Natural Selection at all scales?

Do ant colonies form a kind of “super-organism”? Is that how we should think about what they are doing? If we were all bacterium, or were the size of bacterium, and looked at a human being, wouldn’t we conclude that a human was itself a “super-organism”?

The discrete species that we identify are particular to the scales in space and time that we are accustomed to. Nature itself does not recognize, and behaves without regard to individual species. Natural selection may occur at all scales.


  • Across deep time, the boundary between one kind of organism and another washes away in the history of continuous change.

When in the fossil record does Homo Sapiens begin and its progenitors end? Anthropologists acknowledge that the distinctions we make between Homo Erectus, Neanderthalensis, Habilis, etc. are made largely for convenience. It helps organize the record and discretize the concepts involved. This allows us to create narratives of evolutionary arcs when we think may have understood enough to create a new “chapter” in the story. There are always arguments going around about how to draw the divisions; whether to use the population’s behavior or some element of morphology. That the community can form a consensus about how organize the fossil record does not imply that the record contains any of that organization innately (1).

  • Understanding of the evolution of cooperation among organisms has been advancing for a number of years, and requires that systems be viewed at multiple scales of organization.

Which is the organism, and which is the super-organism? Not just in the case of ants and other highly organized insects, and not just in the case of human constituted super-organisms like cities and nation states, but I also mean to invoke this description for the evolution of multicellularity and every other conceivable case:

Imagine that there exists a population of single-cell bacterium, each only very slightly varied from the other. Now we have to assume a few things: First, that there is some impetus for resource competition. Second, that there is some mechanism that can break the symmetry across the system, i.e. a mechanism to make the population inhomogeneous and allow the formation of competing groups (2).

  • Multiple substrates for the storing of information, each with its proper temporal or spatial scale can be identified.

If there is no particular scale for selection, it follows that there exists no particular substrate for the information being edited by the process – it cannot just be DNA. Recently epigenetic effects have been described which can alter gene expression across two and three generational spans, or even within the lifespan of a single person. They may occur in a person’s body as a result of, for example, changes in diet.

In another example, take the case of behaviors that are transferred within a community: mimetic changes of the kind described by Dawkins (3). The community as a whole has behaviors and other cultural constructions encoded in the traditions being passed between people. At a particular timescale, as long as there exists a mechanism whereby one community with a set of behaviors reproduces or survives in greater numbers than a community with different behaviors, a preference will inevitably be expressed.

It is said that some Native American cultures planned for the next  seven generations. Typically in teaching evolution, the focus will be on the individual and its own success at survival and reproduction, as expected from our limited personal experiences, but instead imagine that there is a set of behaviors that, instead of producing more offspring in only the subsequent generation, somehow produce more offspring not until the second or third generation (4). At that longer timescale, isn’t it the behavior with the lower single generation success rate preferred? Is there any limit to how far forward this can reach? It has occurred to me that it may be very difficult to find a mechanism for selection at very long timescales, though – it seems that the substrate would have to change.

The ways that organisms, superorgansims, and communities are modifying their “genetic” or genetic-analogous information can become very difficult to describe. I think in more general terms it would require a kind of thermodynamics which I’m not even sure exists yet. A kind of thermodynamics that can describe the changing organization (entropy) of a system along with the appearance of new levels of organization (critical states and phase transitions). I can’t go past this point, so I will just recap and stop:

  • Natural selection as it is recognized today is a special case of more general phenomenon that can be described in physical/thermodynamic or information theoretic terms.
  • There is no proper space or time scale for selective processes.
  • As long as all the necessary ingredients for evolution to occur are present at any described scale, it will occur (5).

  1. I suspect it will be so that some changes occur in evolution very quickly indeed, almost discontinuously when taken in the usual deep time scales that we commonly associate with natural selection. I was taught that it takes millions of years for changes in a species to be marked, but I suspect more recent biological research has complicated this. It may be possible for species to change on scales of thousands or even hundreds of years under certain  circumstances. I am not well educated on the particular instances of this kind of event.
  2. This part I simply don’t understand. In its full generality, a theory describing this would not even be recognizably biological – it would be in the language of thermodynamics and phase transitions, as what I am trying to describe here I think is fundamentally a kind of non-equilibrium phase separation. Non-equilibrium because the population is a living system that takes in energy at some rate, and here by phase separation I mean something like oil and water separating due to the energies of interaction between molecules and the entropies of the separated and mixed states of the system as a whole. The particular mechanisms that might produce the inhomogeneity can be incredibly subtle too – I think of it in terms of selection trying to get ‘traction,’ and depending on how well discretized, and how strongly competitive the groups are at the scale of interest there may be very little selection or very clear, well defined natural selective mechanisms; the selective process itself has to emerge in some subtle way from the system’s changing dynamics as the groups are given more definition.
  3. I haven’t read the Selfish Gene. I can only assume his description of mimetics is very similar to this, though probably without my attempt at full generalization.
  4. I’m not completely familiar with exactly how the Monarch butterflies manage their migration. It is my understanding that the entire round trip takes five or more generations.
  5. Which are really necessary!? I don’t know. If I were to build a computational model, how simple could I make it, and still have a model that reproduces evolutionary processes? Is there a simpler one than Conway’s Life?

One thought on “Natural Selection at all scales?

Comments are closed.