Abstract
Understanding the evolution of animal cognitive capacities requires us to study their full range of naturally occurring sequences of behavior. It has long been theorized that cognitive capacities are revealed through the sequential structure of natural behavior, particularly its hierarchical organization. Progress in understanding the origins of this capacity has, however, been limited by a lack of techniques for identifying and measuring hierarchical structure in behavioral sequences. To fill this methodological gap, we introduce here an analysis pipeline for measuring hierarchical structure in sequential behavior. We then establish the validity of our approach by first applying it to chimpanzee percussive tool-use (PTU) sequences and comparing it to markov-simulated control sequences. Secondly, we apply our analysis to a dataset on PTU in humans and compare the hierarchical complexity of chimpanzee and human PTU. Despite decades of speculation, our study is the first empirical demonstration of hierarchical structuring in chimpanzee tool-use. We found chimpanzee PTU is characterized by a level of hierarchical complexity beyond that which can be generated through markov process, but is nonetheless systematically less hierarchically complex than human PTU, as expected. Altogether, our analyses demonstrate the potential for our approach to successfully detect and measure hierarchical structuring in natural sequences of behavior, which we believe will play a pivotal role in shedding light on old questions, as well as opening up entirely new lines of inquiry in the study of human and animal behavior.