Geographic Clustering of Late Pleistocene Morphotypes in the Denisovan Corridor: A Computational Analysis

Title: Geographic Clustering of Late Pleistocene Morphotypes in the Denisovan Corridor: A Computational Analysis Description: This study presents a computational analysis testing whether certain mythological motifs preserve fragmented cultural memories of interactions with non-sapiens hominin species. It tests the hypothesis that motifs with high morphological specificity cluster in regions with documented Late Pleistocene hominin overlap (the Denisovan corridor) at rates significantly higher than in temporal/isolation control regions. Methods. A corpus of 3,743 raw mythological entries from the Denisovan corridor and 2,215 entries from control regions (Australia, San/Bushmen) was assembled from public-domain folklore sources and the Berezkin thematic catalog. Non-English entries were translated into English via an API Large Language Model to ensure compatibility with English-based regular expression extraction. A multi-stage pipeline (v5.1 filter + v1.8 analyzer) was applied to identify and classify high-quality candidates. A behavioral null model was run on six non-mythological texts (1,597 records) to validate pipeline specificity. Results. The pipeline identified 100 high-quality candidates in the experimental corpus (2.67% yield) versus 19 in the control corpus (0.86% yield). This difference is highly significant (z = 7.24, p < 0.000001, Cohen's d = 1.79). Morphotype distributions differed significantly between corpora (χ² = 25.13, p = 0.0007, Cramér's V = 0.28). The control corpus yielded zero Type III (Archaic Dentition) and zero Type V (Cyclopean) candidates (Fisher's exact p = 0.041), a result referred to in the paper as the "Three Zeros." The null model returned zero candidates from 1,597 records (100% rejection rate), confirming pipeline specificity. A targeted search for a Super-Archaic Eurasian Ghost Hominin (SAEGH) signal returned a null result, reported as a documented boundary condition rather than a failure. A distinct Type III signal emerged in Denisovan zones, interpreted in the paper as consistent with a "Second-Hand Memory" hypothesis. A post-hoc relational analysis isolated a cross-culturally validated ectoparasite/grooming motif, presented as an exploratory finding requiring independent validation. Conclusion. The data appears to support the geographic clustering of possible Late Pleistocene hominin signals in the Denisovan corridor and are inconsistent with the hypothesis that these motifs are a universal human cognitive archetype. The SAEGH null result establishes a provisional boundary for deep-time folkloric memory. The grooming motif offers a novel, testable framework for encoded survival protocols. Contents of this record. The manuscript (PDF), the trait matrices for both corpora (CSV), the analysis reports (Markdown), the extraction and classification scripts (Python), the null model corpus list, and the extracted JSONL entries used in the analysis. All data, code, and analysis scripts are provided for replication. Keywords: Hominin Mythology; Collective Memory; Computational Folklore; Denisovan Corridor; Null Model Validation; Paleoanthropology; Digital Humanities. License: Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Related Identifiers: OSF Project — https://osf.io/mdb84/