Creating Chaos

How did an obscure article in a 1960s atmospheric science journal become part of the plot of a 1990s box office hit?

In 1963, Edward Lorenz claimed that the extreme sensitivity of complex systems to initial conditions seemed to make long-term weather forecasting impossible, that is, that sensitivity to initial conditions created unpredictability. In 1993, Michael Crichton used this same notion of unpredictability, now known as chaos theory, to create a bio-technological disaster. The journey from scientific journal to Jurassic Park was long and complex, but clearly demonstrates society's fascination with this new science.

James Gleick's popular 1987 book, Chaos: Making a New Science, traces chaos theory from obscurity to its establishment as a new discipline, looking at the main American players in the scientific community and their important results. Gleick shapes his historical account into an argument: He contends that chaos theory is a revolutionary way of looking at and doing science.

True or not, the scientists studying chaos theory often deliberately invoke the language of revolution. Having noticed this consistent use of language, Davida Charney, an associate professor in the Penn State English department, and I determined to retrace chaos' journey from obscurity, looking at the rhetorical issues of creating a scientific revolution.

By rhetorical issues, we mean the strategies scientists use in their writing to convince others that their findings are not only valid but also important or, in this case, revolutionary. Along with writing style, these strategies include such things as the use of jargon and the coining of new terms, the use of citations, and the creation of a context for their work by shaping their own historical account of the field.

The introductions of journal articles offer a natural place for scientists to shape their story. Generally, scientists use a "create-a-research-space" pattern in these introductions, as documented by John Swales, a linguist who specializes in advanced writing. According to Swales, scientists use four standard rhetorical moves to create a context for their work. First, they demonstrate the interest or importance of the research topic. Second, they selectively review and summarize the previously published research literature. Third, they show that the research is not complete, creating a gap in the previous research. And fourth, the current research is presented as a timely and appropriate "filler."

This model provides a useful pattern for established sciences; yet how do scientists doing "revolutionary" work tie their research to a non-existent past? When we analyzed the introductions to journal articles by two key chaos theory players, Mitchell J. Feigenbaum and James A. Yorke, we found that as the field became more established Feigenbaum and Yorke seemed to follow these standard rhetorical moves closely; in their early articles, however, they came up with a few moves of their own.

Perhaps the most interesting move in both articles is the first, in which writers usually attempt to demonstrate the scientific community's interest in their project by citing previous research. Both scientists in our study used an equation to create interest, but differed in how far they were willing to follow this unusual approach.

For example, in Yorke's earliest article, "Period Three Implies Chaos," from 1975, Yorke used the unusual first move, then returned to the conventional pattern. Since there was no established field, Yorke found very little past research to draw from to show interest in his project. Of the 17 articles he cited after the equation, six had not yet been published, and four were self-citations. So rather than demonstrate interest, Yorke created it. The lure of Yorke's provocative title, which eventually gave the phenomenon its name, is reinforced by a mathematical puzzle: Yorke presented what seemed to be a relatively simple differential equation, which, after several iterations, became chaotic. Yorke then shifted to a traditional second move, establishing some ties to his community.

Feigenbaum had a different approach. He made no attempt to attach his research to an as-yet-insignificant past. In "Universal Behavior" (1983), he made a bold move. In almost a reversal of the Biblical account, he simply created chaos out of order, firmly attaching his article to the other side of the entire enterprise of western science since Newton. "There exist in nature processes that can be described as complex or chaotic and processes that are simple or orderly," he wrote. In the first move of "Quantitative Universality" (1978), his best-known article, he set up a fictitious population as an example of a recursion equation and continued to use it throughout his introduction. Consequentially, his introduction looks like a long mathematical proof rather than a series of rhetorical moves. With the relentless logic of mathematics and without citations, he established his claim for a universal function. Feigenbaum's rhetorical message is clear. His research is something completely different from what had come before.

How well do these rhetorical moves work? From the perspective of a straight citation count, they both work very well. Yorke's paper has been cited over 400 times and Feigenbaum's "Quantitative Universality" a whopping 1,100 times. (To give you an idea of how rare that count is, the average scientific paper is cited only once or twice per year and only two in 10,000 scientific papers are cited more than 500 times in their lifetimes.)

Naturally, how the story is told is only one of many factors that account for the success of an article. To determine a reader's response to the individual rhetorical moves, we had 12 scientists from two large state universities read aloud and comment on various combinations of these two early articles and two more-recent articles by the same authors.

For readers 20 years after the fact, Feigenbaum's brash approach seemed more appropriate. One reader of the article stated: "It reads like an announcement." Another commented: "You rarely see papers that are constructed of raw ideas." On the other hand, Yorke's shift back to a traditional approach seems to backfire. For our readers, seeing the form they expected but little new information led many of them to classify the article as "simple," "generic," or containing "textbook information." Finally, even the claim made in Yorke's provocative title that "Period Three Implies Chaos" was challenged. One reader familiar with the article noted some problems with its content, explaining: "So [Yorke's claim] is [actually] 'Period Three Implies something,' but it's not 'Period Three Implies what-is-generally-taken-as-the-current-definition-of Chaos.'"

Perhaps this comment explains our interest in the intricate negotiation between science and language in chaos theory. Even as theory changes and develops, language, with its fluctuating meaning is a constant. Yorke named the phenomenon "chaos" to attract attention for the new enterprise. And, despite the dethroning of Yorke's "Period Three" phenomenon, and despite some researchers who gruffly say, "We don't call it chaos theory," chaos it remains.

So in making that journey from journal articles to Jurassic monsters, our findings-to-date indicate that, along with presenting interesting science, how one tells the story does seem to matter. And a provocative name doesn't hurt.

Danette Paul is a Ph.D. candidate in English, 103 Burrowes Building, University Park, PA 16802. This essay was prepared from papers she presented at the 1991 Penn State Rhetoric Conference, and the 1993 Conference on College Communications and Composition.

Last Updated March 01, 1995