Notes on Novelty 8: Conclusion – Post evo-devo

Notes on Novelty series:
1. Introduction
2. Historical considerations – before and after evolution
3: The meaning of evolutionary novelty
4: Examples – the beetle’s horns and the turtle’s shell
5: Evolutionary radiations and individuation
6: Levels of description
7: Surprise!
8: Conclusion – Post evo-devo

With the growth of developmental genetics, it is possible to see beyond the view of homologies working at the level of whole organs. The mechanisms that define the ordinate axes of structures, the genetic circuits that pattern them, and the cell types with which organs are formed can be considered. The more that researchers look, the more they will find that the same tools have been used to build a great variety of structures long thought to have independent histories. Discerning what has been conserved and what is novel in the origins of organs and body plans will be possible only with more comparative data, experiments on non-model animals, and targeted fossil discoveries from crucial nodes in the tree of life. (Shubin et al. 2009: 822)

My conclusion is that we find novelty in a subjective sense, based upon what we think we should find in the data. Further knowledge of the underlying or overlaying mechanisms should reduce the surprisal of the trait appearing either in the paleontological record or in a distinct taxonomic group. Only when it doesn’t and in cases where it should, should we start to seek non-Darwinian mechanisms. Much of the debate over novelty in evolution has centred implicitly around what the researchers themselves find interesting, surprising or contrary to the conventional interpretations. Another factor lies in a topic I haven’t discussed here: what counts as “suitably dissimilar”, so a few words on this.

Similarity and its converse, dissimilarity, is often applied to the issues of novelty. Stephen Jay Gould attempted, for example, to apply it in the case of the Cambrian explosion (1990); some fossils’ bodyplans were too dissimilar (using the consistency index of cladistic analyses, C) to put them into existing taxonomic categories (“phyla”); subsequent work did precisely that (Briggs and Fortey 2005). Relying upon morphology – that is, description at a certain grain (indeed, the only grain available to palaeontologists most of the time; cytological data is rare and molecular data nonexistent) – Gould found some traits and overall phenotypes too weird to fit into our classifications. Briggs and Fortey showed over several decades of work summarised in their paper that they could be seen as instances of stem groups of our present clade-based phyla (that is, phyla that have been made monophyletic).

What is similarity in taxonomy? It is true that rarely does this get any kind of analytic or quantitative analysis, especially in palaeontology, but that doesn’t mean there is no analysis underlying it of which the authors are not conscious (or think it too obvious to state). The best analysis of the recognition of similarity in the literature is that of Amos Tversky, who argued that the similarity of one thing to another is the overlap of features minus the unique features to each object, out of a “feature set” (1978). This is now called “Tversky Similarity”:

 [TS]

What is crucial to understanding similarity like this, which was developed for a generic psychological account and not just for science, is that the choice of features to evaluate (which may be entirely unconscious, or a cultural convention, or even biologically imposed) determines what counts as similarity in each case. With a different feature set, different similarity indices might arise, and indeed the history of science has been one of finding the “right” feature set (for example, in the case of elements and the periodic table, Scerri 2008).

The alternative to using apparent similarity, especially at a given grain of description (the gross morphology of the organism) is to use homology. Instead of treating the relations between organisms and specimens as a matter of what they “look like”, identify homologs and organise your groupings on that basis – this is the reason for cladistics. Homologies, not similarities in the eyes of the beholders, set up a baseline for assessing how novel a trait may be, and if we can find the developmental, physiological, molecular and environmental reasons for the novelties thus uncovered, so much the better. All I am criticising is the idea that somehow, in ways we can’t even properly articulate, we need to “go beyond” or “extend” the Darwinian approach. In one sense of course we do; Darwin did not know everything. In this case, we do not; Darwin has been validated again and again regarding novelty.

I am not dismissing the idea that there can be an evolutionary explosion from some novelty or node in an evolutionary tree, although these are more often the result of subjective assessments (the Cambrian explosion has, for example, become less and less of an explosion and more and more of an evolutionary diversification of the ordinary kind, as paleontological evidence has come in). What I am dismissing is the notion that there is some objective sense in which evolutionary novelties are natural kinds. Radiations occur all the time – some clades are speciose and others are sparse. If there is a general principle why this occurs, it is not yet obvious. Maybe some kinds of developmental modularities do cause clades to be more radiative. Maybe, however, is not an explanation.

Science progresses best when it eliminates subjectivity from its categories; anthropomorphism has ever been the bugbear of good science (unless anthropoi are the objects of study, and even there we tend to project ourselves on our subjects, as any anthropologist can tell you). It has taken us over a century to begin to recognise that what Darwin started requires this in biology as in the other physical sciences. So I would like to leave the last word to somebody of a certain weight in evolutionary biology, and who seems to be on the right side of the debate on many issues, Sewall Wright:

“Creative” and “emergent” evolution

The present discussion has dealt with the problem of evolution as one depending wholly on mechanism and chance. In recent years, there has been some tendency to revert to more or less mystical conceptions revolving about such phrases as “emergent evolution” and “creative evolution.” The writer must confess to a certain sympathy with such viewpoints philosophically but feels that they can have no place in an attempt at scientific analysis of the problem. One may recognize that the only reality directly experienced is that of mind, including choice, that mechanism is merely a term for regular behavior, and that there can be no ultimate explanation in terms of mechanism—merely an analytic description. Such a description, however, is the essential task of science and because of these very considerations, objective and subjective terms cannot be used in the same description without danger of something like 100 percent duplication. Whatever incompleteness is involved in scientific analysis applies to the simplest problems of mechanics as well as to evolution. It is present in most aggravated form, perhaps, in the development and behavior of individual organisms, but even here there seems to be no necessary limit (short of quantum phenomena) to the extent to which mechanistic analysis may be carried. An organism appears to be a system, linked up in such a way, through chains of trigger mechanisms, that a high degree of freedom of behavior as a whole merely requires departures from regularity of behavior among the ultimate parts, of the order of infinitesimals raised to powers as high as the lengths of the above chains. This view implies considerable limitations in the synthetic phases of science, but in any case it seems to have reached the point of demonstration in the field of quantum physics that prediction can be expressed only in terms of probabilities, decreasing with the period of time. As to evolution, its entities, species and ecologic systems, are much less closely knit than individual organisms. One may conceive of the process as involving freedom, most readily traceable in the factor called here individual adaptability. This, however, is a subjective interpretation and can have no place in the objective scientific analysis of the problem. [Wright 1931: 159]

Briggs, D. E. G., and R. A. Fortey. 2005. Wonderful strife: systematics, stem groups, and the phylogenetic signal of the Cambrian radiation. Paleobiology 31 (2):94-112.

Gould, Stephen Jay. 1990. Wonderful life: the Burgess Shale and the nature of history. London: Hutchinson Radius.

Scerri, Eric R. 2008. The Past and Future of the Periodic Table. American Scientist 96 (1):52-58.

Tversky, Amos, and Itamar Gati. 1978. Studies of similarity. In Cognition and categorization, edited by E. Rosch and B. B. Lloyd. Hillsdale, NJ: Lawrence Erlbaum Associates:79-98.

Wright, Sewall. 1931. Evolution in Mendelian populations. Genetics 16 (2):97-159.