Introduction: The Paper’s Claim vs. The Critical Question
The study “Using machine learning to classify extant apes and interpret the dental morphology of the chimpanzee-human last common ancestor” claims that Miocene hominoid fossils share dental proportions most similar to gorillas, suggesting the chimpanzee-human last common ancestor (LCA) had gorilla-like teeth. The authors propose this as evidence that humans and chimpanzees evolved their similar dental proportions independently through “parallel evolution.”
But does identifying shared dental ratios between fossils and living apes actually demonstrate gradual evolutionary transformation? Or does this pattern better align with a biological reality where organisms reuse optimized designs across distinct lineages?
Critical Analysis: Reassessing the Evidence
Finding 1: Machine Learning Classifies Fossils by Dental Ratios
The study reports that Miocene hominoids (20–5 million years old) cluster with gorillas in “molar module component” (MMC) and “premolar-molar module” (PMM) ratios, while later hominins like Homo erectus align more closely with humans. The authors interpret this as evidence that the LCA had gorilla-like teeth, with humans and chimps later converging on similar proportions.
Critical Lens:
This analysis assumes without justification that similarity in dental proportions must result from evolutionary descent. Yet the same data could equally support a model where a common engineering principle governs tooth-size relationships in primates. For example, the MMC ratio (M3/M1) could reflect biomechanical constraints for efficient mastication—a functional requirement that a designer might implement across multiple species. The “convergence” narrative relies entirely on the prior assumption of universal common ancestry, which the data neither confirm nor disprove.
Finding 2: High Classification Accuracy for Extant Apes
Supervised learning algorithms achieved >92% accuracy in classifying living apes using linear dental measurements. When applied to fossils, many Miocene specimens were classified as Pan (chimpanzee) using linear metrics but as Gorilla using MMC/PMM ratios.
Critical Lens:
The paper conflates classification with explanation. Machine learning identifies patterns but cannot determine whether those patterns arose through mutation/selection or reflect intentional engineering. The fact that algorithms group fossils with living apes based on dental proportions simply shows that certain ratios are statistically associated with specific taxa—not that those taxa evolved from one another. Notably, the study admits that “similarities in postcanine tooth proportions in extant Pan and Homo dentitions are the result of parallel evolution“—a concession that undermines the very tree-of-life model it seeks to support.
The Bigger Picture: Why Dental Ratios Don’t Build Body Plans
Even if the LCA had gorilla-like teeth, this tells us nothing about how gorilla, human, or chimpanzee body plans originated. Dental proportions are modifications of pre-existing structures, not innovations. The study’s key metric (MMC) measures the relative size of molars—a tweak to an existing system. Such changes align with observed adaptive fine-tuning but fall far short of explaining the origin of teeth themselves, let alone the genetic and developmental programs that build jaws, nerves, and enamel-forming cells.
As biologist Michael Denton notes in Evolution: Still a Theory in Crisis, “The existence of a graded series of forms… does not demonstrate that the series is the result of an evolutionary continuum.” The dental patterns here could reflect a designer reusing effective biomechanical solutions across distinct primate kinds.
Broader Context: The Limits of Similarity
The study’s evolutionary interpretation hinges on the unproven assumption that similarity equals ancestry. Yet this logic struggles to explain why bats and birds (both fly) or octopuses and humans (both camera-type eyes) share analogous structures without common descent. Philosopher of biology Paul Nelson observes that “biological similarity is often a puzzle for Darwinism, not a confirmation,” as it frequently crosses proposed evolutionary boundaries.
The Bottom Line: Patterns ≠ Process
If microbes-to-man evolution were true, we’d expect fossils to document gradual transitions between major body plans. Instead, this paper highlights a persistent pattern: organisms vary within bounds, reusing and recombining existing structures. The dental proportions of Miocene hominoids—like the genetic “similarities” between humans and chimps—reveal a biological reality of shared design principles, not a creative mechanism capable of building life’s complexity from scratch.
Paper Reference:
Using machine learning to classify extant apes and interpret the dental morphology of the chimpanzee-human last common ancestor
Authors: Tesla A. Monson, David W. Armitage, Leslea J. Hlusko
Abstract:
“Machine learning is a formidable tool for pattern recognition in large datasets. We developed and expanded on these methods, applying machine learning pattern recognition to a problem in paleoanthropology and evolution… Miocene hominoids are morphologically most similar in dental size and shape to extant chimpanzees. However, relative dental proportions of Miocene hominoids are more similar to extant gorillas and follow a strong trajectory through evolutionary time… similarities in postcanine tooth proportions in extant Pan and Homo postcanine dentitions are the result of parallel evolution.”