Preprint

Preservation Asymmetry in Upper Palaeolithic Cave Ichnology


Footprint Evidence from Sealed Caves in Northwestern Eurasia

Preprint — Version 1 — November 2026

Observatory Research Team and Inferno Labs

DOI: https://doi.org/10.67767/obs.2026.001

We invite readers, researchers, and domain specialists to suggest edits, additional data, or critical review for this preprint. Join the collaborative research!

Abstract

This study synthesizes footprint evidence from nine naturally sealed caves in northwestern Eurasia, preserving ichnological traces dated from approximately 36,000 to 7,000 BP. Data derive primarily from Pastoors and Lenssen-Erz (2021), supplemented by five post-publication studies applying high-resolution 3D photogrammetry and indigenous tracking expertise. Seven of nine caves record children among visitors, including ages as young as three years. Groups consistently comprise fewer than fifteen individuals, demonstrating mixed-age multi-generational compositions rather than institutional gathering formations. Spatial structuring across all sites shows deepest, most inaccessible zones associated with significant activities (art production, clay modeling, funerary deposits), whereas occupation zones exhibit footprint destruction via erosion or repeated disturbance. Where recording windows can be estimated from calibrated radiocarbon dates, visitation frequency ranges from single events to four episodes across periods extending up to fifteen thousand years. A preservation asymmetry emerges: footprints survive preferentially in deep, undisturbed zones where transient exploratory activity occurred, while zones of sustained occupation tend to destroy ichnological evidence.

Keywords: cave ichnology; Upper Palaeolithic; prehistoric footprints; behavioral archaeology; preservation bias; San tracker methodology; multidisciplinary analysis

1. Introduction

Palaeolithic cave use has attracted sustained scholarly attention across archaeology, palaeoanthropology, and cognitive science disciplines [1][2][3][4][5]. Caves function as intensively studied contexts documenting art production, burial practices, symbolic behavior, and habitation, with their material remains fundamentally shaping understandings of early human cognition, social organization, and ritual life [6][7][8]. However, interpretive traditions have predominantly relied on material culture—pigment deposits on walls, ochre fragments, carved bone artifacts, lithic assemblages—while direct records of human bodies moving through these spaces have received comparatively less systematic examination.

Ichnological evidence provides distinct analytical capacity unavailable from art and artifact records: documentation of specific individuals' age estimates, sex determinations, gait characteristics, group sizes, and movement trajectories at discrete temporal moments. Since the 1960s, ichnological investigations have produced empirical records of who was physically present in cave systems, entry counts, behavioral activities during interior occupancy, and visitation frequencies when combined with radiocarbon dating and stratigraphic evidence [1][2][9][5][10][11]. This evidentiary corpus holds considerable value for reconstructing Palaeolithic cave utilization patterns and has expanded substantially following two methodological developments: application of high-resolution 3D photogrammetry to footprint documentation protocols, and integration of indigenous tracking expertise—particularly from San ichnologists—into prehistoric print interpretation frameworks [12][13][14].

This paper synthesizes available footprint evidence from nine naturally sealed caves in northwestern Eurasia, drawing upon the synthetic volume edited by Pastoors and Lenssen-Erz (2021) alongside five post-publication studies revising or extending its findings. This evidentiary corpus supports empirically grounded accounts of Upper Palaeolithic sealed-cave utilization patterns while identifying structural biases within preserved records having implications for evidentiary interpretation methodologies.

2. Dataset Composition and Geographic Distribution

2.1 Nine Sealed Cave Sites

The dataset comprises nine caves where human footprints were preserved through natural sealing mechanisms—entrance collapse, stalagmite growth, or restricted passage access—subsequently preventing disturbance (Table 1). All sites are documented in Pastoors and Lenssen-Erz (2021), though five entries incorporate post-publication studies revising key identifications or adding newly documented data points.

Geographic clustering shows five of nine caves located in southwestern France (Niaux, Fontanet, Tuc d'Audoubert in Ariège; Pech-Merle in Lot; Cussac in Dordogne), with Chauvet (Ardèche) and Aldène (Hérault) positioned outside this southwestern cluster. This distribution reflects intersection of two preservation prerequisites: limestone karst cave formation geology and natural sealing event occurrence, rather than underlying patterns of prehistoric cave utilization intensity.

