Case 1: Evolution due to differences in variable properties

54 _{The notion of causality employed here is fully compatible with the manipulationist account of causation}

One possible explanation accounting for the difference in height in EC2 is that each entity of the population is in a different environmental context (that is, is different with respect to one intrinsic-variable property) causally relevant to height, yet one of which is correlated between parents and offspring. Apart from this difference we suppose no other differences between the entities of the population. Had each entity of the population been in the same environmental context, they would have had the same height. Let us a call this variant of the model ‘EC2*’. EC2* resembles a case of passive gene-environment correlation distinguished by behavioural geneticists (for more on passive gene-environment correlation see Sesardic 2005).

Figure 2.2. Illustration of the model EC2* – all differences in height between entities are due to some differences in intrinsic-variable properties.

Figure 2.2 illustrates this case. Imagine that the trait ‘head colour’ is causally involved in determining the height of each entity and is intrinsic-invariable within the range of possible environmental states so that two perfect clones growing in any two different states of the

environment have the same phenotype on the trait head colour. Let us also assume that ‘head colour’ is the only intrinsic-invariable trait involved in the determination of height and that each parent has the phenotype ‘grey head’ so that there is no variation at the population level in intrinsic-invariable properties between the entities of the population. Suppose also that each entity transmits its intrinsic-invariable properties (the phenotype ‘grey head’) perfectly to its offspring. Suppose finally that the level of resources received by an entity is causally involved in determining the height of each entity and that there is a gradient of resources in the population (as shown on Figure 2.2): the higher the amount of the resource received, the taller the entity. Each offspring once it is born, moves away from its parent but remains close to it. Because each entity is at a different position along the resource gradient, they all have a different height. Yet because each offspring remains located close to its parents, offspring on average resemble their parent more than they resemble other parental entities.

In this case, we have variation in height, that leads to differences in reproductive output and which is heritable (using the standard regression notion of heritability).55 _{Yet this is clearly}

not a case of ENS, let alone a pure case of ENS. In fact all the differences between entities are differences in extrinsic properties (the amount of resource they obtain depends on their parent’s position and their position) and in intrinsic-variable (height). There is also no population variation in intrinsic-invariable properties, thus these differences cannot be attributed to natural selection as per the different distinctions I made in Chapter 1.

This case, although not original (see the very similar although less detailed case 9 in Godfrey-Smith 2007), is useful for our purpose for three reasons. First, it shows us that

underlying a continuous trait such as height, there can be perfect inheritance of intrinsic-invariable factors (in our case the phenotype ‘grey head’) within the range of possible states of the environment. Observing a continuous trait is thus insufficient reason to claim that ENS does not require perfect inheritance to occur. This is a fairly obvious point but it is worth mentioning.

Second, it illustrates that unless the different phenotypes of a given trait are intrinsic and invariable within a specified range of environmental conditions, natural selection cannot be invoked as a cause of evolutionary change on this phenotype. If height is an intrinsic-variable phenotype within the different environmental states, over a given period of time, in a given population, and there is no variation in the population on an invariable property causally involved in determining the height of individuals, then height is not a property that is subject to natural selection. This is because whether entities survive or reproduce depends entirely on their extrinsic and intrinsic-variable properties. This point is important for our purpose because the fact that a trait is continuous or discrete is not relevant for ENS if this trait is an intrinsic-variable or an extrinsic one. It only matters if this trait is an intrinsic-invariable one. This means that there is not necessarily any fundamental disagreement with the view that natural selection requires the perfect transmission of types over time and Godfrey-Smith’s evolutionary nominalism if the traits under scrutiny, such as height in our case, are not intrinsic-invariable.

Finally, this case demonstrates that even if in a population there is a positive heritability, under the regression approach, and differences in reproductive outputs between the different individuals, this does not necessarily make this population a Darwinian population, since the resemblance between parents and offspring might be due to a correlation between the environment of parents and offspring, rather than caused in the relevant sense by the parents. This suggests that the variance approach to heritability might in some respects be superior to the

regression approach for discriminating the effects of natural selection from drift or correlated responses. This is mainly because the variance approach takes the independent population variable to be an intrinsic-invariable property of entities, which is instantiated by evolutionary genes (see Haig 2012 for details on what this concept entails) while the regression approach takes any population parameter (intrinsic-invariable or intrinsic-variable: that is, any phenotype) as an independent variable. This means that the regression approach to heritability captures more evolutionary processes than natural selection (see Chapter 4 for more details).