constraints imposed by the laws of fundamental physics. What organism would move faster, except for the constraint imposed by the speed of light? What does fundamental physics have to do with the strength of density dependence? Etc. And less-fundamental physical constraints, while they absolutely exist, have a way of turning out not to be constraints at all over sufficiently long time scales. For instance, the evolution of multicellularity freed multicellular organisms from many of the physical constraints (e.g., diffusion-imposed constraints on body size) to which single cells are subject.

#3 is sort of a special case of #1. For instance, all organisms use nucleic acids as their genetic material. That’s a ‘frozen accident’ of history which probably isn’t going to change in future, and which may impose some sort of constraint on what’s evolutionarily possible. Or maybe not. As Francis Crick once said, “Evolution is cleverer than you are.” In general, whenever we claim that X is some kind of absolute constraint on evolution, we have be very careful that we’re not merely expressing the limits of our own imaginations.

I think #4 is a big one, not just in evolution (obviously!), but also in ecology. In the context of community ecology, I’ve lately been wondering if there might not be fairly general reasons to expect natural selection to promote ecological coexistence. That is, if you incorporate adaptive evolution into models of ecological coexistence, so that the model parameters can (co)evolve, are there reasons why you are likely to end up in a region of parameter space permitting coexistence?  These reasons, if they exist, would of course have exceptions. After all, natural selection doesn’t actually ‘care’ about coexistence, it only ‘cares’ about relative fitness. Any coexistence you get due to selection is just a side effect from an evolutionary perspective. But perhaps it’s a particularly likely side effect. With the recent explosion of interest in ‘eco-evolutionary dynamics’, #4 is likely to be thoroughly explored in the near future.

#5 is the inspiration for the title of this post. In physics, the anthropic principle is the idea that observations of the universe must be compatible with the conscious life that observes it. No less than Alfred Russel Wallace prefigured a (strong) version of the anthropic principle in 1904 when he wrote,

“Such a vast and complex universe as that which we know exists around us, may have been absolutely required … in order to produce a world that should be precisely adapted in every detail for the orderly development of life culminating in man.”

By the ecological anthropic principle, I don’t just have in mind the idea that we wouldn’t be here to observe nature unless nature were a certain way. I’m also (actually, mostly) thinking of the fact that we can only observe those bits of nature that are sufficiently stable and persistent (so maybe a broader term like ‘selection bias’ would actually be better). It’s like the old joke in economics, about why you never find $10 bills on the ground (if they were there, someone would’ve picked them up by now). For instance, many food web theoreticians have asked whether observed topological features of natural food webs are such as to promote their stability, implicitly arguing that unstable webs should not persist long enough to be observed. Some population ecologists have made a similar argument to explain why most populations exhibit at least weak density dependence (if they didn’t, they’d be expected to go extinct much faster than they do). But I don’t think the ‘ecological anthropic principle’ has been systematically explored as an explanation for why nature is the way it is, rather than some other way. How strong a constraint does it impose on states of nature? What happens if it opposes natural selection (for instance, Laurence Mueller‘s studies of Drosophila find no tendency for selection to favor stable population dynamics)?

The take-home message here is that ecologists shouldn’t just worry about explaining how the world is. If we want an ultimate explanation for the way the world is, we should probably pay more attention to the way it isn’t.

6 thoughts on “An ecological anthropic principle?

  1. Pingback: ESA 2012 preview: all the good talks are on Friday | Dynamic Ecology

  2. Fun post, but although their form is very similar the physical and ecological anthropic principles feel completely different. The ecological AP, and especially the selection bias variant you discuss at the end, inherently makes human ‘not-special”. For me, it demolishes the common misconception of an evolutionary ladder, and explains that “we think of ourselves as special just because we are us”. The physical AP, though, makes me feel the opposite. The PAP is undoing the copernican revolution and putting us back as the center of the universe: the whole thing is being justified by our existence.

    I am not sure why these two principles cause such a drastically different reaction for me. The only source I can think of is my lack of imagination: I feel like I can imagine and model ecological systems that can’t support human life, I even feel like I might be able to observe them in a laboratory (I use I in a vague sense here, since I would have no idea what to do in a laboratory). On the other hand, I can’t really imagine different universes, I can sit down and calculate the standard model with different parameters, but I have absolutely no intuition pumps for how to think of most of the resulting dynamics. But I am not sure if this is what is causing my uneasiness, I will have to reflect on it at a more reasonable hour of the day.

    Your selection bias principle is awesome, though. I wonder if it’s been applied to the evolution of cooperation. Almost every paper in the field has the same intro: natural selection is tooth-and-claw and yet we see cooperation all around us. Well, we see it because species that don’t cooperate go extinct faster — an easy claim to justify with standard group-selection arguments, and on the level of species, group-selection doesn’t seem that silly. I think it would be fun to combine predator-prey (or some ecological niche competition?), sympatric speciation, and EGT with group-selection into one model that shows that species that are bad at cooperating go extinct faster. So, if you’re taking random time-slices, cooperation will be over-sampled, even more so if you also try to work in some historicity assumptions by sampling only when you have an ancestory of cooperation in the sampler’s.past.

  3. Pingback: Are unstable communities like $10 bills lying on the ground? | Dynamic Ecology

  4. Not a biomechanist, but taking offense. 🙂 (joking, of course)

    You say that #2 is largely irrelevant for determining why eco/evo communities occupy only a particular parameter space. But if we consider supervenience and that the properties of “higher” levels such as ecological communities supervene on the property of “lower levels” such as individuals, and keep working all the way down, we’d arrive at the superveneince of properties of eco/evo communities on physical laws. Now, maybe with so many layers between these two levels, understanding the physical properties of matter will not help you understand ecological communities in any structured, significant way and might probably be a futile exercise. But perhaps understanding the properties of a layer or two above the physical one might?

    • Hmm, not sure I quite see why supervenience of biology on physics undermines my #2.

      I admire evolutionary biomechanics, I think it’s a fascinating subfield. I agree that there are interesting ecological facts that can be explained by appeal to biomechanical constraints organisms face. But even in evolutionary biomechanics, organisms rarely are distributed uniformly within the space of what’s physically possible. So even in evolutionary biomechanics, appeals to physical constraints on what’s possible are rarely the whole story. Rather, interest often centers on why we observe some physically-possible organisms and not others.

      • I guess what I was saying was stretching it a bit too far, but I’ll try to explain better:

        I would argue that because the properties of ecological/evolutionary systems, and of biological systems more generally, supervene on those of physical systems, the answers to why ‘organisms rarely are distributed uniformly within the space of what’s physically possible’ and to ‘why we observe some physically-possible organisms and not others’ also “ultimately” depend upon the properties of physical systems. Even if these answers have to connect to ecology and evolution, that itself again supervenes on physics because all processes of evolution and ecology can be captured by literal physical processes. Another way to phrase this is that they are not truly physically possible, because everything that is physically possible exists. But this is true only in the extremely abstract sense. To actually attempt this would be totally futile, because we won’t come out with any significantly structured understanding of eco/evo.

        At the same time, going down to “lower” levels, maybe not such a big leap as to physics but to some level closer such as biochemistry, might explain why certain parameter spaces in ecological systems are not explored, because the biochemical specifics of the system don’t allow it, for example.

Leave a Comment

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

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

Google photo

You are commenting using your Google 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

This site uses Akismet to reduce spam. Learn how your comment data is processed.