Professor Edward Mendelson arrived late to his Modern Poetry class one day last spring, apologizing profusely for having been delayed by a talk. The talk, he explained, concerned a piece of software someone was showing off to the English department which could, given the text of a book, identify its genre.
This was a surprising point of overlap between my classes—I’d written a nearly identical program as a homework assignment in my machine learning course the previous week. I expected Professor Mendelson to make some insightful remark about the basic predictability of individual writing styles or the narrowness of genre conventions, perhaps some broader point about the extraordinary fact that complex ideas like “romance novel” can be characterized, at least in some narrow sense, by embarrassingly simple statistical models.
“Hogwash,” he said, and began the lecture.
I don’t remember exactly what we discussed that day, but for rhetorical effect I’ll assert that it was Auden’s “Under Which Lyre.” The poem’s antepenultimate stanza reads:
Thou shalt not answer questionnaires
Or quizzes upon World-Affairs,
Nor with compliance
Take any test. Thou shalt not sit
With statisticians nor commit
A social science.
I’m inclined to agree with Professor Mendelson that automatic genre classification provides little insight into language, but what about the broader question—do students of the humanities have something to learn from statistics, or should they refuse to “sit / With statisticians” altogether? As numerical models begin to encroach on central aspects of human identity, and begin to shape our tastes and bodies and relationships, there’s an understandable urge to insist that Columbia ought to preserve the Core Curriculum as a safe space for humanistic thought, free from the threatening determinism of numbers. This is the attitude that seems to inform the science requirement. While the Core responds to a perceived need for science courses, the science requirement’s basic impulse is almost anti-scientific: Students are encouraged to develop a vague appreciation of the heavenly bodies or the human genome without any real grasp of scientific and numerical methodologies.
This is dangerous. The quantity of information, and of numerical information in particular, that we are forced to consume daily has grown tremendously since the Core was established; numerical methods, for better or worse, have as much to say about the human experience as Aristotle does. If the Core Curriculum is to achieve its goal of helping students “understand the civilization of our own day and participate effectively in it,” it must take science seriously. The science requirement demands the same focus that is the strength of the rest of the Core Curriculum, and Columbia should begin by requiring an introductory course in probability and statistics for all students.
Why statistics? That such a course would provide useful technical skills independent of the Core’s goals is self-evident and uninteresting; the point is that the study of statistics has a place in a modern humanities curriculum. The central questions in statistics are fundamentally about the interpretation of numbers—how to question the sufficiency of data, to bound confidence, to formulate hypotheses. These are not simply technical problems, but epistemological ones. The study of statistics goes to the heart the Core’s dual purpose, at once enabling students to understand the foundation of the scientific mind-set and teaching them to apply that mind-set to present problems.
Without the tools to evaluate numerical evidence, there are only two options left to us. One is to fall into the trap that Auden warns against, to trust numbers unquestioningly, to ascribe to the results of “questionnaires / and quizzes upon World-Affairs” the character of natural law, to accept that “romance novel” is a Bernoulli process, and give in to a dehumanized view of human history and the human spirit. But the other option—to reject utterly numerical data as a basis for personal and ethical decision-making—is equally dangerous. We have fought hunger and disease with surveys and averted financial collapse with statistics. These tools are as central to modern civilization as the "Ethics" are, and Columbia students can no longer participate effectively in the world without basic number sense.
The author is a School of Engineering and Applied Science senior majoring in computer science. He will study at Cambridge University on a Churchill Scholarship next year.
