Science Saturday: Early Human puzzle

From the Smithsonian Human Origins Program, a noble quest:  In search of what makes us human. Within the site we find the Human Family Tree, shown below.

Note that, despite this being declared our family tree, there are several question marks littered on the diagram. Note, also, the dashed lines that link several of the species together, indicating potential evolutionary links.

If the diagram appears confusing it's because, it is. Upon clicking on the active links one finds that, while the evolutionary biologists agree that the species listed are part of our "family tree", there is disagreement as to just how the family members should be linked. For Homo sapiens, one finds that there are two competing evolutionary models:  the multi-regional hypothesis and the out of Africa hypothesis.

Yet, just how did we get to where we are today?

By looking at the family tree, above, we see that there are at least 3 paths, and possibly 4, backwards from Homo sapiens. Why? From Who Was Adam?, by Rana & Ross,

Until a few years ago, paleoanthropologists primarily debated whether it was Australopithecus africanus or Australopithecus afarensis [the two long green bars in the chart above] that morphed into Homo...

This debate became more complex in 1999 when an international team of scientists reported on the discovery of a new hominid species, Australopithecus garhi, dated at about 2.5 million years ago. Because A. garhi may have used tools to butcher animal carcasses, some researchers have proposed that this australopithecine gave rise to Homo. Others dispute this interpretation and consider A. garhi an evolutionary dead end. [note that A. garhi, above, has one of the question marks]

These minor controversies pale in comparison, however, to those recently inaugurated with the discoveries of Kenyanthropus platypops [yellow in the chart above] and Sahelanthropus tchadensis [red in the chart above]. Instead of these new hominid discoveries bringing clarification and resolution, they throw the entire field into even greater chaos. If Kenyanthropus is indeed a legitimate homind taxon, then neither Australopithecus africanus nor Australopithecus afarensis led to humanity - Kenyanthropus did... The discovery of Kenyanthropus potentially invalidates all traditional hominid evolutionary trees found in textbooks.

To be sure, evolutionary biologists would simply indicate that this is how science works - by continuing to gather data and learning from that data, sometimes revising previously held theories.

Yet, as Rana & Ross state,

To declare human evolution a scientific fact, paleoanthropologists must do much more than reliably define evolutionary relationships among the hominids in the fossil record. They must also establish an evolutionary pathway that compels widespread acceptance among biologists.

Despite the lack of compelling evidence for human evolution, and despite the fact that there is widespread disagreement with regards to a supposed human evolutionary pathway, we're presented with a "human family tree". I would say that a field in which a single fossil find causes so much controversy is a field which is not robust enough to present such a diagram as anything more than speculation.

Comments

[LotharBot]

Children inherit DNA from two parents, four grandparents, eight great grandparents, and so on. Different pieces of your DNA came through different paths. What that means for evolutionary theory is that there isn't, and shouldn't be, one "master" tree. There are, rather, many trees which combine to make a mostly tree-like graph. The only way to get a nice, clean tree is for enough time to pass that most of the DNA variants coalesce into one. If you end up with an actual tree, what it really means is you've lost the data that points to the far more complex structure. In that sense, wacky structures like you see above -- triangles, branches that reconnect, and dead ends -- are all expected in a good evolutionary "tree".

Now, in the case of early humans, the real problem is that the tree isn't being constructed from DNA at all. It's being constructed from small numbers of fossils and large numbers of assumptions. That set of data is not robust enough to construct a good, consensus tree. There's barely enough data to make haphazard guesses, which is why you get all the question marks and arguments among scholars.

Of course, this won't stop people like our friend BobC* from bashing people over the head with the data and declaring it as certain. And it won't stop certain creationists from citing the sloppy data here as showing a systematic problem in all of evolutionary theory. Neither conclusion is correct. The correct conclusion to take from this is that the data is sketchy when it comes to recent human ancestors, and it is therefore unsafe to draw conclusions from it.

* might I be permitted to gloat just a little longer? I just love scoring a knockout on an internet bully.

[Rusty]

LotharBot,

might I be permitted to gloat just a little longer? I just love scoring a knockout on an internet bully.

You're a funny guy.

Yes, in a sense, the family tree would spread out as one goes back in time. However, if one is positing that the human race came from a small original set (an Adam & Eve), then wouldn't one expect to find the family tree indicate a clear and/or limited pathway? It seems to me that natural process evolutionists assume that human evolution occurred and simply arrange potential ancestors based their genetic similarity with our own. If, indeed, the multiple primate fossils are part of our family tree, then what tests are available to determine which one of them contributed most to our development? (other than genetic similarity combined with temporal and geographical prerequisites?)

[LotharBot]

I don't think any scientists expect to find a clear pathway. On occasion, someone will claim to have discovered the genetic Adam or the genetic Eve, but that's just a publicity stunt for an otherwise very boring statistical analysis of some DNA. The conclusion isn't really even that we all came from "Adam" or "Eve", only that the DNA in a particular spot from others in the same time period has mostly died out (something that's expected to happen over a certain period of time due to genetic drift.) If you look at some other chunk of DNA on a different chromosome, you expect to find an entirely different "Adam" or "Eve". It's generally understood that it's more likely for a particular block of DNA to die out / consolidate if the population is small, so when people find that X many generations ago was when all of gene Y came from the same person, they assume there was a population bottleneck X many generations ago. For the most part, these assumptions are confirmed from multiple different parts of the DNA.

You're correct that the assumption is that human evolution occurred. The process of arranging potential ancestors is far more complicated than simply finding a percentage match, though. The simplest and least reliable method is called "parsimony". Basically, you grab DNA sequences or lists of physical traits from a bunch of stuff, and you try to create a tree that requires the fewest total number of changes. So you group similar things, and then link groups to other groups, and so on. There's a bunch of statistics involved, and in the end, you say "here's the tree."

There's always uncertainty involved. If you run the algorithm using physical data and again using DNA, some things move around. If you use different parts of the DNA, things move again. That's because mutations are random -- sometimes the same mutation happens multiple times, and you have no way to know for sure which mutations happened multiple times and which happened only once. So you might, for example, assume that wings evolved only once and find that your tree requires you to have a particular lung structure evolve twice, when in reality, wings evolved twice and the lung structure only evolved once.

If you have a really good data set, you tend to have a lot of corroborating evidence -- species that are evolutionarily "close" in one block of DNA tend also to be close in other blocks of DNA. If you have two contradictory pieces of data, there are 30 more that all point the same way. But if your data set is limited, as in the case of early humans, there's a lot more outright guessing involved. If you have two contradictory pieces of data, you just have to guess between them. (Since we're normally not doing DNA analysis on early human fossils, this is even more pronounced, because we can't check to see if fossil X had property Y because of diet, age, genetics, or something else.)

A similar principle actually applies to analyzing Bible manuscripts. You can look at the changes between a series of manuscripts and construct a tree. In some cases, where we have very few manuscripts, it's not clear what the original was. In other cases, we can say with great certainty what the original was and where and when the variants occurred. We can say how one papyrus relates to another, and so on.

The methodology is really quite powerful, provided the data set is robust.

[Rusty]

LotharBot,

With such a complicated process it seems that, given the relative small sampling of primate fossils, the notion that we can reconstruct a human family tree is untenable. At the very least, it illuminates the almost ad hoc assumptions that have to be made with regards to the "fact" of human evolution. One has to wonder just what type of data would be required to even begin building a reliable human family tree?

Or, given an interpretation of Genesis 1 & 2 as describing a literal Adam & Eve (which, I think, is the best harmonization of the account with the entire Bible), what evidence would we expect to find (or have to find) in order to validate such a premise?