By defining cumulative cultural evolution as socially transmitted behavioral change that measurably improves performance across repeated iterations, the paper turns accumulation into a testable process. That definition gives Coordination Systems a boundary, not a slogan: transmission alone is insufficient unless later contributions retain and improve on earlier ones. The Cultural Intelligence program therefore has a concrete criterion for whether group memory is actually compounding.
The term was being used for very different things, so the paper writes down criteria. A change in behavior. Transmitted socially. Producing a measurable improvement in performance. Repeated, so improvements stack on earlier improvements.
They then add stricter criteria, including traits no single individual could have invented alone, and apply the whole set to the animal and human cases people had claimed.
Most animal cases satisfy the core criteria and fail the strict ones. The value here is the definition, not the verdict: it turns a word people were arguing about into something you can test a system against.