This chapter discusses the 'gene's-eye' view of evolution associated with Williams (1966), Dawkins (1976), and others. Section 5.1 outlines the origins of gene's-eye thinking. Section 5.2 argues for a distinction between the process of genic selection and the gene's-eye perspective, or viewpoint. Section 5.3 examines 'outlaw' genes, which benefit themselves at the expense of their host organism, leading to intra-genomic conflict. Section 5.4 revisits a theoretical issue from Chapter 3—the tension between the Price and contextual approaches to MLS1—and considers it in relation to genic selection. Section 5.5 looks at the well-known 'bookkeeping' objection, which says that the gene's-eye view merely records the outcome of evolution but says nothing about the underlying dynamics. Section 5.6 asks whether the gene's-eye view can accommodate epistasis and the context-sensitivity of gene action. Section 5.7 briefly revisits the topics of pluralism and reductionism.


Gene's-eye thinking can be traced back to the earliest days of evolutionary genetics. In his famous 1930 book, Fisher's attempt to synthesize Darwinism with Mendelian genetics led him to a novel conception of the evolutionary process. Instead of thinking in phenotype space, as Darwin had done, Fisher operated at the level of the underlying genes. He thought of natural selection as operating on a large population of genes, or gene pool, altering the pool's allelic composition over time. A separate selection coefficient, that is, fitness value, could be calculated for each allele, allowing organisms and their phenotypes to be bypassed. Evolutionary change, on this view, is simply gene frequency change, and natural selection is a force that leads fitter genes to be substituted for their alleles. Dawkins's concept of 'selfish genes' owes much to this Fisherian picture.

Fisher's picture is obviously an abstraction, as is the gene pool concept itself. In a review of Fisher's book, Wright (1930) questioned the utility of the abstraction on the grounds that 'genes favourable in one combination are. . . extremely likely to be unfavourable in another' (p. 84). Such epistatic effects mean that selection is not usefully thought of as operating separately at each locus, Wright argued. He allowed that individual genes could be ascribed selection coefficients, but regarded this as a computational convenience, not a reflection of the real causal forces at work. 'Selection relates to the organism as a whole and its environment and not to genes as such', Wright wrote (1930 p. 156). Contemporary opponents of gene's-eye thinking often make an argument strikingly similar to Wright's (cf. de Winter 1997).

The true utility of the gene's-eye approach only became apparent in the 1960s, thanks to Hamilton's work on the evolution of altruism. Darwin himself had realized that an organism which behaves altruistically will be at a selective disadvantage vis-a-vis its selfish counterparts, ceteris paribus. So it seems that altruism, and the genes which cause it, should be eliminated by natural selection.1 Hamilton saw that the logic of this argument breaks down if altruistic actions are preferentially directed towards relatives. For relatives share genes, so there is a certain probability that the beneficiary of the altruistic act will itself carry the gene for altruism. To determine whether the altruism-causing gene will spread, we need to take into account not just the effect of the gene on the fitness of its bearer, but also on the fitness of the bearer's relatives.2

Hamilton realized that an even simpler way to determine whether a gene for altruism will spread is to forget about organismic fitness and think directly in genic terms, just as Fisher had done. He wrote: 'despite the principle of the ''survival of the fittest'', the ultimate criterion which determines whether a gene G will spread is not whether the behaviour [it causes] is to the benefit of the behaver, but whether it is to the

1 In speaking of a 'gene' that causes altruism, we mean only that the gene increases the probability that its bearer will behave altruistically by some amount. This involves no presumption of genetic determinism, nor a downplaying of environmental effects on phenotype, as Dawkins (1982) rightly stresses.

2 This is encapsulated in Hamilton's famous rule for the spread of altruism, b/c > 1/r, where c is the fitness loss incurred by the altruist, b the fitness gain enjoyed by the recipient, and r the coefficient of relationship between altruist and recipient. The proof of Hamilton's rule relies on certain non-trivial assumptions; see Michod (1982), Grafen (1985), Queller (1992a), or Frank (1998) for details.

benefit of the gene G' (1963 p. 7). So although altruism may seem anomalous from the organism's point of view, it makes perfect sense from the gene's point of view. Inducing its host organism to behave altruistically towards relatives is a 'strategy' that a gene can use to boost its representation in future generations.

