I also greatly appreciate that evograd brought this paper to our attention. Over the next few weeks, I would like to start a discussion here of one of the deeper issues raised by the paper, namely, how to model separate ancestry, in a fashion that the model is not simply an easily-knocked-over straw man or sand bottom clown. Only within the past decade have I begun to see attempts (e.g., by Douglas Theobald, Elliott Sober and Mike Steel, David Penny [e.g., the paper under discussion], David Baum, and a few others) seriously to describe separate ancestry scenarios, so that the evidence can be evaluated in the light of competing historical geometries.
Problem is, nearly all of the modeling of separate ancestry (SA) is being done by authors who think SA is false, and who (perhaps) for various reasons want SA to be false. Genuine advocates of SA, such as YECs and OECs, and some IDers, generally speaking aren’t welcome in mainstream ev bio journals. SA models sketched out by evolutionary biologists therefore tend to be “well, this is good enough, seems reasonable” with the outcome that satisfactory (i.e., thorough and reliable) theorizing tends to suffer. Of course, pro-SA authors are cheering for their preferred outcomes, too – but the competition should occur with the best models on both sides being weighed in the light of the greatest amount of data possible.
So there is a philosophical or analytical imbalance in scientific discussions of this very important question. We can remedy that here, in the short term, although of course it would be best for the topic to migrate into the refereed literature at some point soon.
More work travel for me over the next 4 days, but contacting B. Zhong to obtain his sequence alignment files remains a top priority. Back to this discussion soon. Best to all.
I think this is an opportunity for scientists who differ in opinion. You could publish critiques of such tests and how the methodology is falsified.
This will move the ball forward towards a resolution.
That wasn’t what I was suggesting. If the authors already possessed a particular well-supported phylogeny, there would be little point in generating another one: the history in question would be seen as known or well-established.
Rather, the authors assumed that some tree (i.e., branching pattern of common ancestry) unites the taxa in question. The research challenge therefore is to find the true, or best, tree of relationships. But the geometry or topology they are seeking will (they take this as given) resemble a branching pattern with a single root. Data that do not yield such a pattern will be seen as “phylogenetically uninformative.”
Only very rarely (see my other most recent post in this thread) do authors ask, “Well, IS there a tree here, or some other geometry?”
Yup! Winston Ewert’s dependency graph hypothesis is a good first step in that direction.
But there is little (or zero) chance of the discussion taking place in the peer-reviewed literature, where it should be happening, as long as the authors who want to explore separate ancestry are anathematized as “creationists” or “intelligent design bad guys.” The mainstream biological community must overcome its hostility to dissent, in order to pursue these very interesting questions in the depth and rigor they deserve.
In cases where there is some obvious reticulation, for example, it’s evidenced by quite a bit more than “there are some heterogenous GC sites that produce inconsistent trees”. In other words, I think there can be good evidence for a non-tree-like phylogeny, but the sequences removed from the analysis in that paper wouldn’t cut it.
Some questions raised by IDers are valid, but surely you understand why evolutionary biologists are often “hostile” to IDers? Don’t you think ID has done a pretty good job of destroying most of its credbility over the years?
If you come to my door every day for a year and punch me in the face, and then one day you come to the door and ask me to discuss an interesting philosophical question, would you be surprised that I don’t immediately invite you in for tea and biscuits? I think ID needs to take a lot of steps to build up the many bridges it’s burnt if it wants to suddenly have a seat at the table.
Todd Wood seems to be working to create a model for separate ancestry (SA), as are any number of baraminologists. I think it’s tenuous work at this time, more in the data accumulation and “initial thinking” stage rather than the ready-to-test model creation point. Are there many SA models in the ‘still viable’, ‘ready to evaluate’ stage?
I haven’t thought about it more since my last comment (#40), which I don’t think you’ve replied to. I still don’t see exactly how a seperate ancestry model could explain the data (without adding in lots of untestable auxillary assumptions).
So, the authors do not meantion models that have been put forward. The model they propose for special creation does not take functional constraints on sequences into account. They reason semantically and logically about this, but do not actually simulate models.
At the very least, they should:
Simulate the models that have been proposed (and there are several)
Propose and test some obvious ones that might falsify their test (and there some obvious ones).
Instead, they make the false claim that there have been no models proposed. That is just not true. So, they end up arguing against a straw man. The fact of the matter is that there are some special creation models that work just fine to explain the data, because they basically mirror common descent.
The real question is to delimit which special creation models appear ruled out by data, and which ones are not. You can’t treat them all as one class, and decry that there are no models to test. There, in fact, are models to test, and a clever scientist would want to find models that would break their test. So in the end, there appear to be major validation flaws in this study. It might be a valid test, but we cannot tell from the work they have done, and its significance is surely overstated.
This does not unsettle our confidence in common descent. There is much stronger evidence elsewhere. This is study, in my opinion, does not succeed. Other evidence, however, is very strong.
They imply that no useful (read: testable) predictive models have been proposed by ID, and I think that’s true. I’ve never seen any kind of rigerously defined seperate ancestry model that would make predictions about this kind of data from IDers or creationists.
I don’t think they really overstate their results. They make it clear that they’re comparing the evolutionary model to a non-evolutionary (seperate ancestry) model, and say:
This clearly does not ‘prove’ that yet unknown models are impossible, but the theory of evolution leads to extremely strong predictions, and so the onus is now on others to propose testable alternatives.
The question of how functional space interacts with the sequences to produce different shapes of explored sequence space is an interesting one, but it seems unlikely to me that it would produce the same results. That’s one testable alternative, so it would be great if someone could investigate that, but as I say, I won’t hold my breath that it’s a good alternative.
I think so too. ReMine’s model really wasn’t predictive. Baraminology (aka Kinds, Discontinuity systematics) models aren’t sufficiently well described to understand the biochemical similarities between Kinds/Baramins.
I’m not sure what sort of functional constraints one can assume with special creation. However, the ‘protein space’ accessible via common descent is very likely much smaller than one accessible to a creator. Denton’s model adds a constraint about physical laws ‘attracting’ toward particular sequences, but his also assumes common descent.
This might not be true. There was a recent paper published by Vattay,Kauffman and co that finds that many proteins exist at the “critical point of the metal-insulator transition” and this might be the hall marks of a universal mechanism of charge transport in living organisms. If this is true, then the protein space becomes much more smaller. I am atttching an article on this.
The more “fine tuned” a molecule/eco system is, the smaller the set of possible variants to choose from.
We cant say that for certain… For example, in our experience DNA/RNA seems to be a single/limited solution to a problem. We don’t know if other solutions exist.Perhaps they do, perhaps they don’t… However we can say that the no: of solutions will be more limited as the functional requirements increase.
Depends on how you define accessible. Design involves constrains set by the purpose of the designer. For example, when designing a car, the designer has access to many kind of chasis. However, due to constraints of economy (availability of spares etc), this is reduced to a very small no: of options. In this case one constraint of economy could be to achieve maximum diversity which can be supported by one planet. Design of systems involves optimization to an environment
The only way to tell what is accessible to a creator is by knowing the constraints involved. We don’t even know all the biological constraints, forget constraints of purpose.
This is an example of “scientific arguments” addressing things beyond the scope of science.
No personal experience… only arguments for common design by scientists based on the universal genetic code (except fro a few exceptions which prove the law)… I dont see any reason to disbelieve them.
If this is false. i would like to know more.
Edit: Lie changed to false to avoid speculations on intention.