Carnegie Mellon University

A Humanities-Centered History of Early AI:
Herbert A. Simon and His Books

Author: Avery Wiscomb

Degree: Ph.D. in Literary & Cultural Studies, Carnegie Mellon University, 2021

How should we regard the relationship between literature and artificial intelligence (AI)? Most scholarship in literary studies that examines AI does so by analyzing texts created by machines or by studying AI's representation in pulp science fiction and utopian and dystopian fiction. Some AI researchers have also suggested that books, art, television, and film have shaped the kinds of AIs scientists create. However, what has been less examined is how reading literature has influenced the scientists who invented such technologies.

This dissertation fills a gap in literature and science studies by describing how literature played an influential role in the life and work of one of the founders of AI, Herbert A. Simon (b. 1916 – d. 2001). To do so, I created a novel archive of Simon's papers available at Carnegie Mellon University Libraries, totaling about 37,000 documents. I also compared Simon's writing and related documents to about 450 books from his library donated to Carnegie Mellon University after his death. My approach to analyzing the Herbert Simon Collection (HSC) adapts a process called "critical search" (Guldi). It combines traditional and digital approaches to self-reflexive search and text-data mining practices in historical archives for guided close reading. In reading for the significance of Simon's source texts, I also draw from citation analysis, theory analysis, and translation studies.

Chapter One introduces the ongoing difficulties in tracing the influence of literature on contemporary science fields such as AI and presents the Simon case. Chapter Two describes a history of Simon's education as a reader, from his childhood to his training in political science at the University of Chicago. In Chapter Three, I read Simon's first book Administrative Behavior (AB), through the humanist lens of the ancient Greek and Roman tragedies he read, loved, and translated. Chapter Four describes how Simon "transmigrated" the individual he proposed in AB into his Logic Theorist program. Finally, in Chapter Five, I narrate how Simon sought out a theory of free will in conversation with the author Jorge Luis Borges.

I find that reading and translating imaginative literary texts activated Simon's receptivity to diverse methodological approaches and contributed to his wide-ranging impact as a researcher. Simon's scientific accomplishments, I argue, were profoundly shaped by his humanities and literary training. Understanding aspects of Simon's work from this perspective offers an alternative, humanities-centered history of early AI, focusing on the significance of literature in scientific and technological innovation.