I fully admit that a supernatural deity could create life so that it all falls into a nested set of clades. However, we have this thing called parsimony.
Most people would argue that the ability to predict planetary orbits using Newton’s equations are strong evidence for the existence of gravity. However, would we be justified in saying that gravity may not exist, and that the evidence is also consistent with gravity fairies moving planets about in a way that exactly mimics our equations of gravity?
Closer. I am not even discussing a possible cause. All I am discussing is the range of graph structures that can produce the evidence we see. This range is much broader than only an evolutionay tree
If anything, we are talking about reverse engineering. You are looking to see what intelligent design would have to do in order to produce results that look like evolution. Compare this to the theory of evolution which predicts a nested hierarchy from first principles. If all we had was knowledge of mechanisms involved in evolution and genetics we would predict that we should see a nested hierarchy before comparing any species. This isn’t the case with intelligent design. There is absolutely no reason one would argue for a nested hierarchy based on the first principles of intelligent design. That’s the big difference here.
one might not go out of their way to look for a nested hierarchy, but it seems plausible most intelligently designed artifacts will form nested clades when forced to do so
the real question is whether the evolutionary data is significantly more cladish than intelligently designed artifacts
and at least my analysis shows id artifacts are even more cladish than evolved artifacts
thus insofar as my analysis reproduces what professionals do, what they should actually be harping on is how badly the data fits a clade structure in order to support evolution
in other words, the perfectly nested clades actually appears to be better evidence for ID than it is for evolution
I see several problems with your analysis. Take my feedback with a large scoop of salt, of course, because I am neither a biologist nor a phylogeneticist.
(1) Bayesian priors. Biologists have observed well-established nested clades at the genus and family level that have diverged over relatively short periods of time. They therefore have a strong Bayesian prior that nested clades explain the ancestry of divergent forms over longer periods of time.
(2) Sequence data. The vast majority of the cladistic work in biology has been done with sequence data. Your critique, based as it is on gene accretions rather than sequence data, therefore fails to land any blows.
(3) Inapplicability of DAGs to existing genomic data. The DAG model relies implicitly on complete and error-free gene catalogs. The gene catalogs in the real world are neither. In the absence of suitable data, the DAG model is not ripe for consideration.
(4) DAGs are susceptible to overfitting. The universe of candidate DAGs is extremely large even for very limited datasets. Trees are in fact a tiny subset of all the possible DAGs. If DAGs provide several orders of magnitude more models to consider than trees, then the fact that I might find a better fit from the DAG pool than from the tree pool is quite likely an artifact of DAGs’ having more parameters to play with.
I don’t think this is quite right. The trees are built using sequence data, but phylogenetic signal is measured based on following discrete traits along the tree, and a gene can be one of these traits. So my simulation builds the kind of trees that are analyzed for phylogenetic signal.
Please correct me if wrong, this is what I’ve been able to gather from my investigation so far.
They are some good points, but still everyone is ignoring the glaring problem. Regardless of how wonderful or not my own analysis is, someone has to do it to even say the phylogenetic signal is evidence for evolution. Without my analysis, you all are stuck up a creek! The big circular question begging creek.
Klax
(The only thing that matters is faith expressed in love.)
84
Evolution needs no more evidence than it had two centuries and in fact over two millennia ago. It’s common sense.
Could you explain why we would expect a nested hierarchy and only a nested hierarchy from intelligent design?
The question remains if your analysis is legit. It would seem better to use the methods other scientists use for determining phylogenetic signal.
It would also be interesting if you used something like automobiles as your ID artifacts. You could use traits like turbos, brand of tire, number of cylinders, and so on.
I appreciate your irenic approach to the discussion, and I hope you feel I have reciprocated. You do raise an interesting question from the frequentist perspective. Given the relative paucity of data for evaluating the virtually infinite number of models that could be considered, I think the biologists are on solid ground by (implicitly, perhaps) taking a Bayesian approach.
i am pretty sure their bayesian approach is merely circular argument by another name
you and the other interlocutors continually miss the fundamental point
yes, a tree is a more specific graph than a general DAG
BUT the problem is any data can be forced into a tree
that’s the fundamental problem
even @T_aquaticus admits that any data can be forced into a tree
so, the corrective action is supposedly to measure the phylogenetic signal
BUT from what I have seen so far, the phylogenetic signal is not actually a signal
it cannot signify whether the data is actually a good fit for a tree like history
it just regurgitates the fact we can fit any data to a tree
case in point is my demonstration that random data has a better phylogenetic signal than evolved data
i think i need to see indication that professional scientists understand this issue before i can take seriously the nested clades argument. otherwise, the fact that i an amateur can see a fundamental flaw in evolutionary theory that professionals cannot indicates something is rotten in the state of evolutionary theory, and casts the whole field in a bad light
i don’t believe i have a YEC or creationist bias. i think i am coming at this issue pretty objectively, and am happy to be proven wrong
