Neanderthals made tar. Birch bark tar, used as adhesive for hafting stone tools to wooden handles, has been found at multiple Middle Paleolithic sites. The production process requires sustained heating of birch bark in a low-oxygen environment — conditions that do not occur naturally and must be deliberately created and maintained.
The question is how much cognition this required (arXiv:2305.09751). Not in the vague sense of “were Neanderthals smart?” but in the specific, measurable sense: how many concurrent variables must the producer track simultaneously? Working memory capacity — the number of independent items that can be actively maintained and manipulated — is one of the few cognitive metrics that can be operationalized and compared across species and across time.
Petri nets formalize the process. Each step in tar production is a transition. Each material state is a place. Each token represents a unit of material or a condition that must be maintained. The concurrent tokens — conditions that must hold simultaneously — map directly to working memory demands. If four conditions must be maintained at once (fire temperature, bark placement, oxygen exclusion, timing), the process requires working memory capacity of at least four.
The analysis finds that Neanderthal tar production required working memory loads comparable to modern human preferences — roughly 4 to 7 concurrent items. Not higher than modern humans, but not substantially lower either. The process is within the range that modern adults comfortably manage.
This does not prove cognitive equivalence. Working memory capacity is one axis of cognition among many. But it falsifies a specific claim: that Neanderthal technology is explicable by simple sequential reasoning with minimal concurrent tracking. The tar recipe is irreducibly parallel — multiple conditions must hold simultaneously, not sequentially. The recipe is the evidence, and the Petri net is the ruler.