Happy Newtonmas

Over the past 50 years or so, there have been many attempts to give a general metaphysics of evolution, ranging from axiomatisation (by Mary Williams, at the height of the “theories are axiomatic systems” period*), to “logical necessity” cases (such as Lewontin’s three conditions for natural selection), to “units of selection” arguments, most closely associated with George Williams and RIchard Dawkins. In each of these, and other, attempts, there has always been the presumption that there is a fixed hierarchy of ranks and units in biology. These are the “forms” of biology: replicators, interactors, species, genes, cells, and so on.

The odd thing about this is that as people were asserting that essentialism is dead (see the article on species linked above), they were being essentialists about concepts and units and ranks. Ernst Mayr, for example, who asserted that species individually (the species taxon, as he put it) have no essences, nevertheless asserted that the concept of species (the species category) did so. He was an essentialist about the species concept. Likewise, the gene centrism of a Dawkins is essentialist about the replicator concept. And so on.

Now one of the reasons why people adopted the hard and fast categories is that they usually were specialists in groups, such as mammals, birds or insects, where these categories had a real purchase. This is often referred to, mostly by botanists, as the “fur and feathers” or “vertebrate” or just “animal” bias. But another is just that they were seeking what used to be called the Characteristica Universalis, or the most general universal and formal language for the domain in question. It is a general disposition of those in the west to do this (and despite suggestions to the contrary, I cannot see how one might apply the Eastern metaphysics fruitfully in the domain of science). It is a constant temptation to try to ground ideas in unchanging and agential categories. We like species because they do something. We like replicators because they are the ultimate doers. These categories apply in ways that make sense of both the world, and our need for constancy. Coherence is not gone.

Until you stop focussing on the “obvious” cases, and start paying attention to as many as you can find. I have what I call the “esoteric method”: look for cases that don’t fit the current categories and then go look and see if that is more general than you might think. For example, in his 1942, Ernst Mayr referred to nonsexual organisms as “aberrant” when discussing the adequacy of his “new” “biological” species concept (122, 129). Today we know that not only are most organisms not sexual, which would mean most of them are not arrayed in species, but that the sexuality of species even in the small twig of the phylogenetic tree that is metazoans is not constant: many groups have either got hybridisation, or asexuality, or both. Nor is gene exchange confined to sexual species – between species gene flow is common, and even among asexuals lateral transfer is frequent. In fact the sort of species Mayr expected to exist are rare, except among some groups of vertebrates (oddly, the group Mayr studied, birds, often hybridise).

Over the past 50 years these essentialistic categories have become harder and harder to support empirically, as we have learned of more and more exceptions. Some, such as John Dupré, have argued for a pluralism of conceptions in biology due to the polytypic nature of the instances to which these categories are applied. It’s just a brute fact of biology that none of these categories are universal, and so biologists must avail themselves of whatever conception works in a particular case (to make this more concrete: species are sexual isolates when that works, but in, say, bacteria, they are phenetic clusters or something else).

Some years ago, I published an idea that I think might be the resolution to this (2003) in which I argued that species is like any other property of organisms, something that has evolved in its own way. The reason there is no universal notion of species for the same reason there is no universal notion of leg: species, like legs, are the outcome of evolution. In other words, these kinds themselves evolve. This applies also to other apparently universal aspects of biology: genes, or rather replicators, cells, individuals, and so on. It is not the case that, as Dupré thinks, that anything goes, but that there are evolved modalities, as I called them – ways of being whatever it is that we are trying to understand. This applies not only to the organisms and their traits, but to the kinds of organisms, and even to the kinds of kinds. Taxa, units, ranks, entities, systems – all these are evolved, and so to understand what it means to be, say, a bird species or a eukaryote gene, you need to understand the evolutionary relations of that group.

Last year, Peter Godfrey Smith published an interesting book that argues that the sole precondition for a Darwinian perspective on the world is that there are populations. Because we are disposed to see biology in terms of agency, we want agents, but that is, PGS holds, a remnant of the oldthink of teleology that Darwinism replaced. I think he’s well on the right track, although he still thinks that this means we cannot have types or classes. I think that classes are merely local and evolved. We are in a reading group covering his book right now, so as we work through it, I’ll probably add some more.

One thing I do want to say now, though, is that there is a prior problem knowing what a population is. For instance, to know that an ensemble of individuals form a population, you need, minimally, to show they are of the same species because you don’t get a population that spreads across two or more species, unless they are causally connected reproductively (in which case they might be classed as the same species anyway). Moreover, you already need to know the sort of object/organism that counts as an individual for that group in order to identify it as a population. This is not always so easy, in the case of colony organisms. While PGS is rightly arguing that there are no ranks or special units, only populations (which comprise individuals that have heredity and ecological differences, leading to evolution**), it seems to me that he still requires there to be some sort of types or equivalence classes, even if there are no universal kinds of types.

In part, this is something that comes out of the death of the essentialism story: it is often assumed that if one abandons essentialism, one loses access to any kind of equivalence class in biology (i.e., natural kinds; we aren’t worried about conventional classes or functionally defined classes), and that is what PGS assumes too. But it is my view that biology always uses types, which are defined or rather ostended by identifying an exemplar and then looking for clusters of properties. This is what PGS says we should be doing, but he does not see these as types. I do. By finding these clusters of properties (and even more the underlying developmental traits and heredity), we are then able to determine what a population is, and what individuals are, by a process of iterative induction (start with a case that is presumably exemplary and then make inductive generalisations from that until they fail).

What bothers people who think in terms, not of binaries as Chris Schoen suggested, but of absolute levels or entities that do not change, is that evolution leaves us gasping and dealing with vague boundaries, shifting kinds and so on. I feel for them, but it is really biology that does this, and always has. What really is novel about evolutionary thinking is that we know not only that the appearances change, but that the forms, and the forms of forms also change. However hard to come to grips with, we must. And the solution to this vagueness is phylogenetic thinking. If you know where a species or an organism is placed on an evolutionary network (allowing for the moment that the tree topology sometimes fails), then you know what sorts of sorts it will fall into, or if you find that it doesn’t, that sets up an interesting research project.

More as it occurs to me.

* Williams was a student of the originator of the Axiomatic Method for the sciences, Joseph H. Woodger.

** Evolution includes a lack of change by stabilising selection or developmental entrenchment (which I think may be a subset of the former). We need not presume that selection always causes change (but if there is a lack of change, I think we should presume that is due to selection).