Commentary on Hull

Abstract: 58 words
Main Text: 1000 words
References: 75 words
Total Text: 1180 words


A more pluralist typology of selection processes


Bence Nanay

Department of Philosophy

University of California, Berkeley

Berkeley, CA. 94720



Instead of using only one notion of selection I argue for a broader typology of different types of selection. Three such types are differentiated, namely simple one-step selection, iterated one-step selection and multi-step selection. It is argued that this more general and more inclusive typology might face more effectively the possible challenges of a general account of selection.


As evolutionary approach is getting more and more popular in a large number of different disciplines, there is a strong need for a typology that would show what is similar and what is different in all these evolutionary models. The endeavor of the target article is of key importance in this respect, since it outlines a conceptual framework in which different selection processes of different research fields could be compared.

The authors of the article restrict themselves to three such phenomena though. Besides, they draw a strict division line between them and the rest of evolutionary processes, claiming that only these three processes count as selection processes. Even if they do not exclude the possibility of this list becoming longer in the future, they do not discuss any phenomena not fulfilling the criteria for being a selection process.

To put it differently, the target article focuses on what selection is, whereas I would like to address the differences between various forms of selection. According to the article a phenomenon is either selection or it is not. My question is about what type of selection it is. Selection is not a ‘yes or no’ phenomenon; it has different degrees, different types.

I suspect the general line of critical responses to the target article will consist of arguments why one of the three processes does not fit the general definition of selection provided by the authors. The more pluralistic typology I tend to favour would offer a way of defense against these objections. Even if one of the processes covered by the authors turns out not to fulfil all the criteria of selection as defined in the article, it might fulfil a weaker criterion of selection.

At this point three types of selection processes has to be differentiated. I use the term multi-step selection as a synonym of what the target article considers as selection proper, defining it as "repeated cycles of replication and environmental interaction so structured that environmental interaction causes replication to be differential". Besides multi-step selection, two further selection processes are included in the typology.

I call iterated one-step selection the process whereby there are repeated cycles of replication and environmental interaction, but where environmental interaction does not cause replication to be differential. In other words, environmental interaction does not influence the next replication. In the case of simple one-step selection, replication is followed by environmental interaction, but no new replication occurs after that: this is the end of the story.

The differences of these three processes are summarized bellow:

An important example of simple one-step selection is the development of the central nervous system (Changeux 1985, Edelman 1987, Adam 1998). From the initial variety of neural connections some are selected by environmental interaction, whereas other connections disappear during the development. No further replication takes place after the environmental interaction, since no (or very few) new neural connections are formed later.

It is more difficult to find a clear example of iterated one-step selection. Nevertheless, the otherwise highly controversial explanatory model of evolutionary epistemology seems to imply a version of such selection process, which can be examined without committing to the credibility of evolutionary epistemology as such (Popper 1972, Campbell 1974). According to this theory, all thinking processes can be characterized by repeated cycles of blind variation and selective retention. A variety of thoughts is produced continuously and blindly, but environmental interactions decide which thought will survive. These environmental interactions, however, do not have any impact on the next variation of thoughts (this characteristic of variation is dubbed as "blind" by Campbell).

Single one-step selection plays a central role in an earlier typology of selection processes outlined by Darden and Cain (1989). According to them, the criterion of selection is that replication is followed by environmental interaction, but this cycle need not be repeated. One-step selection is not only included in their typology, but it is also conceived as the paradigmatic case of all selection processes. The target article argues powerfully against the central role of single one-step processes in describing selection, but these arguments do not support their conclusion that "single cycle selection is (at most) a limiting case" of an account of selection processes. In my view, both one-step selection à la Darden and Cain and multi-step selection discussed in the target article could be included in the same typology as simple and more complex case of the same phenomenon.

It is important to point out that there is no strict boundary between the three types of selection processes outlined above. During the development of the central nervous system, for example, more selection takes place after the initial simple one-step selection, since some new neural connections are formed even in the adult brain. The distinction between iterated one-step selection and multi-step selection is even more difficult to draw. In the latter environmental interaction causes replication to be differential, whereas in the former it does not. This effect of environmental interaction on differential replication, however, is a matter of degree. The case in which environmental interaction has no influence on how differential the next replication will be, and the case in which it fully determines what replicators can and cannot replicate in the next generation (in gene-based biological selection for example) are just the two extremes. It can be argued that most cases of selection involving repeated cycles lie somewhere in between these two extremes. Operant conditioning might be an example for this.

If the difference between iterated one-step selection and multi-step selection is indeed a matter of degree, then a more pluralistic typology could provide the conceptual framework for examining both types of selection processes as well as the gradual transition between them. Even if this transition is not gradual, the pluralistic typology would provide a better framework for comparing multi-step selection with simpler selection processes. But if it is, the pluralist approach becomes probably the only way of defending the general project of accomplishing the endeavor of the article: of giving a general account of selection.


Adams, P. R. (1998) Hebb and Darwin. Journal of Theoretical Biology 195: 419-438.

Darden, L., & Cain, A.J. (1989) Selection type theories, Philosophy of Science 56:106-129.

Campbell, Donald (1974) Evolutionary Epistemology. In: The Philosophy of Karl Popper, ed.: Schilpp, P. A. LaSalle: Open Court. pp. 413–463.

Changeux, Jean-Pierre (1985) Neuronal Man: The Biology of Mind. New York: Pantheon.

Edelman, G. M. (1987) Neural Darwinism: The Theory of Neuronal Group Selection. New York: Basic Books.

Popper, Karl (1972) Objective Knowledge: An Evolutionary Approach. Oxford: Clarendon.


I am grateful to Frank Sulloway and Eors Szathmary for their helpful comments.