Temporal range extends from Aurignacian period (Chauvet, approximately 36,000 uncalibrated BP) through Mesolithic period (Aldène, attributed to approximately 10,000–7,000 BP), encompassing Gravettian, Solutrean, Magdalenian, and post-glacial periods. Dating quality exhibits marked variation and should not constitute uniform chronological framework. Nevertheless, dataset spans periods with demonstrably different demographic conditions—from Aurignacian European metapopulation potentially numbering low thousands [15]—through Late Glacial population expansion phases.

No footprints in this dataset receive direct radiometric dating. All chronological attributions derive from associated materials—charcoal deposits, bone fragments, cultural diagnostic elements—attributing footprints to contemporaneous periods through spatial and stratigraphic association. Date formats mix uncalibrated radiocarbon years (BP) and calibrated calendar dates (cal BP) exactly as presented in source publications. When uncalibrated dates appear in usage, actual calendar age exceeds stated values, sometimes substantially exceeding them.

Table 1. Sealed caves dataset (nine sites, chronological order):

Cave Location Approximate Date Subjects Recorded Behaviors
Chauvet Ardèche, France ~36,000–26,000 BP (Aurignacian/Gravettian) 1 boy (~8 yrs) Solo movement through deep cave; torch cleaning stops (charcoal dated ~26,000 BP); art-making in deep chambers; footprints cross bear and wolf tracks [9]
Pech-Merle Lot, France ~25,000 BP 5 individuals; 17 footprints (2 men, 2 women, 1 boy ~9–10 yrs) Transit corridor movement [16] Four moving westward; one woman carrying load moving eastward. Boy steers away from drop-off. Single transit episode only. Rejects Lemozi's (1929)[17] priestess-and-child ceremonial interpretation.
Cussac Dordogne, France 28–31 ka cal BP (Middle Gravettian) 7 footprints; adults wearing soft leather footwear Sole cave documenting footwear use—moccasins [18] 1.6 km gallery containing engravings; human remains positioned in bear hibernation nests. Footwear function as protection or symbolic/social marker remains unknown.
Ojo Guareña Burgos, Spain ~19 ka BP – ~4.3 ka BP; 4 episodes/~15,000 yrs 9–11 individuals (Chalcolithic); 1,200+ prints Round-trip movements documented; individual inspecting side passages; four separate visits spanning ~15,000 years; no association with symbolic spaces in deep sector.
Tuc d'Audoubert Ariège, France 17,200–16,500 cal BP (Magdalenian) 14 subjects (maximum 5 visits; 2–6 per visit) Adult male-female core pair traversing ten locations; male handling bear bones; children segregated in Galerie des Petits Pieds; clay bison deposited in deepest chamber; fully structured spatial utilization pattern.
Basura Liguria, Italy ~12,340 BP ~30 prints; 3 juveniles in Sala dei Misteri: ~3 yrs, ~6 yrs, ~12–13 yrs Crawling locomotion documented; morphometric reanalysis showing exclusively juvenile presence, no adults in chamber [11] Clay balls identified as children's play activity, rejected as ritual deposits per Blanc (1960)[19] original interpretation.
Niaux Ariège, France ~13,000 BP (Magdalenian) Two girls (~9 and ~12 yrs, Pales identification); San trackers identify one child (~12 yrs) Movement through Diverticule des Empreintes; San trackers read as single child moving, not ritual dance performance [5]
Fontanet Ariège, France 12,770 ± 420 BP (Magdalenian IV) 69+ prints; predominantly juveniles (statures 114–142 cm); 1 adult Zone 1 (deep): children's play, sliding, clay throwing activities. Zone 2 (near entrance): art production, hearth usage—footprints destroyed by water runoff.[10] Exceptional preservation including fingerprints, skin creases, fingernails recorded. Canid tracks present alongside human prints. San trackers identified 13 individuals; morphometric count yields 2–6—discrepancy remains unresolved methodological question between approaches.
Aldène Hérault, France Mesolithic (~10,000–7,000 BP) Family group; 400+ prints over 50 m Torch management activities documented; planned speleological exploration with anticipated return trip.


Chapter references: Cussac: Chapter 4 + Ledoux et al. (2021) + Ledoux (2019). Basura: Chapter 14 + Citton et al. (2017)[11]. Pech-Merle/Niaux/Fontanet: Chapter 6 + Pastoors et al. (2016) + Ledoux (2019). Aldène: Chapter 15. Chauvet: Garcia (1999). Tuc d'Audoubert: Chapter 13. Ojo Guareña: Chapter 17.