Interestingly, Hamilton showed that altruism can in fact be understood from the organismic viewpoint too. Though behaving altruistically reduces an organism's personal fitness (by definition), it increases its inclusive fitness. An organism's inclusive fitness is defined as its personal fitness plus the sum of its weighted effects on the fitness of every other organism in the population, the weights being determined by the coefficient of relationship r. Given this definition, natural selection will favour those organisms with the highest inclusive fitness. So instead of thinking of genes trying to maximize the number of copies they leave, we can think of organisms trying to maximize their inclusive fitness. Most people find the gene's-eye approach more intuitive than the inclusive fitness approach, but mathematically they are equivalent (Michod 1982; Hamilton 1996; Frank 1998).

Gene's-eye thinking was developed further by Williams (1966) and Dawkins (1976), who argued that all organismic adaptations, not just pro-social behaviours, are ultimately for the benefit of the underlying genes. Dawkins also emphasized 'outlaw' genes, such as segregation-distorters and transposons, which spread despite their negative effects on the host organism's fitness (and thus on the fitness of all genes at unlinked loci in the same genome). Recent research has revealed outlaws, or 'selfish genetic elements' (SGEs), to be much more common than was originally thought; they constitute one of the strongest arguments for the utility of the gene's-eye view (Pomiankowski 1999; Hurst et al. 1996; Hurst and Werren 2001; Burt and Trivers 2006).

Dawkins offered another, quite different argument for treating the gene as the unit of selection, namely that genes are 'replicators'. Though entities at other hierarchical levels can reproduce, hence form parent—offspring lineages, the fidelity of reproduction is typically lower than that of DNA replication. This is especially true for sexually reproducing entities, where offspring contain a mixture of genetic material from two or more parents. Only genes have sufficient permanence to qualify as units of selection, Dawkins argued; organisms and their phen-otypes are temporary manifestations. Despite the prominence Dawkins attached to this argument, arguably it confuses the unit of inheritance with the unit of selection. For as we saw in Chapter 1, replicators in

Dawkins's sense are not strictly needed for evolution by natural selection at all.

The influence of gene's-eye thinking has been enormous, particularly in behavioural ecology. Nonetheless, certain conceptual questions remain. Can all evolutionary change be understood from a gene's-eye viewpoint? Should the gene's-eye view be equated with genic selection, or are these concepts distinct? How does the gene's-eye view relate to the 'hierarchical' picture of evolution developed by multi-level selection theorists? These and other issues are explored below.


Proponents of gene's-eye thinking have been guilty of a certain ambiguity. Sometimes they present their view as an empirical thesis about how evolution happened, sometimes as a heuristic perspective for thinking about evolution. Williams (1966) suggests the former interpretation, for he contrasts genic selection with group selection. Williams argues that if group-level adaptations turn out not to exist in nature, as he suspects, then 'we must conclude that group selection has not been important, and that only genic selection. . . need be recognized as the creative force in evolution' (p. 123—4). Dawkins's early work takes a similar line. Discussing Wynne-Edwards's theory that reproductive restraint in bird species evolved by group selection, Dawkins (1976) argues that a 'selfish gene theory' can explain the data better. The implication is that if Wynne-Edwards's theory were right, the selfish gene theory would be wrong and vice versa.

However, Dawkins (1982) adopts a different line, claiming not to be propounding 'a factual position . . . but rather a way of seeing facts' (p. vi). The selfish gene theory is not an empirical alternative to orthodox Darwinism, he claims; rather, it is an alternative perspective that is often heuristically valuable. We can think of evolution either in the orthodox way, in terms of selection between organisms (or other 'interactors'), or in the gene's-eye way, in terms of selection between genes. There is no empirical issue at stake—both are valid perspectives on a single set of facts.

The idea that the gene's-eye view is a heuristic perspective, not an empirical thesis, is closely bound up with the distinction between 'replicators' and 'interactors', discussed in Chapter 1. As we saw, Dawkins and Hull argue that replicators and interactors play complementary roles in the evolutionary process. Organisms and groups are interactors but genes are replicators; so to oppose genes to organisms, or to groups, as rival units of selection is to commit a category mistake. Organism and group-level selection are both ways by which genes can spread in a population, Dawkins argues (1984 p. 162). In a similar vein, Buss (1987) argues that there is no incompatibility between the gene's-eye view and multi-level selection; for any selection process, at whatever level, can also be viewed from a genic perspective.

