For nearly two decades, a TV game show called Pyramid aired in North America. The show featured two competing teams, each team consisting of two contestants: an ordinary civilian contestant and a celebrity. In the show’s first round, both teams’ members viewed a pyramid-shaped sign that displayed six category titles, some straightforward like “Where You Live” and others less conventional like “Things You Need to Feed.” Each team then had an opportunity to compete for points in 30-second turns. The goal was for one team member to gain points by identifying a word or phrase related to the category from clues provided by the other team member. For example, a target phrase for the “Where You Live” category might be “zip code,” and the clue might be “Mine is 94705.” “Things you Need to Feed” might include both “screaming baby” and “parking meter.”
The team that won the first round advanced to the “Winner’s Circle,” where the game was turned around. This time, only the clue giver was shown the category name and had to suggest concepts or instances belonging to that category so that the teammate could guess the category name. Clues like “alto,” “soprano,” and “tenor” would be given to prompt the teammate to guess “Singing Voices” or “Types of Singers.”
As the game progressed, the categories became more challenging. It was interesting and entertaining to hear the clue receiver’s initial guess and how subsequent guesses changed with more clues. The person giving clues would often become frustrated, because to them their clues seemed obvious and discriminating but would seem not to help the clue receivers in identifying the category. Viewers enjoyed sharing in these moments of vocabulary and category confusion.
The Pyramid TV game show developers created a textbook example for teaching about categories—groups or classes of things, people, processes, events or anything else that we treat as equivalent—and categorization—the process of assigning instances to categories. The game is a useful analog for us to illustrate many of the issues we discuss in this chapter. The Pyramid game was challenging, and sometimes comical, because people bring their own experiences and biases to understanding what a category means, and because not every instance of a category is equally typical or suggestive. How we organize reflects our thinking processes, which can inadvertently reveal personal characteristics that can be amusing in a social context. Hence, the popularity of the Pyramid franchise, which began on CBS in 1973 and has been produced in 20 countries.
Many texts in library science introduce categorization via cataloging rules, a set of highly prescriptive methods for assigning resources to categories that some describe and others satirize as “mark ’em and park ’em.” Many texts in computer science discuss the process of defining the categories needed to create, process, and store information in terms of programming language constructs: “here’s how to define an abstract type, and here’s the data type system.” Machine learning and data science texts explain how categories are created through statistical analysis of the correlations among the values of features in a collection or dataset. We take a very different approach in this chapter, but all of these different perspectives will find their place in it.
Cataloging and programming are important activities that need to be done well, and prescriptive advice is often essential. However, we believe that understanding how people create psychological and linguistic categories can help us appreciate that cataloging and information systems design are messier and more intellectually challenging activities than we might otherwise think.