All data originate from Pastoors and Lenssen-Erz (2021) unless otherwise specifically cited.

3. Multi-Generational Groups and Child Activity Evidence

Seven of nine caves document children among visitors, establishing presence of very young children across majority of sites as consistent pattern. Group compositions remain small (typically fewer than 15 individuals), mixed-age, demonstrating multi-generational family group structures rather than institutional gathering formations.

At Basura, morphometric reanalysis conducted by Citton et al. (2017)[11] revised original Blanc's (1960)[19] identification of five individuals of mixed ages occupying Sala dei Misteri chamber to three juveniles exclusively, with no adults present in that specific room. Youngest subject measured approximately three years old moving through chamber on quadrupedal locomotion. Clay balls adhering to chamber walls, originally interpreted by Blanc as ritual deposit items, now understood as product of children's playful manipulation and recreation activities.

Ledoux (2019)[10] documented 69 or more footprints at Fontanet dominated by juvenile trackmakers, with estimated statures ranging from 114 to 142 cm. Footprint evidence records children moving in multiple directions under low ceiling constraints, sliding on clay surfaces, throwing clay projectiles—patterns representing sustained play behavior rather than directed transitory movement or ceremonial activity formulations.

Single child of approximately eight years moved alone through deep chambers at Chauvet, stopping periodically to clean torch whose charcoal deposits received radiocarbon dating to approximately 26,000 BP [9] At Pech-Merle, San trackers collaborating with Pastoors et al. (2016)[16] identified five individuals originating from seventeen footprints, including boy aged nine to ten years steering trajectory away from dangerous drop-off point. Trackers also identified woman carrying load moving in opposite direction—pattern indicating transit corridor function rather than ceremonial space usage. This revision overturned Lemozi's (1929)[17] reading of identical prints as representing "priestess and child."

Pech-Merle identifications originated from Tracking in Caves project, bringing three ǂKhomani and Ju|'hoansi San trackers—Tsamgao Ciqae, /Ui Kxunta, and Thui Thao, all CyberTracker-certified Master Trackers—into European cave systems to read prehistoric footprints independently. Methodological design isolates indigenous tracking expertise from researcher expectation parameters: trackers examine footprints independently, discuss readings in Ju|'hoansi language, reach consensus before presenting findings to research team who maintain silence throughout entire analysis phase.[13] Epistemological basis for treating these readings as scientific evidence originates from Liebenberg's (1990)[12] demonstration that indigenous tracking employs both systematic tracking (inductive-deductive reasoning from physical evidence) and speculative tracking (hypothetico-deductive or abductive reasoning from environmental context), rendering methodology epistemologically equivalent to formal scientific method [14]

Beyond Pech-Merle, where trackers identified five individuals from seventeen prints—including woman-carrying-load reading overturning Lemozi's (1929)[17] ceremonial interpretation—the project contributed revised or new identifications at Tuc d'Audoubert, Niaux, Fontanet, and Bàsura sites [20][5]

At Niaux, Pales's morphometric reading identified two girls of approximately nine and twelve years in Diverticule des Empreintes passage; San trackers' independent reading of identical tracks [5] found single child of about twelve years moving in slow, controlled motion—divergence between two methods treated as open result rather than settled fact within project documentation. At Tuc d'Audoubert, younger children segregated in Galerie des Petits Pieds while adults conducted structured activities elsewhere. At Aldène, group composition described as consistent with family undertaking planned speleological exploration with anticipated return trajectory.

Groups consistently remain smaller than fifteen individuals. Age ranges demonstrate heterogeneity throughout. At Tuc d'Audoubert, core group includes male-female pair traversing ten locations together systematically. At Cussac, adults wore soft leather moccasins while moving through 1.6 km gallery containing human remains positioned in bear hibernation nests and extensive engraved panels [18] Footwear constitutes single clearest marker of intentional social preparation for cave visitation within complete dataset. No site documents elderly individuals in deep or physically demanding zones.

Group composition inferences derive from foot morphology analysis and stature estimation methodologies—methods distinguishing children from adults with reasonable confidence levels but not reliably separating elderly individuals from younger adults on morphometric grounds alone. Consistent absence of elderly-attributed prints in deep zones therefore compatible with two explanations: actual demographic pattern in which older individuals did not undertake deep cave excursions, OR methodological limitation in which elderly prints exist but remain undetectable as distinct age category through current analytical approaches.

