Whether you can observe a thing or not depends on the theory which you use. It is the theory which decides what can be observed.

— Albert Einstein, 1926

[Gold reminds us we must not forget] … the striking reformation of evolutionary theory implied by the well-documented genetic and developmental homologies alone. De Robertis expresses this key argument in the final line of his 1997 article on the ancestry of segmentation: “The realization that all Bilateria are derived from a complex ancestor represents a major change in evolutionary thinking, suggesting that the constraints imposed by the previous history of species played a greater role in the outcome of animal evolution than anyone would have predicted until recently.” (Gould 2002: 1152) [De Robertis, E.M. 1997. The ancestory of segmentation. Nature 387: 25-26. See also, De Robertis, E.M., G. Oliver, and C.V.E. Wright. 1990. Homeobox genes and the vertebrate body plan. Scientific American, July, pp. 46-52; De Robertis, E.M., and Y. Sasai. 1996. A common plan for dorsoventral patterning in Bilateria. Nature 380: 37-40.]

(….) Hughes (2000, p. 65) has expressed this cardinal discovery of evo-devo in phyletic and paleontological terms: “It is hard to escape the suspicion that what we witness in the Cambrian is mainly tinkering with developmental systems already firmly established by the time these Cambrian beasts showed up.” (Gould 2002: 1155) [Hughes, N.C. 2000. The rocky road to Mendel’s play. Evol. and Develop. 2: 63-66.]

— Gould, Stephen J. The Structure of Evolutionary Theory. Cambridge: Harvard University Press; 2002; p. 1152; 1155.

As it turns out, the miracle of complex life is more amazing, yet ironically simpler, than anyone ever expected. Researchers now know that life’s building materials are few, and they were “invented” near the dawn of animals. More specifically, a surprisingly small number of genes—”tool kit genes”—are the primary components for building all animals, and these genes emerged at a time before the Cambrian Explosion, some 600 million years ago. Thus the amazing diversity of the animal kingdom is the result of the flexibility of a small number of building blocks that have existed for eons.

This means, for example, that the gene that controls the formation of an arm on a human is the same gene that controls the formation of a wing on a bird, a fin on a fish, and a leg on a centipede, and that this gene has been around since the first animals grew the first appendage of any kind. Some prominent scientists have argued that if we could rewind the tape of life and start over again, the result would be a totally different world from that which exists today. They are wrong. Tool kit genes conserve the essence of animals, and they react to ecological cues in very consistent ways [emphasis added].

— Carroll, Sean B. Endless Forms Most Beautiful: The New Science of Evo Devo. New York: Norton; 2005: Inside Dustjacket.

We now need to confront the question of whether the biological community — or at least the large proportion of it — has come to accept a theory of evolution that is based on a broadly parallel error. Our case studies on the action of natural selection all involve microevolutionary changes occurring within particular lineages hundreds of millions of years after the origin of the major body plans of which the species concerned represent variations. Many of these case-studies are well known, especially the evolution of industrial melanism in Biston (Bishop and Cook 1980), the evolution of pigmentation patterns in Cepaea (Jones, Leith and Rawlings 1977) and the evolution of Batesian mimicry in several lepidopterans (Turner 1977). Many paleontological case studies are also restricted to particular lineages, with studies on the horse (Simpson 1951; MacFadden 1992) and the mollusks of Lake Turkana (Williamson 1981) being among the best known. While such studies are usually transspecific, and therefore in the realm of ‘macroevolution’, they are only a very short distance in that direction from an origin-of-body-plans perspective. (Simpson (1944) used the term ‘mega-evolution’ for the biggest-scale evolutionary events such as body plan origins, but this term has not become widely adopted.)

So, this book is starting with an exhortation to the reader to believe that current evolutionary theory, based on natural selection and adaptation in present-day lineages is, at the very least, incomplete; and this exhortation is based on the drawing of a parallel between the processes of development and evolution. (Arthur 1997: 2-3)

(….) Regardless of timing of early [Cambrian] divergences, it appears that no phylum-level body plans have arisen in the animal kingdom in the last 500 my. This contrasts with the situation in plants, where teh angiosperm body plan arose relatively recently (probably about 130 my ago: see Hickey and Doyle 1977; Crane, Friis and Pederson 1995). Perhaps this difference relates to a difference in developmental-genetic control mechanisms in the two kingdoms, with some genes controlling the determination of animal body axes and other key processes of early ontogeny being more ‘generatively entrenched’ (Wimsatt 1986) than their nearest equivalents in plants. (Arthur 1997: 7)

(….) [O]ur current (neo-Darwinian) theory of evolution is incomplete…. In fact, neo-Darwinian theory is incomplete even when assessed against its own criteria. The essence of the neo-Darwinian view is that the evolutionary process is of a two-fold nature, involving the production of organismic novelties (of whatever sort) ultimately by mutation and the sieving of these by natural selection. (Arthur 1997: 9)

(….) The main problem with neo-Darwinism in its current form is that its theoretical structure is extremely lopsided. There has been sustained development of quantitative models of the action of selection, from the pioneering work of Fisher (1930), Haldane (1932) and Wright (1931) up to recent work such as that of Charlesworth (1994); while the mutational and developmental production of the variants being sieved by selection has continued to be treated by too many evolutionists as a ‘black box’, despite the numerous advances that have been made in developmental genetics in recent years. Essentially, the individual and population levels have been treated as quasi-independent. The fitness of mutant genotypes have been considered to be crucially important in models of selection, while the ways in which fitness effects are produced … have been largely disregarded. (Arthur 1997: 9-10)

This situation should of course be considered undesirable by all evolutionary biologists, including the strictest of neo-Darwinians, but how serious a problem the lack of a mutational/developmental component of evolutionary theory is perceived to be depends on the extent to which the ‘perceiver’ is a gradualist. If, despite the views put forward herein, all evolution proceeds through the accumulation of very minor variations — an extreme view popularized by Dawkins (1986) — then it may not be too much of a deficiency in the theory to simply assume that mutation perpetually generates morphologies that are slight variants on the existing form. But to anyone proposing the existence of one or more radical morphogenetic phases in evolution, the need for an adequate picture of the genetic architecture of development and of the ways in which this is altered by mutation becomes compelling. Hence the feelings of dissatisfaction that many evolutionary developmental biologists have with neo-Darwinism. There is nothing wrong with elaborate models of selection, but a detailed quantitative statement of how existing types are sorted and selectively eliminated (or held in a state of stable equilibrium) cannot pretend to be a complete theory. (Arthur 1997: 10)

Ironically, most of the alternative approaches to evolution that have proliferated in the last few decades have allowed the focus on destructive rather than creative forces to persist. The neutral theory of molecular evolution (Kimura 1983) — arguably within a broad neo-Darwinian world view — concentrates on the stochastic loss of neutral and nearly neutral alleles produced in an unspecified way by mutation. Punctuated equilibrium (Eldredge and Gould 1972) is a pattern, not a process, and may simply be a geological reflection of the standard neo-Darwinian mechanism of allopatric speciation, although some authors (e.g. Williamson 1981) have suggested otherwise…. (Arthur 1997: 10)

(….) The only approach [as of 1997, at the time of this writing] to evolution that has attempted to focus on creative forces has been that of Evolutionary Developmental Biology. I use this label (… Hall 1992) to cover the work of a heterogeneous group of biologists including, among others, von Baer (1828), Thompson (1917), de Beer (1930), Goldschmidt (1940), Waddington (1957), Gould (1977b [2002]), Raff and Kaufman (1983), Buss (1987), Arthur (1988), Thomson (1988) and Raff (1996). (Arthur 1997: 11)