other Australopithecus species and sometimes clusters with no other taxa in the distance correlation results. What should we make of these results? First, the inconsistency between clustering methods does not seem unique to the postcranial characters, since we see the same inconsistencies when using the full craniodental set or the high relevance (a ≥ 0.5) craniodental characters. In our previous work, we found more consistency with the full craniodental set and a slightly larger taxon sample, suggesting that the present taxon sample may still be too small to reveal reliable clusters. Ideally, cluster analyses work best when there are well-defined clusters with multiple points in each. In many of the cases we present here, a possible cluster is accompanied by a heterogeneous set of outlying points that do not form a single cluster. This is not ideal for cluster analysis. Ironically, it may be further discoveries of nonhuman ape taxa that help to clarify the boundaries of humanity. Despite these inconsistencies, when we look at the frequency with which taxon pairs cluster together in our partitions, we can see fairly clear trends. There is a group of Homo taxa, excluding Homo floresiensis, that almost always cluster together. This group includes H. naledi, H. sapiens, H. neanderthalensis, H. erectus s.l., and H. heidelbergensis. Not surprisingly, the evidence for cultural sophistication is strongest here (Ross et al. 2023), and we suggest these taxa can be accepted as Near Certain Humans (or certain in the case of Homo sapiens) (NCH). Surprisingly, despite cultural evidence that Homo floresiensis is human (Wise 2005), it clusters with NCH in an average of only 45.5% of the partitions. In past craniodental analyses, we have also seen Homo floresiensis not clustering with other Homo taxa, but since the known character states for H. floresiensis amounted to only 32.2% of the full craniodental character set, we speculated that the paucity of data artifactually separated H. floresiensis from other human taxa. Here, 60.3% of the postcranial characters are scored for H. floresiensis, and we still do not see a close clustering of H. floresiensis with other Homo taxa. We thus must consider the likelihood that H. floresiensis differs anatomically from other humans or might exhibit developmental pathologies, as suggested by Rupe and Sanford (2017). Still, with the combined character set of 482 characters, Homo floresiensis clusters with other Homo taxa in all distance correlation partitions, all medoid partitioning, and in the two-cluster fuzzy analysis. Our postcranial characters generate a greater degree of uncertainty regarding the status of Au. sediba. In the present craniodental analyses and in previous studies Au. sediba has always clustered with Homo taxa (except for H. floresiensis), but with the postcranial characters Au. sediba clusters with no other taxa or with other australopiths. This is not surprising since the postcranial skeleton is known to differ considerably from NCH like H. sapiens, H. erectus s.l., and H. neanderthalensis. With 72.4% of Au. sediba’s postcranial character states scored, this clustering is unlikely to be a statistical artifact. With the combined craniodental and postcranial characters, Au. sediba clusters consistently with Homo taxa again, as it does with the craniodental characters alone. This difference between clustering with postcranial characters versus the combined character set cannot be attributed to a bias towards craniodental characters, because only 49% of Au. sediba’s craniodental characters are scored. If anything, the combined character set should reflect a bias towards postcranial Gorilla Chimpanzee Ardipithecus Au. afarensis Au. africanus s.l. Au. sediba H. floresiensis H. georgicus H. habilis H. naledi H. heidelbergensis H. erectus s.l. Neandertals H. sapiens Gorilla - 100% 100% 34% 27% 5% 9% 0% 0% 0% 0% 0% 0% 0% Chimpanzee 100% - 100% 34% 27% 5% 9% 0% 0% 0% 0% 0% 0% 0% Ardipithecus 100% 100% - 34% 27% 5% 9% 0% 0% 0% 0% 0% 0% 0% Au. afarensis 34% 34% 34% - 89% 34% 30% 18% 21% 21% 18% 18% 18% 18% Au. africanus s.l. 27% 27% 27% 89% - 46% 36% 30% 32% 25% 23% 23% 23% 23% Au. sediba 5% 5% 5% 34% 46% - 55% 64% 77% 68% 55% 59% 55% 55% H. floresiensis 9% 9% 9% 30% 36% 55% - 46% 59% 50% 41% 41% 41% 41% H. georgicus 0% 0% 0% 18% 30% 64% 46% - 84% 75% 82% 86% 82% 82% H. habilis 0% 0% 0% 21% 32% 77% 59% 84% - 91% 75% 80% 75% 75% H. naledi 0% 0% 0% 21% 25% 68% 50% 75% 91% - 84% 89% 84% 84% H. heidelbergensis 0% 0% 0% 18% 23% 55$ 41% 82% 75% 84% - 96% 100% 100% H. erectus s.l. 0% 0% 0% 18% 23% 59% 41% 86% 80% 89% 96% - 96% 96% Neandertals 0% 0% 0% 18% 23% 55% 41% 82% 75% 84% 100% 96% - 100% H. sapiens 0% 0% 0% 18% 23% 55% 41% 82% 75% 84% 100% 96% 100% - Table 2. Frequency of co-clustering. For each pair of taxa, the frequency with which they occur in the same cluster is shown based on 44 different clustering treatments described in this paper. WOOD AND BRUMMEL Hominin Baraminology Reconsidered 2023 ICC 264
RkJQdWJsaXNoZXIy MTM4ODY=