4. Six Consistent Behavioral Patterns

Six behavioral patterns recur across nine-site dataset with sufficient consistency to characterize as typical of Upper Palaeolithic sealed-cave use in northwestern Eurasia region:

Small, specific groups. No cave documents anything approaching crowd-sized assemblies. Largest documented group comprises fourteen individuals at Tuc d'Audoubert, distributed across multiple visits and multiple chambers. Most sites document fewer than six individuals per visit event.

Children in deep zones. Children—including very young children—present in deepest accessible areas at seven of nine sites. This represents single most consistent feature throughout dataset.

Spatial structuring of activity. At every site permitting sufficient data assessment, different activities occur in different cave regions. At Fontanet, Zone 1 (deeper, no artwork) preserves children's play activity, while Zone 2 (closer to prehistoric entrance) contained hearths, occupation evidence, and artistic productions—though footprints lost to water runoff processes.[10]

Deepest points for most significant activities. Clay bison modeling at Tuc d'Audoubert, engravings and funerary deposits at Cussac, art production at Chauvet—all located at deepest or most difficult-to-reach points within cave systems.

Rare visitation. Where stratigraphic or dating evidence permits assessment, visits occurred rarely and episodically. Only Ojo Guareña documents multiple distinct visits—four episodes across approximately fifteen thousand years. Most caves document single visit event occurrences.

No elderly in deep zones. No site documents elderly individuals in deepest or most physically demanding cave zones. Pattern consistent across all nine sites. Group compositions inferable from foot morphology and stature estimation—methods distinguishing children from adults with reasonable confidence but not reliably separating elderly from younger adults morphometrically. Consistent absence therefore compatible with two explanations: actual demographic pattern wherein older individuals avoided deep cave excursions, OR methodological limitation wherein elderly prints present but undetectable as distinct age category.

5. Depth, Activity, and Preservation Mechanisms

Footprints preserve in deep cave zones because deep zones more likely remain undisturbed—they exceed casual access thresholds, beyond flooding pattern ranges, beyond later occupation impacts. Fontanet case demonstrates mechanism directly: Zone 2, closer to prehistoric entrance showing hearths, occupation, and art-making evidence, lost footprints to water runoff processes, while Zone 1, deeper and lacking occupation evidence, preserved sixty-nine or more prints in exceptional detail including fingerprints and fingernail impressions.[10]

This creates preservation asymmetry: places people lived and worked—zones of sustained, repeated activity—tend to destroy footprint evidence. Places people explored, played, or passed through—zones of transient, non-habitation activity—tend to preserve ichnological remains. Sealed-cave dataset therefore structurally biased toward recording exploratory and play behaviors rather than whatever activities may have occurred in entrance zones, occupation areas, or repeatedly visited chambers.

Table 2. Depth, activity, and preservation patterns:

Cave Depth Range Activities Recorded Preservation Mechanism
Chauvet Deep (~150 m+) Solo child, torch work, art chamber visits Rockfall sealing entrance
Pech-Merle Mid-depth Transit corridor movement Entrance collapse
Cussac Deep (1.6 km gallery) Shod adults, engravings, burials Entrance sealed
Ojo Guareña Very deep Exploration, round-trips Restricted deep access
Tuc d'Audoubert Deep (10 locations) Structured pair, clay bison, children's gallery Entrance collapse
Niaux Deep Two children moving Sealed passage
Basura Mid-depth Children crawling, clay balls Stalagmite growth
Fontanet Mid (Z1) to deep (Z2) Z1: children's play; Z2: art/hearths (prints lost) Entrance collapse
Aldène Mid-depth Family exploration, torch management Sealed entrance


6. Recording Windows and Visitation Frequency

Visitation rarity assessable by examining temporal window during which each cave could record footprints—period between earliest dated footprint (or associated evidence) and sealing event (or latest dated evidence).

