ILCB Lunch talk was held by Tecumseh Fitch the 01/30/2018
An understanding of both the neural mechanisms involved in language, and their evolutionary history, requires incisive comparisons between humans and nonhuman animals. Ideally, such comparisons are grounded in an explicit, computational framework encompassing both formal and neural components. I review work comparing humans with nonhuman primates, other mammals, and birds carried out in the last decades, much of it using artificial grammar learning to explore the perception of phonology and syntax.
This research suggests the following two hypotheses: First, the phonological continuity hypothesis holds that sequential processing of syllables is supported by equivalent, homologous mechanisms in humans and other animals. This set of mechanisms allows combination via concatenation, and supports sequential processing at the finite-state (regular) computational level. Second, the dendrophilia hypothesis suggests that humans are unusual in our ability to process complex hierarchical structures in multiple domains (language, music, etc). These hierarchical abilities require computational power at the supra-regular level (above finite state), and supports the abstract structures needed for phrasal syntax and semantics. I propose that these general hierarchical abilities are supported neurally by the great enlargement of Broca’s area in our species, and the broadening of its connections to most of the parietal and temporal lobes. Broca’s region in humans acts as a domain-general “stack”, an auxiliary memory supporting supra-regular computation in both language and music.
Tecumseh Fitch