This is a compelling analysis, but it raises certain questions. First, is it really true that a gene's-eye perspective is possible on any selection process? Why should this be so? This question is explored in Section 5.5. Secondly, if the gene's-eye view is simply a different way of thinking about orthodox Darwinism, what become of outlaw genes, which are not explicable in terms of advantage to the individual organism? The existence of outlaws formed part of Dawkins's original case, but they sit badly with the idea that the genic and orthodox approaches are equivalent.

This latter problem can be resolved by distinguishing sharply between selection processes that occur at the genic level, and a gene's-eye view on selection processes that occur at other levels. In cases of outlaws, the selection process itself is at the genic level—for there are fitness differences between the genes within the same organism.3 In cases of organismic-, group-, or colony-level selection (for example) this is not so. But since selection at these higher levels typically leads to overall gene frequency change, it is possible to view the selection process from the gene's-eye perspective—even though the process itself does not take place at the genic level. So we must distinguish the process of genic selection, which is relatively infrequent, from the changes in gene frequency that are the product of selection at other levels, which are ubiquitous (or nearly so).4

The label 'genic selection' will therefore be reserved for selection between the genes within a single organism or genome, rather than for any selection process that leads to a gene frequency change. This understanding of 'genic selection' is increasingly used in the literature, for example, by Maynard Smith and Szathmary (1995), Sober and

3 Proponents of the replicator—interactor approach would express this by saying that where outlaws are concerned, the gene is both replicator and interactor at the same time; see Reeve and Keller (1999) for discussion of this move.

4 The qualification 'nearly so' is necessary for reasons explained in Section 5.5.

Wilson (1998), Hurst and Werren (2001), Gould (2002), and others. It follows that genic selection must not be equated with the gene's-eye view; these are separate concepts.

Hurst (1996) observes that the expression 'selfish gene' has undergone a shift in meaning since Dawkins first introduced it. Originally it denoted any gene that spread by natural selection, irrespective of the selective mechanism; later it came to be used for outlaws or SGEs—which spread at the expense of other genes in the same genome. Obviously, on the former usage 'selfish genes' will be much more common than on the latter; for most genes that spread do so by cooperating, not competing, with the rest of the genome, that is, by selection on higher-level units. The ambiguity noted by Hurst corresponds precisely to the distinction between the process of genic selection, and a gene's-eye perspective on selection processes that occur at other levels.

Classical kin selection is an example of a process that does not occur at the genic level, but on which a gene's-eye perspective is nonetheless valuable. By inducing its host organism to behave altruistically towards relatives, a gene for altruism can spread in a population, as discussed. But the gene is not an outlaw—it does not harm the interests of other genes in the host organism. On the contrary, since donor and recipient have identical coefficient of relatedness at every locus in the genome, all the genes stand to gain equally from the altruistic behaviour.5 So genic selection is not the force leading the altruistic gene to spread. However, it is still useful to adopt the gene's-eye view, and think of the altruistic behaviour as a strategy designed by the gene to boost its transmission. The alternative inclusive fitness approach to kin selection is much less intuitive, as noted above.

Buss (1987) attributes to M. Wade the remark: 'kin selection teaches not of the importance of the gene as a unit of selection, but of the family group as unit of selection above the level of the individual' (p. 184n). Wade's remark is insightful, for it highlights a confusion that many selfish gene theorists have fallen prey to, and ties in with the distinction between the process of genic selection and the gene's-eye perspective. (However Wade's own view—that kin selection is a type of group selection—is not universally accepted; others prefer to treat kin selection as type of individual selection, by regarding an individual's

5 This assumes that donor and recipient are relatives in the ordinary sense. If they are not relatives but both happen to share the altruistic gene for some other reason then matters are more complicated, for the coefficient of relationship will then differ at different loci; see Okasha (2002) for discussion of this point.

relatives as part of its environment; see Chapter 6. But whichever of these views we favour, neither implies that kin selection involves selection at the genic level.)

The process/perspective distinction is crucial for assessing the objections to gene's-eye thinking. Numerous authors have pointed to biological phenomena, such as epistasis, heterosis, and epigenetic inheritance, which they claim cannot be accommodated by a gene's-eye approach (Wright 1980; Sober and Lewontin 1982; Sober 1984; Michod 1999; Jablonka and Lamb 1995; Avital and Jablonka 2000; Gould 2002). But it is not always clear whether they mean that the phenomena do not involve the process of genic selection, or cannot usefully be viewed from the gene's-eye perspective. These claims are quite different.

Was this article helpful?

0 0

Post a comment