Table 3. Recording windows and visit counts:

Cave Earliest Date Latest/Sealing Date Window Duration Visits Confidence Level
Chauvet ~36 ka BP (art, uncal.) ~26 ka BP (torch, uncal.) ~10,000 ¹⁴C yrs 1–2 MEDIUM
Pech-Merle ~25 ka BP (art, uncal.) Unknown Unknown ≥1 LOW
Cussac ~31 ka cal BP (bone) ~28 ka cal BP (bone) ~3,000 cal yrs 1 HIGH
Ojo Guareña ~19 ka BP (UP) ~4.3 ka BP (Chalcolithic) ~15,000 yrs 4 HIGH
Tuc d'Audoubert ~17.2 ka cal BP ~16.5 ka cal BP ~700 cal yrs 1–2 HIGH
Niaux ~13 ka BP (cultural) Unknown Unknown ≥1 LOW
Basura ~12,340 BP (uncal.) ~12,340 BP Unknown ≥1 MEDIUM
Fontanet 12,770 ± 420 BP (uncal.) Unknown Unknown ≥1 MEDIUM
Aldène ~10 ka BP (cultural) ~7 ka BP (cultural) ~3,000 yrs 1 LOW


Only three caves—Cussac, Ojo Guareña, and Tuc d'Audoubert—possess recording windows based on calibrated radiocarbon dates with direct association to footprint contexts. Chauvet contains two radiocarbon dates in uncalibrated years, relationship ambiguous between art (36 ka) and child's torch (26 ka): possibly two visits separated by millennia, OR dates bracket single accessibility period. For remaining five caves, no recording window estimable—only minimum single visit established.

Three high-confidence cases demonstrate: Cussac, one visit across approximately 3,000 calibrated years. Ojo Guareña, four visits across approximately 15,000 years—significantly, cave's continuous accessibility across this period itself evidences non-use: site remaining open and visitable for fifteen millennia entered only four times. Tuc d'Audoubert, one to two dated visit episodes across approximately 700 calibrated years—while trackway analysis (Chapter 13) admits up to five visits by groups of two to six individuals (fourteen unique subjects across two hundred fifty-five prints).

These figures constitute foundation for frequency analysis. Remaining caves consistent with rare visitation but do not independently establish pattern conclusively.

7. Discussion and Implications

Preservation asymmetry fundamentally structures ichnological evidence availability for archaeological interpretation. Deep, undisturbed zones preserve footprint evidence precisely because those zones experienced minimal subsequent human activity—exploration, play, brief visits—whereas zones of sustained occupation, repeated habitation, and intensive activity destroy ichnological traces through erosion, trampling, and sediment disturbance.

This creates interpretive challenge: surviving footprint evidence disproportionately represents transient, non-habitation activities rather than regular domestic or economic behaviors. Archaeological interpretations relying exclusively on footprint evidence risk systematic over-representation of exploratory, recreational, and ceremonial activities while under-representing routine subsistence behaviors, craft production, and domestic routines occurring in entrance zones and occupation areas.

Multidisciplinary methodology combining high-resolution 3D photogrammetry, morphometric analysis, and indigenous tracking expertise demonstrates substantial analytical capacity gains. San tracker readings at Pech-Merle, Tuc d'Audoubert, Niaux, Fontanet, and Bàsura produced revised identifications overturning earlier archaeological interpretations [17][19], establishing epistemological equivalence between indigenous tracking methodologies and formal scientific procedures when applied systematically and independently [12][13].

8. Conclusions

Nine naturally sealed caves in northwestern Eurasia provide direct evidence of Upper Palaeolithic human behavior spanning approximately 36,000 to 7,000 BP. Dataset demonstrates consistent patterns: multi-generational family groups, child presence in deep zones, spatially structured activities, and rare visitation frequencies. Preservation asymmetry reveals structural bias favoring exploration and play behaviors over routine domestic activities.

Methodological integration combining traditional archaeological analysis, high-resolution 3D documentation, and indigenous tracking expertise produces superior interpretive outcomes compared to singular disciplinary approaches. Future research should prioritize expansion of comparative datasets, refinement of morphometric aging methodologies, and continued collaboration with indigenous tracking communities.

