Abstract
This paper argues that the contextual approach to natural selection does not offer an estimation of the contributions of individual and group selection to evolutionary change in multi-level selection scenarios, and that this is so because the term “group selection”, as defined by the contextual approach, does not refer to a process taking place at the group level. In the contextual analysis framework, this term simply denotes an evolutionary change that takes place due to the fact that, overall, individual types do not share similar contexts or environments, and the only way to claim that such an evolutionary change is a result of selection is by admitting that “group selection” is in fact a kind of frequency-dependent selection, i.e. a selection process taking place at the individual level. Therefore, under the names “individual selection” and “group selection”, the contextual approach actually isolates two aspects of the relation between individual types and their environment, and not two distinct levels of selection.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
An argument against the distinction between MLS1 and MLS2 cases has been advanced in (Goodnight 2013). However, since the philosophers whose positions I take issue with here assume this distinction, I will do so as well.
Some of my statements below will be tributary to the choice of these values for the individual character, and would have to be slightly modified in order to accommodate other values. However, that does not affect at all the logic of my argument, which concerns the contextual approach in general.
For simplicity, I assume that parents die immediately after reproduction and the offspring are identical to their parents with respect to the focal trait.
This doesn’t mean that deviations from Condition 3 are confined to empirical cases. Ideal or hypothetical cases will also yield error terms when the fitness functions are not pre-designed—as is the case for my Case 1—to perfectly fit into Eq. 1.
Actually, his example is directed towards a variation of the contextual approach called “the neighbor approach” (see Okasha 2006), but, by Glymour’s own admission, “with suitable minor changes in the arguments” his argument holds for the contextual approach in general (Glymour 2008, p. 840). Therefore, I will treat loosely his argument here as if it were talking about the contextual approach in general. This will not affect the meaning of Glymour’s argument, since the numbers themselves are far less important here than the principle.
Here is a quote that clearly indicates this: “Given the presupposition that group selection acts whenever belonging to properties causally influence reproductive success [i.e. whenever there are group effects on individual fitnesses], random group formation (and hence sample covariances that lie close to zero) does not preclude group selection from acting, and does not prevent group selection from influencing the rate at which the population phenotype changes” (Glymour 2008, p. 844).
One should also bear in mind at this point that in such cases the average individual fitness \( \bar{w} \) within the global population remains constant regardless of the fact that the population is divided into equal groups or not.
Of course, as biology is the science of exceptions, there is an exception to this rule, and we will discuss it at the very end of this paper.
In my Case 1, this implicit assumption is embedded in the fitness functions of the two types.
In this particular case, this means that the individuals in the two groups are likely to encounter the same type of predators, and the frequency of these encounters is also similar.
Again, it is with the help of the version of the contextual analysis known as “the neighbor approach” that Nunney’s point can be formalized. But since the two approaches are operationally identical (only differing by the choice of variables), they share the same strengths and weaknesses with respect to the principles of their functioning.
Brandon goes on to add “Indeed it is a type of drift”, but I will not go into the details of this here.
It is surprising that philosophers discussing the contextual approach can equate the effects on fitness of the group character and the effects on fitness of “group membership” (Okasha 2006, p. 88) and still claim that “group selection” is involved in the contextual approach. This is tantamount to equating “selection on group membership” with group selection; but, obviously, what we would call “selection on group membership” means nothing more than the fact that two individual types are overall subjected to different environments, and this seems to eliminate both the groups and the “selection” from the entire discussion. Critiques of the contextual approach fall prey to the same imprecision, as is obvious when Glymour (2008, p. 836) treats the group character as being a “belonging to” variable, and yet continues to speak of group selection.
The Price approach has its own serious conceptual problems (and not just problems concerning precision or reliability), but that will not be discussed here (see Jeler 2013).
References
Brandon, R. (2013). Natural selection. In E. N. Zalta (Ed.), The Stanford Enciclopedia of Philsophy (Winter 2013 Edition), http://plato.stanford.edu/archives/win2013/entries/natural-selection/.
Darwin, C. (1859). The Origin of Species. London: John Murray.
Glymour, B. (2008). Correlated interaction and group selection. British Journal for the Philosophy of Science, 59, 835–855.
Goodnight, C. J. (2005). Multilevel selection: The evolution of cooperation in non-kin groups. Population Ecology, 47, 3–12.
Goodnight, C. J. (2013). Defining the individual. In F. Bouchard & P. Huneman (Eds.), From groups to individuals (pp. 37–54). Cambridge, MA: MIT Press Books.
Goodnight, C. J., Schwartz, J. M., & Stevens, L. (1992). Contextual analysis of models of group selection, soft selection, hard selection and the evolution of altruism. American Naturalist, 140, 743–761.
Heisler, I. L., & Damuth, J. (1987). A method for analyzing selection in hierarchically structured populations. American Naturalist, 130, 582–602.
Jeler, C. (2013). Do Elliott Sober’s arguments for group selection really account for the causal nature of natural selection?, Filozofia. Journal for Philosophy, 68, 319–331.
Nunney, L. (1985). Group selection, altruism, and structured-deme models. American Naturalist, 126, 212–230.
Okasha, S. (2004a). Multi-Level selection, covariance and contextual analysis. British Journal for the Philosophy of Science, 55, 481–504.
Okasha, S. (2004b). Multilevel selection and the partitioning of covariance: A comparison of three approaches. Evolution, 58, 486–494.
Okasha, S. (2005). Altruism, group selection and correlated interactions. British Journal for the Philosophy of Science, 56, 703–725.
Okasha, S. (2006). Evolution and the levels of selection. Oxford: Clarendon Press.
Okasha, S. (2011). Reply to Sober and Waters. Philosophy and Phenomenological Research, 82, 241–248.
Price, G. (1972). Extension of covariance selection mathematics. Annals of Human Genetics, 35, 697–701.
Acknowledgments
I would like to thank an anonymous reviewer and Charles Goodnight for their remarks on an earlier draft of this paper that have seriously helped me clarify my point. I would also like to thank Samir Okasha for his comments on an even earlier draft.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Jeler, C. Is there such a thing as “group selection” in the contextual analysis framework?. HPLS 36, 484–502 (2015). https://doi.org/10.1007/s40656-014-0044-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40656-014-0044-8