Computational and Neurological Questions of Language w/ Cory Shain

Although Cory Shain (currently at MIT, soon to be at Stanford) studies language, therefore making him a “linguist,” his research could easily be classified as belonging to a number of other disciplines. To understand the computations responsible for language processing, he engages heavily with computer science. To study the functional organization and architecture of language in the brain, he uses methods of neuroscience. To round out the complexities of his research, he pulls from the theories and insights of cognitive science and psychology. Thus, Cory Shain actively questions the boundaries between language and other parts of human communication and cognition by remaining curious about how language works, not just what language is.

In this episode we first dig into the section of his research that’s dedicated to understanding the problem of difficulty in language processing. What mediates the difficulty of processing cost: a sentence's predictability or its frequency?

Then, we come to a truly awe-some question: how and why do we define what counts as language? Is language solely the processes that a specific network or cortex carries out? Can language encompass our intuitions into what someone else might be thinking? Why is there a common assumption that a language-specific network exists, and how does such an assumption influence our understanding of both the brain and of language?

No matter your previous understanding of language processing, the language network, or theory of mind, you will finish this episode having learnt something new about language, the brain, dependency locality, the importance of similar results across studies with varied methods, or perhaps the overlap between industry and academia.

Cory Shain's publications:

Eve Federenko

Language and thought are not the same thing: evidence from neuroimaging and neurological patients

Broca-Wernicke Theories: A Historical Perspective

Rebecca Saxes

Ben Deen