Evolution does not have direction...

After the post spent harrumphing about scala naturae thinking, I should perhaps follow up a bit to explain my understanding of evolution as it relates to questions of teleology and contingency. As so often, do not assume that evolutionary biology is my speciality, but as a phylogeneticist and systematist I cannot avoid having an opinion. Also, this is once more mainly meant to put my thoughts into a coherent form, but maybe somebody will find them helpful or controversial enough to notice.

I would like to make the following points:

1. As a whole, evolution does not have a direction or a goal. Teleological thinking is a severe misunderstanding of evolution. Consequently, there are generally no 'primitive' or 'advanced' species in biology. This one should not be particularly controversial among biologists but it is sometimes disputed among non-biologists.
2. But! There are a few special cases where I find it helpful and justified to think in terms of directionality and more primitive or more advanced solution to an adaptive problem. Again, in a few special cases only, under a set of very specific circumstances.
3. Finally, the question of inevitability versus contingency. There are those who argue that if the tape of history were rewound and played again, evolutionary history would unfold completely differently because every step is contingent on the ones that came before, and thus we would find completely different life forms in an alternate history or, for that matter, on different planets. On the other side there are those who argue that if the tape of history were rewound and played again, things would come out pretty much the same. I lean towards the latter position - to a degree.

Today, let's deal with the first point. I will expand on the other two later.

There appears to be a limited number of reasons to think that evolution has a goal or moves into a certain direction. The first is clearly wishful. Many people, and not only those who claim that god(s) exist and used evolution to bring humans about, like to see humans as the inevitable end product of a march of progress towards greater complexity and ultimately sentience. However, this is seemingly supported by a second one. Looking across deep time, it sure appears as if life evolved in the direction of increased complexity: first there were only bacteria, then much more complicated eukaryotic cells, then multicellular organisms, then big animals and land plants...

As a side issue, we may wonder whether complexity is as easily defined and quantified as it seems. Suggestions include the genome size of an organism (which is obviously not a good metric because most of the genome of eukaryotes is demonstrably uninformative noise), the number of different genes (hard to quantify without a fully annotated genome but perhaps a better idea), or the number of cell types an organism has. There are problems with all of them but for the sake of the discussion I will grant that we can have a general hunch about complexity, e.g. bacterium < mushroom < elephant, even if it gets fuzzy at smaller scales, e.g. fruit fly vs. elephant, and I will grant that more complex organisms evolved later in the history of the planet. This latter observation can hardly be called false, but the question is what it really demonstrates.

We can suggest two models, let us call them the 'march towards complexity' model and the 'random walk' model, contemplate what observations we would expect under each model, and then compare them against what we see in nature.

Under the random walk, we would expect evolutionary change to go into both directions, towards complexity and towards simplicity. So, does the observation of increasingly complex life forms at later stages of evolutionary history contradict the model? Actually, no. We have to keep in mind that (a) simplicity must have a lower bound beyond which a cell cannot exist any more, at least not as an independent organism*, and (b) life must obviously have started at that lower bound.

If we now visualize what evolution would look like if it were a random walk starting at a lower bound, we realize that it would look precisely like what we see. At the beginning, there were only simple, small organisms. They randomly evolve upwards and downwards, and over time the morphospace of 'big and complicated' is filled from below. But the organisms can never evolve further down from where they started, and thus it appears as if evolution is moving towards complexity.

Of course, the march of progress model fits this observation just as well. In this case evolution starts with small and simple and then inexorably moves towards big and complex, giving us the very same impression of moving up, only in this case it would be correct. So at this moment nothing is decided.

What differentiates between the two models is the behaviour of individual lineages. Under the march of progress, we would expect all groups of organisms to evolve towards complexity; maybe some of them do it slower than others but they would not regress and become simpler and 'more primitive'. Under the random walk, we would expect members of all manner of diverse lineages to evolve either towards complexity or towards simplicity.

Put like that, it should be abundantly clear which of the two models fits reality better. We merely have to make a list of all the lineages which have become simplified or, if you cannot avoid thinking in terms of primitivism, 'regressed', and we will soon run out of space on our sheet.

Just to pick a few examples, as mentioned in my previous text post the fork ferns are now known to have secondarily lost their roots (which seems like a rather big thing to give up once you have obtained them). In fact they were so reduced in their complexity that they were for a long time considered to be living fossils from the Silurian era that never had them to begin with.

Our best scenario for the tiny rosette herbs of the quillwort genus Isoetes is that they are descendants of much larger, woody lycophytes of the order Lepidodendrales.

Duckweed. Their bodies are basically thalli, merely green blobs floating on the water. Two of the genera have even lost their roots. Would you believe that these genera have been demonstrated to be Araceae, in other words that their ancestors must have been something somewhat on the lines of these guys? Talk about becoming smaller and simpler in body plan!

And of course the monocots as a whole: a major lineage of flowering plants that has given up secondary growth, that is the ability to form wood, the ability to become a large tree. If you see land plant evolution as a march of progress from algae towards rainforest trees this should make you think.

And those are only plants. Think of all the animals that have become parasitic, reducing their bodies more and more in the process. Think of cave dwellers losing their sight, of flightless birds, of the various vertebrate lineages having lost their legs...

In summary, it sure looks as if individual lineages do not move towards greater complexity. They go here or there wherever opportunities beckon, even if that means going 'backwards' in terms of complexity, even if that means losing organs that are useful to many other species. The random walk model is a good fit to our observations, the march of progress model isn't.

As indicated at the beginning, it follows logically that the terms primitive and advanced don't really have any meaning; for that, there would have to be a directionality to evolution, and there isn't. There is only what has managed to survive, and a simple organism has survived just as well as a complex one. (And it has probably better chances of surviving the next mass extinction event!) There is no evolutionary progress, instead there is simply diversification.

Footnote

*) Parasites can, of course, be considerably simpler than independent organisms, but that assumes that the independent organisms that they parasitize on exist already, and those still have a lower complexity bound if they want to be functional.