June 2021 Faculty Research Spotlight
Dr. Leszek Pawlowicz and Dr. Chris Downum | Breakthrough in Archaeological Applications of Machine Learning
Typologies are central to archaeology. Everything from spear points to pottery designs to sandals to societies have been classified into types. Anthropologist Julian Steward in 1954 even proposed a typology of typologies, writing an article called “Types of Types.” I cannot think of many archaeological endeavors that do not use some form of typology. This is necessary, because typologies reduce complex and otherwise baffling variation into a limited number of comprehensible categories. They help us make sense of what author Henry James once called “the intolerable buzz of reality.”
But how well do typologies work? Do they oversimplify, and thereby potentially mislead? Do archaeologists apply them consistently? How does a person learn a typology? Do typologies change through time, and if so, by how much? What are the rules of typology, and who documents and enforces them? These and other related questions have intrigued archaeologists since the beginning of archaeology. They are also some of the questions that I and Anthropology adjunct faculty Leszek Pawlowicz recently set out to explore, using a form of Artificial Intelligence known as machine learning.