Observability and pharmacological calibration in psychoactive ethnobotany: ceremony duration, admixture selection, and the discovery of ayahuasca

Traditional psychoactive plant ceremonies across 11 indigenous traditions on five continents show a near-perfect log–log correlation (r = 0.977) between ceremony duration and the pharmacokinetic duration of the active compounds, spanning seven pharmacological classes including DMT, mescaline, psilocybin, ergolines, salvinorin, ibogaine, and GABA-A-active compounds. A catalogue of 118 Amazonian admixture plants reveals that pharmacologically active species cluster at the extremes of purpose observability, while candidate plants concentrate in the middle. Agent-based simulations indicate that guided iterative search, not random trial-and-error, could realistically discover the DMT plus MAO-I combination within centuries to millennia in the Amazonian flora.

Study at a glance

Abstract

This Zenodo record holds the preprint manuscript, dataset, analysis code, ancillary process-archaeology catalogues, and the full pre-submission audit trail for an ethnobotanical synthesis study evaluating when traditional psychoactive plant knowledge tracks pharmacological reality. The paper addresses three quantitative tests of psychoactive ethnobotanical claims: (i) Cross-cultural comparison of ceremony duration versus pharmacokinetic effect duration across n = 11 independent indigenous traditions spanning five continents and seven pharmacological classes at the receptor-system level (DMT-family tryptamines including oral and insufflated routes; mescaline-type phenethylamines; psilocybin / 4-hydroxy tryptamines; ergoline lysergamides; salvinorin κ-opioid agonists; ibogaine NMDA / σ / multi-target ligands; GABA-A-active compounds including kavalactones and muscimol). Ceremony duration tracks pharmacokinetic duration log–log with Pearson r = 0.977, p = 2.5 × 10-7, n = 11; slope 1.010; 95% CI for r [0.910, 0.994]. (ii) A 118-plant Amazonian admixture catalogue assembled from 70 years of ethnobotanical documentation (Schultes 1957, Luna 1986, Ott 1994, contemporary mestizo ethnographic sources), cross-classified by purpose observability (Observable / Mixed / Non-observable) and pharmacological validity (Active n = 12 / Candidate n = 106). Active admixture plants cluster at the extremes of the purpose-observability distribution while candidate plants concentrate in the middle (χ2(2) = 20.17, p = 4.16 × 10-5), a bimodal rather than monotonic pattern. (iii) An agent-based simulation comparing six search strategies for the DMT + MAO-I combination against a 1600-candidate Amazonian flora baseline. Guided iterative search converges on the target in centuries to millennia under realistic Amazonian search-space parameters; pure random trial-and-error fails within archaeologically plausible budgets. What changed from v6 (2026-05-11) to v7 (2026-05-21): v7 incorporates a full pre-submission audit pipeline. The canonical dataset ceremony_pharmacokinetics.json was patched to correct three wrong DOIs (Dinis-Oliveira mescaline review DOI/venue; Hasler 2004 psilocybin DOI that previously resolved to an unrelated smoking-cues paper in the same Psychopharmacology volume; MacLean 2013 salvinorin DOI + venue) and two citation-bundling defects (Griffiths 2006 + Johnson 2008 disambiguated for clinical psilocybin; Aporosa 2014 promoted over Aporosa 2022 as the primary load-bearing citation for Fijian yaqona). The manuscript received the F1 fix (pharmacological-class enumeration reframed at the receptor-system level, n = 7), the abstract was trimmed to ERA's 250-word limit, the AI declaration was expanded to cover analysis/figure-generation code and the audit pipeline, and an administration-route count precision fix was applied. A multilingual citation-verification audit reversal restored the McKenna, Luna & Towers 1986 América Indígena citation (Spanish-language original, MAPS bibliography entry 14323) after an earlier 1986→1995 correction was found to be based on incomplete English-language indexing. The v7 deposit also includes a full audit-trail subdirectory documenting the dataset audit (8-phase pre-deposit gate), dataset patch, manuscript F1 fix, the McKenna reversal, and the Option Y ancillary audit (8 files, 269 entries, all citation defects surfaced and resolved through the audit pipeline). What's in the bundle: a single zip (psychoactive_era_v7_bundle_2026-05-21.zip) containing the manuscript, dataset, ancillary process-archaeology catalogues, analysis code, and audit trail. The manuscript (manuscript_era_v1.md) and a readme (README.md) are also uploaded as standalone files for direct viewing. Reproducibility: the compute scripts in code/ read JSON inputs from dataset/ (relative paths); all headline numbers in the manuscript are produced by these scripts against the JSON inputs in this bundle. Audit-trail tag chain (in the source git repository): psychoactive-dataset-audit-2026-05-20 → psychoactive-dataset-patched-2026-05-20 → psychoactive-manuscript-f1-fixed-2026-05-20 → psychoactive-final-audit-2026-05-20 → psychoactive-mckenna-1995-fixed-2026-05-21 (the initial fix that was subsequently reversed) → psychoactive-mckenna-1986-restored-2026-05-21 (the canonical-citation reversal). The pipeline reversal is documented in audit/PROTOCOL_HARDENING_LESSONS.md. This is the version cited from the manuscript's Data Availability statement. Earlier versions remain accessible via the concept DOI for historical reference. Author: Elliot Allan, Deep Time Research Institute (deeptime-research.org). ORCID: 0009-0008-8541-0944. Manuscript currently in submission to Ethnobotany Research and Applications.