How to Delight With Statistics
Edward Tufte's books make train schedules irresistible.
Etymology, the study of word origins, often interests people otherwise uninterested in language. The reason, surely, is that etymology is tethered to ordinary life in ways that are easy to grasp. Anyone can enjoy knowing how a $10 bill came to be called a sawbuck (the Roman numeral "X" that appeared on early notes reminded people of the wooden sawbuck used in carpentry), or how the word "robot" came into English (it comes from the Czech word "robota," meaning "drudgery," and was part of the title of a widely popular 1920 Czech play), or how the Indo-European root for "beech tree," "bhago-," gave us the word "book" (Germanic tribes used beech staves to carve runes on). We like etymologies because they tell us stories.
There would be no telling of stories at all, of course, without grammar. But people uninterested in language issues are content to remain uninterested in grammar. Grammar, the genetic blueprint of meaning, operates at a level of abstraction. Nouns and verbs and other parts of speech are considered not as familiar individuals but as members of different species. Their interactions are considered as elements of an ecosystem. Grammar is hard to visualize, and it is harder to enjoy.
Or at least it was, until a man named Edward R. Tufte began to publish his work about the display of information in the early 1980s. Tufte has produced three books that make it possible to see the dynamics of linguistic grammar in a brilliantly synesthetic way--books that succeed in this task by not having grammar as their ostensible subject at all.
E dward Tufte is a professor at Yale who teaches courses in statistics and information design. In 1983 he produced an elegant book with the unprepossessing title The Visual Display of Quantitative Information, which both explained and demonstrated certain rules for presenting numerical data in a graphically arresting manner. This tour de force--written, designed, and published by Tufte himself--was immediately recognized as a classic by data wonks and makers of fine books alike. It has never been out of print.
In his analysis of such improbably compelling genres as railroad schedules and balance sheets, chemical symbols and weather summaries, Tufte reprinted numerous graphic displays that epitomized the statistical draughtsman's conceptual art. In the French engineer Charles Joseph Minard's 1861 encapsulation of Napoleon's Russian campaign, for example, a compact space revealed the diminishing size of the army, its day-to-day geographical location, the direction of its route, the passage of time, and the gradual drop in the temperature. It is a map, Tufte noted, that illustrates "how multivariate complexity can be subtly integrated into graphical architecture, integrated so gently and unobtrusively that viewers are hardly aware that they are looking into a world of four or five dimensions."
T ufte did not explicitly spell out his ultimate intentions in that first volume of his trilogy, but it has turned out that the whole series has been produced according to a precise schema--as he now indicates in the introduction to Visual Explanations, which was just published by his company,. The Visual Display of Quantitative Information had been devoted to pictures of numbers, he explains--"how to depict data and enforce statistical honesty" whether the subject is traffic deaths or the distribution of galaxies. The second book in the series, Envisioning Information (1990), was about pictures of nouns--that is, about representational rather than numerical reality: the depiction of cartographic information, diagrams, and signage. (What's the best way to show sunspot activity, for example, or the working of a subway system?) The third book, Visual Explanations, is about pictures of verbs--that is, about displays that illustrate dynamic processes and can therefore function as explanatory narratives.
T hese explanatory narratives, these pictures of verbs, can offer remarkable insight when properly conceived. Tufte's inclusion of John Snow's famous map of the 1854 cholera epidemic in London is a case in point: Snow's street-by-street tracking of cholera deaths pointed unmistakably to the culprit--the water coming out of the Broad Street pump. The pump handle was removed at once, and the epidemic ceased.
In contrast, these verb pictures can also smother insight when ill-conceived--as were the charts prepared by the manufacturer Morton Thiokol to illustrate the behavior in cold weather of the O-rings on the space shuttle Challenger. These charts, which failed to convince NASA officials not to launch the shuttle, were confusing--dense with invitations to second-guess, even though crucial information that would have delayed the launch was undeniably "there" (somewhere). The O-ring itself, Tufte makes clear, was not the only design element whose malfunction contributed to the Challenger tragedy. After presenting his own version of what the charts should have looked like, Tufte writes, "There are right ways and wrong ways to show data; there are displays that reveal truth and displays that do not."
What makes a visual display effective? Tufte lays out rules about such things as density and emphasis, chartjunk and clutter, layering and hierarchy, color and parallelism--all of which have synesthetic grammatical analogs in writing. (Hierarchy, for instance, is the building of large structures out of small ones, be they nautical charts out of soundings or sentences out of morphemes.) Aware of the parallels between visual grammar and written grammar, Tufte frequently makes direct comparisons that enhance one's understanding of each. In one place he shows how the rhetorical shape of a passage in Gibbon's Decline and Fall of the Roman Empire, with its paired images and a ribbing of parallel verbs, finds a visual counterpoint in works of landscape design and even dance notation. (To get the full picture, so to speak, you'll have to look at the book yourself.) An entire chapter is devoted to the concept of "smallest effective difference," showing how distinctions that are subtle but clear (two light colors on a map, say) can be far more powerful visually than any amount of heavy-handed contrast--as useful a lesson in writing as it is in graphic design.
W hat one does not expect in a book of this kind--a book that is technical even as it is beautiful--is that the author will be not just fastidious but also an evocative and wonderfully quirky writer. Tufte is all of these. In his discussion of one historic work--Jean Buchon's comparison of American rivers in his 1825 Atlas Geographique--Tufte notes how the depictions of the Mississippi and the Amazon must curl around the page, since there's no room to show them stretched out like the other rivers. Actually, what Tufte says is, "[T]he two rivers meander boustrophedonically around a tight frame." That word boustrophedon describes writing that goes from left to right on the first line, then right to left on the second, then left to right on the third, and so on; it comes from a Greek word describing the turning in a field of an ox and plow. Tufte, too, likes his etymologies.
Cullen Murphy is managing editor of the Atlantic Monthly and also writes the comic strip Prince Valiant.