References

  1. ↑ 1.0 1.1 Pales, L. (1976). Les empreintes de pieds humains dans les cavernes: les empreintes du réseau nord de la Caverne de Niaux (Ariège). Paris: Archives de l'Institut de Paléontologie Humaine, Mémoire 36.
  2. ↑ 2.0 2.1 Duday, H. & Garcia, M.-A. (1985). L'homme et la caverne. Les Dossiers histoire et archéologie 90: 28–34.
  3. ↑ Leroi-Gourhan, A. (1986). Gestes et paroles. Paris: Editions Albin Michel.
  4. ↑ Clottes, J. & Lewis-Williams, D. (1998). Les chamans de la préhistoire: transe et magie dans les grottes ornées. Paris: Seuil.
  5. ↑ 5.0 5.1 5.2 5.3 5.4 Pastoors, A., Lenssen-Erz, T., Ciqae, T., Kxunta, U., Thao, T., Bégouën, R., Biesele, M. & Clottes, J. (2015). Tracking in Caves: Experience-Based Reading of Pleistocene Human Footprints in French Caves. Cambridge Archaeological Journal 25(3): 551–564. https://doi.org/10.1017/S0959774315000021
  6. ↑ Breuil, H. (1912). Quatre cents siècles d'art pariétal. Montignac: Centre d'Études et de Documentation Préhistoriques.
  7. ↑ Bordes, F. (1972). Prehistory. London: Weidenfeld & Nicolson.
  8. ↑ White, R. (1982). Uralic Languages and their Speakers. Helsinki: Suomalais-Ugrilainen Seura.
  9. ↑ 9.0 9.1 9.2 Garcia, M.-A. (1999). Human footprints in the Chauvet cave. International Newsletter on Rock Art (INORA) 24: 1–4.
  10. ↑ 10.0 10.1 10.2 10.3 10.4 Ledoux, L. (2019). L'ichnologie préhistorique et les traces d'activités au sein des cavités ornées: le cas de la grotte de Fontanet (Ariège, France). PhD dissertation, Université de Bordeaux.
  11. ↑ 11.0 11.1 11.2 11.3 Citton, P., Romano, M., Salvador, I. & Avanzini, M. (2017). First occurrence of archaeological human footprint cast: an experimental study. Journal of Archaeological Science: Reports 14: 686–693. https://doi.org/10.1016/j.jasrep.2017.07.001
  12. ↑ 12.0 12.1 12.2 Liebenberg, L.W. (1990). The Art of Tracking: The Origin of Science. Claremont: David Philip Publishers.
  13. ↑ 13.0 13.1 13.2 Biesele, M. (2021). Trackers' consensual talk: Precise data for archaeology. In Pastoors, A. & Lenssen-Erz, T. (eds.), Reading Prehistoric Human Tracks, pp. 385–397. Cham: Springer. https://doi.org/10.1007/978-3-030-60406-6_19
  14. ↑ 14.0 14.1 Lenssen-Erz, T. & Pastoors, A. (2021). Tracker-based reading of prehistoric footprints. In Pastoors, A. & Lenssen-Erz, T. (eds.), Reading Prehistoric Human Tracks, Chapter 6. Cham: Springer. https://doi.org/10.1007/978-3-030-60406-6_6
  15. ↑ Bocquet-Appel, J.-P., Demars, P.-Y., Noiret, L. & Dobrowsky, D. (2005). Estimates of Upper Palaeolithic meta-population size in Europe from archaeological data. Journal of Archaeological Science 32(11): 1656–1668. https://doi.org/10.1016/j.jas.2005.05.006
  16. ↑ 16.0 16.1 Pastoors, A., et al. (2016). Tracking in caves: the San trackers in the Pech-Merle cave. Quaternary International 430: 184–195. https://doi.org/10.1016/j.quaint.2016.02.046
  17. ↑ 17.0 17.1 17.2 17.3 Lemozi, A. (1929). La Grotte-temple du Pech-Merle: un nouveau sanctuaire préhistorique. Paris: Picard.
  18. ↑ 18.0 18.1 Ledoux, L., et al. (2021). Tracks in the dark: ichnological evidence of footwear in Cussac Cave. Archaeological and Anthropological Sciences 13: 111. https://doi.org/10.1007/s12520-021-01362-0
  19. ↑ 19.0 19.1 19.2 Blanc, A.C. (1960). Le palline d'argilla della Grotta della Bàsura. Rivista di Studi Liguri 26: 183–189.
  20. ↑ Lenssen-Erz, T., Pastoors, A., Ciqae, T., Kxunta, U., Thao, T., Bégouën, R., Biesele, M. & Clottes, J. (2018). Reading human spoor in ice age caves with San hunters. In Puckett, R.F. & Ikeya, K. (eds.), Research and Activism among the Kalahari San Today (Senri Ethnological Studies 99), pp. 103–127. Osaka: National Museum of Ethnology.
BY
Observatory Research Team
Collaborative Research Desk
The Observatory Research Team coordinates open preprints, cross-disciplinary field synthesis, and collaborative research initiatives.
Inferno Labs
Research & Evaluation Laboratory
Inferno Labs produces software to support verification, inference, statistical testing, corpus benchmarking and project management. Inferno Labs uses LLM in its managed production channels.