A central property of human language is its hierarchical structure. Humans can flexibly combine elements to build a hierarchical structure expressing rich semantics. A hierarchical structure is also considered as playing a key role in many other human cognitive domains. In music, auditory-motor events are combined into hierarchical pitch and/or rhythm structure expressing affect. How did such a hierarchical structure building capacity evolve? This paper investigates this question from a bottom-up perspective based on a set of action-related components as a shared basis underlying cognitive capacities of nonhuman primates and humans. Especially, I argue that the evolution of hierarchical structure building capacity for language and music is tractable for comparative evolutionary study once we focus on the gradual elaboration of shared brain architecture: the cortico-basal ganglia-thalamocortical circuits for hierarchical control of goal-directed action and the dorsal pathways for hierarchical internal models. I suggest that this gradual elaboration of the action-related brain architecture in the context of vocal control and tool-making went hand in hand with amplification of working memory, and made the brain ready for hierarchical structure building in language and music.

人类语言的一个核心特性是它的层次结构。人类可以灵活地组合元素来构建表达丰富语义的层次结构。层次结构也被认为在许多其他人类认知领域中起着关键作用。在音乐中，听觉运动事件被组合成层次的音高和/或节奏结构来表达情感。这样的层次结构建设能力是如何演变的？本文从自下而上的角度研究了这个问题，基于一组与动作相关的组件作为非人类灵长类动物和人类认知能力的共享基础。尤其，我认为，一旦我们关注共享大脑结构的逐渐详细说明，语言和音乐的层次结构构建能力的演变对于比较进化研究来说是易于处理的：用于目标导向动作的层次控制的皮质-基底神经节-丘脑皮质电路和分层内部模型的背侧通路。我建议在声音控制和工具制造的背景下逐步阐述与动作相关的大脑结构与工作记忆的放大相辅相成，并使大脑为语言和音乐中的层次结构构建做好准备。