Evolution in Four Dimensions (E4D; Jablonka & Lamb Reference Jablonka and Lamb2005) argues that evolutionary biologists experience only one dimension of evolution – genetics – and proposes that epigenetics, culture, and language are additional evolutionary dimensions. Similarly, we experience four physical dimensions, but string theory in theoretical physics suggests that there are really ten dimensions (Woit Reference Woit2002). In both cases, the extra dimensions are attractive to theorists, but their experimental value is as yet unproven.

Evolutionary theory's power lies in its breadth. Its scope is such that a microbial ecologist (ADB) and a neuroethologist (ZF) have a meaningful common language. Any major extension to evolutionary theory, which E4D aspires to provide, should be broadly applicable to most organisms. This may be why combining evolution with genetics (the Modern Synthesis) and developmental biology (evo-devo) has been successful.

E4D is often irritating in its zeal to demonstrate its thesis that genetics has been overemphasized. It often uses disparaging terms (“genetic astrology,” p. 62), and makes argumentative statements (the alleged need to “abandon” the central dogma, p. 153).

E4D makes its strongest case for epigenetic inheritance systems, because they have the greatest potential for generality. Epigenetic inheritance is a cellular phenomenon, and all living things have cells that could be influenced through epigenetic mechanisms. Genetics and epigenetics could be grouped together as “cellular inheritance systems.” That there are several heterogeneous epigenetic mechanisms diminishes the prospects for wide applicability, however (indeed, “epigenetics” originated as a catch-all term meaning “not genetics”; cf. Morange Reference Morange2002).

The evidence for epigenetic inheritance having major impacts on evolution is limited, however. E4D frequently uses thought experiments to make its arguments. The concrete examples, such as an epigenetic morph of the flower Linaria vulgaris that has been stable for many generations, are more convincing than the thought experiments; and some of the interpretations of the real examples are questionable. Chapter 7 describes experiments in which silver foxes were selected for tameness. This selection for behavior generated morphological changes. It was hypothesized that stress caused the activation of “dormant genes” (Belyaev et al. Reference Belyaev, Ruvinsky and Borodin1981a; Reference Belyaev, Ruvinsky and Trut1981b). As far as we can determine, this hypothesis has not found strong empirical support during subsequent decades of further research (Gulevich et al. Reference Gulevich, Oskina, Shikhevich, Fedorova and Trut2004; Lindberg et al. Reference Lindberg, Björnerfeldt, Saetre, Svartberg, Seehuus, Bakken, Vilà and Jazin2005; Trut Reference Trut1999). An alternative hypothesis is that the morphological changes result from selective pressures on genes that have a common influence on both behavior and morphology, which can be tested as part of ongoing research on silver fox genetics (Kukekova et al. Reference Kukekova, Trut, Oskina, Kharlamova, Shikhevich, Kirkness, Aguirre and Acland2004; Reference Kukekova, Trut, Oskina, Johnson, Temnykh, Kharlamova, Shepeleva, Gulievich, Shikhevich, Graphodatsky, Aguirre and Acland2007).

Such matters raise practical concerns of how one might predict whether a particular organismal feature is likely to be inherited by epigenetic mechanisms, and if so, by which one. Epigenetic mechanisms also lack clear rules for determining how features will be inherited across generations, in contrast to the clear understanding of genetic inheritance.

Cultural (Ch. 5) and symbolic (Ch. 6) can be grouped together as brain-based inheritance systems. Thus, the book's argument that these should have equal status to genetics in evolutionary theory is immediately weakened, because not only do the organisms involved need brains, they need complex brains with particular properties. The cultural dimension applies only to a very limited number of animals (Whiten & van Schaik Reference Whiten and van Schaik2007), and the symbolic dimension applies only to humans.

To make incorporating brain-based inheritance into evolutionary biology worthwhile, Jablonka & Lamb (J&L) need to show that the similarities of cellular and brain-based inheritance are greater than the differences. But the differences are profound, and trying to put both systems in a common framework obscures more than it reveals. By analogy, a jet plane and a tricycle are both forms of transport. Some aspects of their behavior can be described in common terms: Knowing the velocity of the plane or the tricycle allows one to calculate its distance traveled over a given time, for example. But understanding how a jet plane works requires extensive understanding of aerodynamics; understanding a tricycle does not. Similarly, understanding a brain-based inheritance system will require a deep understanding of neurobiology and ethology; a cellular inheritance system will not.

There is no substantive discussion of nervous systems in E4D, which seems to consider animals as generic information processors. Animals are often constrained (Breland & Breland Reference Breland and Breland1961; Wells & Wells Reference Wells and Wells1957) or specialized (Healy et al. Reference Healy, de Kort and Clayton2005) in their cognitive abilities, including what they can learn, because of the particular neural circuitry of each species, which results from selection for capabilities that are relevant to each species' ecology (Healy et al. Reference Healy, de Kort and Clayton2005). As with epigenetics, there are not yet general rules to help us to make strong predictions about which behaviors or symbols are liable to be transmitted across generations, or in which species such behaviors are liable to be important. For example, many animals vocalize, but vocal learning (surely relevant to cultural inheritance) is limited to three orders of birds (songbirds, parrots, and hummingbirds; Baptista & Trail Reference Baptista and Trail1992), cetaceans (Deeckea et al. Reference Deeckea, Ford and Spong2000), bats (Boughman Reference Boughman1998), elephants (Poole et al. Reference Poole, Tyack, Stoeger-Horwath and Watwood2005), and humans. Why is vocal learning present in these taxa, but not others?

The Baldwin effect is mentioned as a way that behavior might influence genetics, but E4D admits there have been no experimental tests of these phenomena (p. 311), nor are there suggestions for testing the Baldwin effect (apart from suggesting Drosophila as an experimental organism). E4D also tries to distance symbolic inheritance from the meme concept, although we struggled, and ultimately failed, to understand why.

E4D is provocative in both the worst and best senses of the word. Despite our many reservations, disagreements, and outright annoyances with this book (provocative in the worst sense), it led us to find interesting research that was new to us, and forced us to consider our theoretical points of view carefully (provocative in the best sense). Ultimately, the book's biggest missing piece is that it does not suggest a research program for empirical evolutionary biologists. In contrast, the “gene's eye view” (Dawkins Reference Dawkins1976; Wilson Reference Wilson1975), which E4D criticizes so strongly and at such length, galvanized biology because it generated myriads of testable hypotheses.