Computer Science as a Liberal Art

What place does computer science have at a liberal arts college? This question seems urgent to me, a senior math-CS major at Reed, as our young department seeks to establish its place in the college at large and as it struggles within itself over the nature of its subject. I think that many of the concerns about the role of computer science stem from misunderstandings over its nature, both as a subject and as taught at Reed. Computer science, as practiced at Reed and throughout the academy, is not solely vocational training but is additionally an essential intellectual pursuit with broad applications to humanistic inquiry.

A rough definition: computer science is something like the study of computation. Computations—algorithmic processes for solving problems—are likely as old as thought; their structure has been studied at least since the invention of the abacus in Sumer in the third millennium BCE. Computer science, in this sense, cannot be understood merely as the study of computers, as in the literal physical devices we all use every day. Instead, it is the study of the very idea of mechanistic problem-solving. Modern computers are an important example of such tools, so inevitably feature centrally in the discipline as both method and object, but the stakes of the discipline must be understood as far older. Of course, as with all academic disciplines, its modern borders are historically contingent—computer science is entangled with mathematics and philosophy, with physics and linguistics, with sociology and political science. Any field with something to say about computers, the artifacts, or the computation process, or which uses computers or computations as tools for studying its own objects, ought to be a natural ally.

Indeed, the study of computation addresses numerous fundamental, humanistic inquiries. In CS 382 (Algorithms and Data Structures) and 387 (Computability and Complexity), we discuss which questions can possibly be answered by rote computation—and by contrast, which questions require genuine creative thought. In CS 315 (Ethics and Public Policy) and 386 (Private and Fair Data Analysis), we question the role of computer technology and algorithmic decision-making in society, asking when the state ought to act with discretion and when it ought to act by mere procedure. In CS 378 (Deep Learning), 384 (Programming Language Design and Implementation), and 394 (Principles of Compiler Design), we study different means of representing processes of computation, and hence processes of thought—and the limits that any such translation must inevitably have. I think no one would doubt that these questions and many more like them have a place in the liberal arts.

In Plato’s Republic, he says that education should be “the art of orientation”—aimed not to impart knowledge onto students, but to orient them so that they can seek knowledge for themselves. By this standard, too, computer science has a clear place in the liberal arts. At its best, it teaches us to precisely communicate our thoughts and arguments, to reason rigorously and creatively about deep questions, and to critically evaluate tradeoffs between conflicting arguments or design choices. At times we are able to work with mathematical precision, and at times we have to make subjective or even philosophical judgments. That the methods of the field are so varied is evidence of the naturality of its object: questions about computation reward you for studying them from different perspectives, with as many tools as you can muster.

In this light, we must understand that computer science is not just coding or software engineering. Certainly, coding is one of the central tools of the discipline—analogous to the role of ethnography in anthropology, or primary source analysis in history. It is a tool we use to express computational patterns, either to specify how a computer should behave or to communicate our ideas precisely to other computer scientists. Many computer scientists spend much of their time coding, but so too do many anthropologists spend much of their time performing fieldwork. I do not mean to denigrate the importance of coding to the discipline—computer science is enormously valuable for software engineers, in the same way that anthropology is valuable for journalists—but software engineering is not its raison d’etre.

Computation is as natural an object of study as matter, or life, or politics, or literature; the methods of computer science are as valuable as those of chemistry, or biology, or political science, or English. Ada Lovelace, the author of the first modern computer program and hence one of the originators of the modern discipline of computer science, wrote that “the Analytical Engine weaves algebraical patterns just as the Jacquard loom weaves flowers and leaves.” Computer science is nothing more or less than the study of the weaving of such patterns.

Riley Shahar