Derwood on "Baraminology"

I wrote this up a few years ago...

***

"We look at the same data, just not under the same metaphysic as you."

“It’s all metaphysics, and my metaphysic is the best one!”

So the argument goes in this ‘debate’ – the evolution accepter, beholden as they are to the ‘metaphysic’ of Naturalistic materialism, cannot see how the creationist metaphysic – supernaturalistic antimaterialism, is far superior. It is, after all, premised on Scripture, and Scripture is Inspired.

And so, if one views evidence, in the words of Henry Morris, “the right way” – that is, through the eyes of the creationist metaphysic – one will see the Truth of the creationist claims.

Well, let’s take a look at this creationist metaphysic in action. I will let the objective, rational reader determine if this metaphysic is the superior one when dealing with issues scientific…

When I was a graduate student working on molecular phylogenetics, I discovered a series of articles in the Creationist peer-reviewed literature * dealing with the same subject. 

The authors of these articles were applying computer algorithms to molecular data to determine the relationships between creatures that descended from the ‘kinds’ that were Created and were later allowed to live on the ark.

These and other papers lay out the creationist version of systematics, called Baraminology (or Discontinuity Systematics), which utilize standard computer programs and reproducible analyses using molecular data. These ‘baraminologists’ have set up an entire field of study, complete with its own bible-based terminology and concepts.

The first paper, “A Mitochondrial DNA Analysis of the Testudine Apobaramin,” 1997, DA Robinson, CRSQ 33:4 p. 262-272, examines the relationships between turtles, and establishes or at least lays out some important criteria for establishing affinity of species (baramina) – patterns of mutation bias, gaps between ingroup and outgroups, topological congruence of cladograms using differing parameters and analyses, and strong bootstrap support for the arrangements. The author was able to determine using these methods – which are essentially the same as those used by systematists – that all turtles are related via descent from a created kind, but could not resolve lower-level relationships.

The third paper dealt with cat phylogeny, and just expanded on earlier ‘proof of concept’ papers.

But the second paper was of great interest to me.

“A Quantitative Approach to Baraminology With Examples from the Catarrhine Primates,” 1998, D. Ashley Robinson and David P. Cavanaugh, CRSQ 34:4 p. 196-208, was the very subject I was working on.

Much of the paper consists of quoting/referring to Scripture, which is odd for a scientific paper but not, I assume, for a scientific paper premised on the supernaturalistic metaphysic, and outlining their justification for their “baraminic distance” criterion. This takes up about the first 4 pages. The baraminic distance is essentially equivalent to the materialistic genetic distance measure, it is just called something else.

Those pages are, save for the references to Scripture, well written and exhibit a great deal of thought. The paper gets interesting, however, when we get to the Materials and Methods section on p. 201. The title of the paper and several sentences in the introductory portion indicate that the interest here is in the Old World monkeys, not the human-ape question. Indeed, they discount that question altogether:

“Since Scriptures clearly imply that humans were specially created (Genesis 1:26-272 , 22), and thus phylogenetically distinct from other organisms, we utilize the human-nonhuman primate relationship as a control.”

This will be of interest later.

Their data consisted of 12s rRNA gene sequences, chromosomal characters, morphological characters, and ecological characters. The data were analyzed individually and as a total evidence dataset using standard phylogenetic analysis software. 

It is the results and discussion in which the metaphysic of supernaturalism comes into play.

For those of you that do not know, when you set up a data matrix for analysis you utilize what is called an outgroup – a taxon that is not closely related to the group under study – for use as a ‘yardstick’ of sorts. For example, when analyzing primates you might use rabbit as an outgroup. Interestingly, as quoted above, the baraminologists use human as the outgroup in their analyses.

Outgroups must be designated prior to running the analysis, or the results will appear strange. If you designate the wrong taxon as the outgroup, your results will be strange indeed (you can, of course, run analyses without an outgroup, but these analyses were not utilized by the baraminologists).

