Irreversibility is generally accepted as a principle in evolutionary biology. The theoretical explanation for this observable phenomenon has simply been the claim that history does not exactly repeat or reverse itself in a complex world. But this is a very general principle, applying to all phenomena, not only the biological.
The proposed model eliminates the issue of irreversibility in skeletal evolution. If we think of evolution as an elaborative process, irreversibility is puzzling. One problem is that there is so much undeniable evolutionary reduction. If we think that, for example, the posterior vertebrae were gradually added, their observable gradual reduction would be an instance of reversion. With the proposed model, the column of vertebral segments was not gradually formed through a process of accumulating micromutations; its observable gradual reduction is not a reversion but a process completely different from the formative process.
Wih the suggested model, the evolutionary constraint against reversal is tangible and intrinsic. Evolutionary change takes place only through change in the regulation of the growth of a given set of structural elements. Once the expression of a structural element is suppressed, through some heritable disturbance in the process of ontogenetic differentiation, it cannot be restored, except through atavism. This is a process of distortion resulting from a one-way random degenerative process unrelated to the process which created the primitive complex of symmetrical skeletal elements. In an evolutionary sense, the number of parts is increased only through parabiosis, a mechanism too crude to add useful new parts to a complex skeleton.
If, to explain irreversibility, we only claim that the same complex of evolutionary events cannot recur due to mathematical improbability, how then can we explain parallelism and convergence?
The phenomena of parallelism and convergence (similar evolutionary developments in different lineages) can only be explained by a theory that explains evolutionary constraints. If morphology is somehow constrained to evolve in a limited number of channels, rather than being limitlessly plastic and elaborative, then similarities among divergent lineages are not such marvelous coincidences. Where evolution is a matter of reduction rather than elaboration, the range of possibilities is greatly narrowed. The main constraint upon evolution--in the particular arena under consideration, the historic evolution of segmented organisms--is that new parts are not added; new structures form only from parts which are already present.
forward to Incipience and Preadaptation