This chapter explores the putative connection between knotting and linguistic abilities in recent human evolution. As a first step, it gives mathematical grounds to the idea that it makes sense referring to “knowledge of knotting” or “knotting grammars,” as customarily done in the case of language, yet maintaining open whether the underlying computational mechanisms are specific for each domain or shared. After reviewing some pieces of evidence of the use of scaffolds by early modern humans that clearly entail complex knotting activity—and which seems lacking in the case of Neanderthals, the chapter concludes that a cognitive architecture with a shared computational facility accessible to both knotting and language, is to be preferred both on theoretical and empirical grounds. Comparative data with other non-human species, as well as patterns of comorbidity in the context of linguistic impairments, are offered that support the stance

This chapter considers two competing views about what modularity might consist of, which the authors refer to as sui generis modularity and descent-with-modification modularity. Descent-with-modification helps make sense of the considerable phylogenetic continuity that has been documented in recent years, in terms of comparative psychology, comparative neuroanatomy, and comparative genomics. The chapter also considers language, the canonical putative module, and its relation to cognitive systems. Cognitive mechanisms for spatial and temporal representation seem to run deeply through the structure of the linguistic system. The notion of descent-with-modification, once recognized, has significant implications for how one can assesses debates about modularity. The descent-with-modification perspective suggests caution for inferring the absence of modularity from many studies of "normal" cognition. Descent-with-modification also suggests that one should expect the hallmarks of ancestry even in the very machinery that makes abstract linguistic representation possible.