value in the clustering patterns even when the cluster assignment for every taxon might not be reliable. Finally, the HN character subset reminds us that wildly inconsistent clusterings with poor average silhouette widths should not be trusted, and a different character set should be sought. The exceptionally poor results with these characters also support the basic humanity of Neandertals. Most young-age creationist scholars agree that Neandertals are human (Ross et al., in review), but some old earth creationists claim that Neandertals are not human (e.g., Rana and Ross 2015). If that were correct, I would expect that characters that distinguish H. sapiens from Neandertals would be the best at classifying apes and humans, but that is definitely not the case. It must be emphasized again that these results are not properly a test or falsification of character-based essentialism, since we have no reason to believe that any of these characters here tested would or should be essential. Instead, the character subsets evaluated here can be seen as a form of character weighting, where “important” characters were selected using essentialist thinking. In that respect, this procedure has been found wanting. These character subsets repeatedly fail to distinguish human from nonhuman in a reliable fashion, which is an expectation of essentialism. If we are to identify truly essential character traits, we must begin with a more deliberate approach to identifying the sorts of traits that reasonably might be essential to humans. Such an approach must consider theological and biological issues in targeting the most likely essential traits. We must also consider the possibility that the theological essentialism described in Scripture is a systems-based essentialism and will not reduce to specific sets of discrete characteristics. What we can see clearly in this study is that naïve, character-based essentialism, with its quick and easy selection of “essential” traits, is unreliable. ACKNOWLEDGEMENTS This work was sponsored in part by a grant from the Genesis Fund and by donations to Core Academy of Science. I am also grateful to Lucinda Hill and Southern Adventist University for providing access to the cast skeletons of chimpanzee and gorilla in their collection. REFERENCES Ahern, J.C. 2005. Foramen magnum position variation in Pan troglodytes, Plio‒Pleistocene hominids, and recent Homo sapiens: Implications for recognizing the earliest hominids. American Journal of Physical Anthropology 127:267–276. DOI: 10.1002/ajpa.20082. Beaudet, A., R.J. Clarke, L. Bruxelles, K.J. Carlson, R. Crompton, F. de Beer, J. Dhaene, J.L. Heaton, K. Jakata, T. Jashashvili, K. Kuman, J. McClymont, T.R. Pickering, and D. Stratford. 2019. The bony labyrinth of StW 573 (“Little Foot”): Implications for early hominin evolution and paleobiology. Journal of Human Evolution 127:67–80. DOI: 10.1016/j. jhevol.2018.12.002. Biddle, D.A. 2016. Differences between apes and humans. In D.A. Biddle, D.A. Bisbee, and J. Bergman (editors), Debunking Human Evolution Taught in our Schools, pp. 1–31. Genesis Apologetics. Clarke, R. 2013. Australopithecus from Sterkfontein Caves, South Africa. In K.E. Reed, J.G. Fleagle, and R.E. Leakey (editors), The Paleobiology of Australopithecus, pp. 105–123. New York: Springer. DOI: 10.1007/97894-007-5919-0_7. Clarke, R.J. 2019. Excavation, reconstruction and taphonomy of the StW Figure 7. Orthogonal views of the 3D MDS results for the simple matching distances using the LC character subset. or a mix of living and fossil apes (FA characters). The procedure implemented here for defining a set of essential characteristics that can be used to include and exclude taxa from the category “human” does not produce reliable results. The character states are found in nonhuman fossil taxa, or the character states fail to include even the most obviously human fossil forms (Neandertals). Evaluating these essentialism-derived character subsets with clustering methods produce more interesting results. The consistency with which Homo sapiens clusters with other human taxa such as Neandertals and H. erectus seems reassuring, but different clustering methods frequently produce alternative clusterings of relatively equal quality using the same subset of characters and distance metrics, as judged by average silhouette width. Similarly, clustering of relatively high average silhouette width often place obviously nonhuman taxa (particularly Paranthropus robustus) in the same cluster as Homo sapiens. None of the character subsets is immune to these problems. Although these results seem to be entirely negative, there are important conclusions from this work that can inform baraminology studies. First, we can see that even though all character samples are not of equal quality, there is a remarkable consistency to the taxon clustering from these character subsets (except for the deliberately nonsensical HN characters). Even as we must be cautious due to the poor quality of the clustering and the frequency with which nonhuman and human taxa occur in the same cluster, it remains notable that certain taxa always or frequently cluster together. In the larger world of statistical baraminology, we sometimes have little choice but to use character matrices containing relatively few characters, but these results suggest that even that small sample may reveal something of WOOD Essentialism and Human Kind 2023 ICC 100
RkJQdWJsaXNoZXIy MTM4ODY=