So, when the baraminologists ran neighbor joining analyses on the data, they used human as the outgroup. NJ methods assume a constant rate of evolution, which is not indicated by either fossil or molecular evidence and so has fallen out of favor. Though they do not specifically state that they designated human as outgroup, this is what must have happened. This is because the order of the taxa in the dataset can influence the arrangement produced in NJ analyses. For example, I analyzed one of my datasets and I got an arrangement similar to the one seen in the CRSQ paper. Human is first in that dataset, so I cut and pasted it last, re-ran the analysis, and Human got stuck somewhere in the middle of the cluster (however, when I ran a bootstrap analysis, human grouped with chimp). However, when I designated a new world monkey as outgroup, I got the ‘accepted’ arrangement – human + chimp. Making human the outgroup produces an arrangement similar to the one in the CRSQ paper – NJ analyses by default use the first taxon as the outgroup unless designated otherwise.

And what follows from that is the production of weakly supported topologies, since they tried to force the data to conform to a ‘non-natural’ topology. The node linking chimps and gorillas was supported with only 53% bootstrap support. That is fairly low. In a paper not constrained by the antimaterialism metaphysic, in which human is not the outgroup, chimps join gorilla with 96-100% support, depending on the data used. Forcing the data to fit a preconceived notion based on a metaphysic produces statistically significant error.

They mention in the abstract “We have found that baraminic distances based on hemoglobin amino acid sequences, 12S-rRNA sequences, and chromosomal data were largely ineffective for identifying the Human holobaramin. Baraminic distances based on ecological and morphological characters, however, were quite reliable for distinguishing humans from nonhuman primates.”

The description of the morphological analysis sounds impressive – 43 characters. The morphological characters, however, I believe, were specifically selected to produce the desired results. Why do I say this? Because this paper:

Mol Phylogenet Evol. 1996 Feb; 5(1): 102-54. Primate phylogeny: morphological vs. molecular results. Shoshani J, Groves CP, Simons EL, Gunnell GF.**

Was known to the authors. It contained an analysis of not 43 characters, but 264, and this analysis grouped human with chimp.

The other data, ecological data, is the most subjective and should produce no surprise when it was this data that provided the baraminologists their ‘strongest evidence' for a separate human baramin. And what were some of these data? Things like percent foliage in diet, monogamy, population group size and density, home range size, etc. It looks to me like these data too were chosen to produce a desired outcome, for what exactly does “monogamy” have to do with descent?

Indeed, the authors state in their Discussion section:

“Character selection, not the method of analysis, is expected to be the primary factor affecting baraminic hypotheses. False conclusions can be reached unless baraminically informative data has been sampled. Since we have no a priori knowledge regarding which characters are more reliable for identifying holobaramins, it is important to evaluate the reliability of a wide variety of biological data for inferring baraminic relationships.”

And later:

“it is interesting to note that the ecological and morphological criteria were the most adept at distinguishing humans and the most highly correlated, indicating that the datasets in the strongest agreement were the most reliable.”

Yes, that is interesting – the most subjective and limited criteria are the most reliable for giving the creationist the arrangement they want…

That is, they have to pick data that give them the results they want – those that conform to Scripture.

Creationism’s metaphysic in action…

What I did not mention is this, from the section on selecting characters:

“With the exception of the Scriptural criterion no single data set is sufficient to define the holobaramin.”

Translation: Scripture gives us the answers, we need to find the data that will conform to these answers.

The ‘superior’ metaphysic in action.

*I had contacted the authors of this paper in 1999 asking for reprints and neither replied to my requests. I had to buy the issues from CRSQ. Later, after reading in the paper that the data sets were available from the authors on request, I sent an IM to DA Robinson while online one day. First he pretended not to know what I was talking about. After he acknowledged co-authoring the paper, he said something that astounded me – he said that he didn’t think the data sets even existed anymore!