Thursday, April 18, 2024 - 6:00pm to 8:00pm

David E. Dunning

History and Sociology of Science, University of Pennsylvania,

In this talk I will discuss two episodes from my current book project, a history of the rise of mathematical logic and its connection to early computer science and AI. I focus on shifting and competing notational practices in order to show how methods of writing have shaped understandings of natural and artificial reasoning alike. After briefly introducing the larger project, I will share two cases. I’ll discuss Victorian self-taught mathematician George Boole’s efforts to rewrite logic in algebraic notation, revealing his specific vision for how and why his system should be learned. Many of his readers were very enthusiastic, but rather than straightforwardly adopting his system, they tended to follow him in taking notation itself as a fruitful arena for innovation. After presenting Boole, I’ll turn to the watershed 1943 paper “A Logical Calculus of the Ideas Immanent in Nervous Activity,” by Warren McCulloch and Walter Pitts. The authors drew on recent mathematical logic and navigated “typographical necessity” in order to present abstract neurons as a system for writing logical propositions. By taking seriously the centrality of writing in the project that launched neural network techniques, I aim to show how modelling the mind mathematically was not simply about understanding or imitating intelligence, but rather reimagining thought as a fundamentally notational phenomenon.