New morphological analysis of artifacts from Folsom‑period sites shows that Native American groups on the western Great Plains made and used two‑sided “binary lots” in structured games roughly 12,800–12,200 years ago. This finding changes how scholars and curators should evaluate claims about the origins of dice and early probabilistic practice.
What the artifacts are and where they came from
The objects identified are two‑sided gaming pieces — called binary lots — usually carved from bone or wood and marked so each face reads differently when thrown. Archaeologists applying the new criteria have confirmed just over 600 such pieces from 57 sites across 12 U.S. states, with key finds coming from Folsom‑period sites in Wyoming, Colorado, and New Mexico. The dated range (about 12,800–12,200 years before present) places them roughly 6,000 years earlier than the earliest known dice from Mesopotamia and the Indus Valley.
These pieces were not cube dice but small, handheld tokens thrown together; players counted how many pieces landed on a particular face. That binary outcome system resembles a multi‑coin toss rather than the multi‑faced dice familiar from later Old World contexts.
How a morphological checklist altered artifact classification
Doctoral researcher Robert Madden at Colorado State University developed a systematic morphological test by returning to ethnographic records: he used a 1907 documentation of 293 historic Native American dice sets as the baseline checklist. Applying that checklist to museum collections and existing excavated objects allowed previously ambiguous artifacts to be reassessed against visible, repeatable criteria—distinct face treatments, consistent size ranges, grouped recovery patterns, and characteristic wear consistent with repeated throwing.
That method matters because many of the objects had sat in collections without confident classification. Madden’s approach turned qualitative ethnographic descriptions into explicit diagnostic conditions; the result was confirmation of more than 600 binary lots across dozens of sites, which strengthens continent‑wide interpretation rather than relying on a single site or anecdote.
Games, social function, and contrast with Old World gambling
Contextual evidence indicates these pieces were used in one‑on‑one contests without any institutional “house” advantage. Excavation contexts and associated artifacts suggest the games functioned as neutral, rule‑governed mechanisms for social exchange, alliance‑building, and communication between groups with limited regular contact. Unlike later commercial gambling systems, the archaeological signature points to gifting and reciprocal social processes rather than marketized wagering.
Comparatively, Old World dice traditions that appear millennia later are embedded in different material forms and socio‑economic settings; Mesopotamian and Indus examples are often six‑faced or multi‑faced and enter historical records in courtly, administrative, or ritual contexts. The Ice Age evidence therefore represents an independent emergence of repeatable randomization devices and social practices around chance, not a simple diffusion from the Old World.
How to read a putative “dice” — a quick decision checklist for researchers and curators
When deciding whether an object should be cataloged as a gaming piece, use multiple, coordinated signals rather than a single attribute. Below is a concise comparison of diagnostic signals, their implications for interpretation, and practical thresholds for museum or publication action.
| Signal | What it implies | Action / Confidence threshold |
|---|---|---|
| Distinct face treatments (paint, incision, burn) | Deliberate two‑state design for binary outcomes | If present, flag as candidate; proceed to context check |
| Grouped recovery (multiple pieces in one locus) | Used together in games rather than single utilitarian tools | Strong support; elevate to probable gaming set |
| Wear patterns consistent with throwing | Repeated handling and impact, not tool use | Increases confidence; combine with face treatment |
| Associated radiocarbon dates / stratigraphy | Places artifacts in cultural sequence and enables comparisons | Required for chronological claims; without it, classify cautiously |
| Ethnographic analogues or historic sets | Provides behavioral model and morphological reference | Essential for converting candidate to confirmed using Madden’s checklist |
Short Q&A for immediate questions
How certain are the dates? The date range (~12,800–12,200 BP) comes from stratigraphic associations at Folsom contexts and calibrated radiocarbon results tied to the sites where the pieces were found; individual pieces without direct dating should be treated as provisional.
Are these the same as modern six‑sided dice? No. These are binary lots intended to produce two outcomes per piece; players scaled probability by throwing groups of pieces rather than relying on multi‑faced geometry.
Does this mean formal probability theory existed here? Not formalized mathematical theory, but the archaeological pattern shows intentional, repeatable engagement with randomness and scalable outcomes—practical probabilistic reasoning that predates Old World examples by millennia.
Next research checkpoints: investigate how variations in shape, material, and marking across sites affected game rules and social roles, and work directly with tribal communities to situate continuity and meaning. For anyone writing about origins, the practical test is simple: apply morphological and contextual criteria before attributing invention or primacy to any single region.
