If you like the science on this weblog, I highly recommend Laelaps, Brian Switek's contribution to the ScienceBlogs network. Where I am more micro and anthro oriented he is more macro and spans the whole tree of life. I'm really glad he's on ScienceBlogs; Laelaps adds to the diversity in an interesting way.
In any case, I wanted to point to this long post, Troodon sapiens?: Thoughts on the "Dinosauroid", it mulls over many concepts and evolutionary processes. Brian highlights the alternative views of the paleontologists Stephen Jay Gould and Simon Conway Morris. While Gould emphasized historical contingency and overall stochasticity of evolutionary process, Morris tends to lean toward the power of selection generating convergent adaptations. Some have suggested that Morris' views are influenced by his Christianity. Brian does note that this tendency toward teleology has correspondences with the talking points of the Intelligent Design movement, but I think it is important to observe that Richard Dawkins has come down on the side of Morris and against Stephen Jay Gould on this question. So the alignments here can be rather confusing to an outsider (some critics of adaptationism argue that the Oxford school of evolutionary biology, of which Dawkins is a representative, owes a great deal to William Paley's arguments from design, simply substituting the theistic god for the blind engineer of selection).
On a philosophical note, I do think these arguments about contingency and inevitability have to framed within the context of time and space. Assuming enough evolutionary time and a large enough effective population size it is imaginable that contingency and constraint can eventually be circumvented as selection slowly explores every nook and cranny of the adaptive landscape. But of course that is assuming particular parameters; over shorter time periods stochastic forces can be critical in explaining the madness. What is a large effective population? And what is a long time period? There are assumptions often unspecified in these debates between those who argue for contingency and those who argue for inevitability, and I think quite often that results in talking past one another.
That should be "talking past one another".
I disagree that enough time circumvents contingency - contingent variation is built upon by further change, becomes necessary, and then indispensable. I doubt any amount of evolutionary development would cause the fundamental flaw of the mammalian eye to disappear, and our renal organs do not seem capable of working as efficiently as reptiles'.
Likewise, I don't see jellyfish evolving into vertebrates no matter how much time they're given. Time's arrow only points one way, and the mutations necessary to change a fundamental aspect of their physiology are too improbable.
those lineages might go extinct over a short period of time ;-)
I have read both Conway Morris and Gould on this. My sympathies are mostly with Gould, for non-biological reasons.
I don't think that convergence has that power that CMorris thinks. For example, Gould argued early on that placentalness has no particular advantage over marsupialness. It's just that at the time when the land bridge formed in Panama (and even more, at the time when Europeans reached Australia) the placentals had been honed by a lot of severe selection and decimation, so they were more fit than the particular marsupials they met. (One marsupial, the Virginia opossum, has expanded its range as far as Canada in recent decades).
At the same time, as CMorris points out, convergence produced marsupials similiar to placentals, e.g. a marsupial tapir. But note that the "marsupial wolf" is sometimes called a "marsupial tiger" and in many respects seems more like a hyena. In other words, convergence is pretty rough.
So as I understand, the question is whether there were "deep" random turning points (like marsupial vs. placental), where a major decision is made basically on the QWERTY principle. Convergence doesn't affect this.
CMorris also does mention the hox genes, though, which are very deep and are shared in some form by a high proportion of animate creatures (e.g. flies and mammals) -- maybe all. One of his arguments against Gould's Burgess shale speculations is that a lot of those creatures were more strange visually than they were biologically, and at the deep level were pretty similar.
In short, I had no way of deciding.
It's just that at the time when the land bridge formed in Panama (and even more, at the time when Europeans reached Australia) the placentals had been honed by a lot of severe selection and decimation, so they were more fit than the particular marsupials they met. (One marsupial, the Virginia opossum, has expanded its range as far as Canada in recent decades).
right, the same process is seen when noting that more taxa regularly overran north america from eurasia when beringia was open then vice versa.
"our renal organs do not seem capable of working as efficiently as reptiles'."
Huh? How are you defining "efficiency" in renal-organ function? Many reptiles can excrete more nitrogen with loss of less water than mammals, but that is a function of liver biochemistry more than kidney function (nitrogenous wastes are packaged as urate rather than urea). Mammalian kidneys can produce urine that is more concentrated (as a solution) than the kidneys of reptiles (which cannot produce urine more concentrated than blood plasma).
Simple explanation of the differences in vertebrate renal systems
Mammals have a more advanced system to compensate - poorly - for a much less advanced biochemistry. Reptiles can secrete uric acid instead of urea, which means they can rid themselves of nitrogenous wastes while losing only a tiny fraction of the moisture mammals do in the process. This is a large part of why reptiles do so well in arid environments.
Of course, a truly advanced organism would be capable of nitrogen fixation and wouldn't have the waste problem in the first place.
The problem I have with Morris's convergent evolution argument is that the dinosaurs were around for hundreds of millions of years yet they never (as far as I know) evolved into anything vaguely 'humanoid'. Similarly, if you go back in time a few million years, neither had the mammals. It's only that we're living now, when there are humanoids around contemplating their existance, that encourages us to believe that the humanoid design was somehow inevitable. In some respects, it's rather like the Anthropic Principle.
Having said that, I seem to be alone in not having a problem with the Anthropic Principle.
Right. That's what I said: the reptile "advantage" is in biochemistry, not kidney function.
The most self-sufficient organisms, I think, are the cyanobacteria that can fix both nitrogen and carbon; I would not call them "advanced."
I just wanted to say "Thanks!" for the link and the kind words, Razib. I definitely agree that in considering convergence in evolution (as well as contingency and the SCM/Gould debate) the time/space/details of what we are talking about need to be defined if we're going to make any sense of it all, although I have to admit I laughed a little when I heard Conway Morris suggest that "Dinosauroids" and humans could have coexisted in the video from the BBC program.
But the biochemical differences give them a renal system that is overall far more effective. Their kidneys are simpler because they're much, much better. We had to develop a lot of sophisticated kludges to cope with the fundamental flaw in our renal system's design.
But surely the fact that Placentals had endured 'harsher selection pressures' is why they were better then the marsupials? The outcome was determined by the placentals being better adapted, which agrees with he selectioniost arguments made by Conway-Morris.