Philosophy of Biology and Cognitive Sciences

The variation upon which natural selection and genetic drift act at the population level is modelled as random in classical evolutionary theory models. However, the field of evolutionary developmental biology or evo-devo aims at revealing the developmental mechanisms underlying variational patterns. Variation is highly structured by developmental properties, which link genotypic sequences to phenotypic outcomes. This structuration results in probabilistic patterns, which are one of the main targets of evo-devo. In this talk I want to provide a philosophical frame for understanding the explanatory role of the properties of developmental systems explored in evo-devo. Since the field of philosophy of biology is strongly concerned with the causal and explanatory structure of evolutionary theory, I believe that providing a philosophical frame for the new explanatory agendas coming from the so-called extension of the Modern Synthesis of evolutionary biology is of philosophical interest for this event. Departing from a propensity interpretation of variational probabilities (Nuño de la Rosa and Villegas, manuscript), I will explore the different ways in which the models of evo-devo represent variational tendencies of developmental systems in order to explain evolutionary change and innovation, arguing that the explanatory relevance of such tendencies leans on their typological nature. In order to do so, I will use as illustration a very simple model of a Genotype-Phenotype map: the ARN model (Fontana 2002).

 

In this model, the genotype represents a RNA nucleotide sequence, development corresponds to the folding process of the ARN molecule, and the RNA final shape or secondary structure is taken as the phenotype. The model provides with the space of all possible genotypes and assigns a phenotype to each of them. Importantly, the mapping from genotype to phenotype is not of a one-to-one kind. Unlike most models in evo-devo, where probabilities are qualitatively characterized due to the complex developmental interactions that give rise to complex phenotypes, in this model, variational probabilities can be simply quantified. That is, the probability of a mutation giving rise to a phenotypic change is well-defined. This model represents the relevant patterns of variation explained by developmental systems, independently of the impact of selection in the frequencies of phenotypes. Through it, it is possible to study the developmental properties that affect variation, such as modularity, robustness and evolvability.

 

The propensities of developmental systems are responsible for the variational patterns represented in evo-devo models such as the ARN Genotype-Phenotype map. These propensities are typological insofar as the properties represented in the model are relations between types, that is, they represent the variational properties of genotypes and phenotypes. For instance, variability models include as the input space the genotypic space, i.e. the space of all possible genetic perturbations of a developmental type, as well as all the possible developmental pathways mapping genetic onto
heritable phenotypic variation. Moreover, insofar as they include non-realized ontogenies, developmental types cannot be interpreted as mere abstractions from populations of individual ontogenies, contrarily to what authors such as Lewens have defended (2009).

 

Typological propensities play an important explanatory role in the philosophy of evolution literature. On the one hand, much of the philosophical discussion on causation in biology is based on dispositional accounts. On the other hand, the causal processes relevant for biological explanations seem to be not easily reducible to individual (token) causes. Roberta Millstein (2003) argues that there is a distinctive way of individuating biological kinds that play a causal role in evolution: it is the case of populations. She claims that population genetics, a discipline at the core of classical evolutionary theory, describes evolution, where causes such as selection and genetic drift take place, as a population-level process, and that this process is grounded in populational propensities.

 

Contrarily to Millstein’s claim that the relevant types for evolutionary theory are populations, I will lean on the RNA map in order to show that developmental types are the causally relevant factors involved in the generation of evolutionary variation. I will defend that the explanatory power of such evo-devo properties partially relies on their reestablishment of a certain kind of typological thinking without necessitating the –always controversial– reification of types. I will then argue that this kind of typology provides evo-devo explanatory models with greater counterfactual robustness, enabling them to address how-possibly questions about the generation of variation and innovations more satisfactorily than populational explanations.

 

References

Fontana, W. (2002). Modelling ‘evo-devo’ with RNA. BioEssays, 24(12), 1164-1177.

Lewens, T. (2009). What Is Wrong with Typological Thinking? Philos Sci 76:355–371.
Millstein, R. L. (2003). Interpretations of probability in evolutionary theory. Philosophy of Science, 70(5), 1317-1328.

Nuño de la Rosa, L. and Villegas, C. (manuscript). Chances and propensities in evo-devo: variability, evolvability and random variation.