Characteristic curves and counting machines: Assessing style at the turn of the century

Abstract

From 1887 to 1904, Science and The Popular Science Monthly published a series of articles on the characteristics of composition, articles pointed out an uneasy alliance between scientific methodology and traditional rhetorical assumptions about the nature of writing. This turn-of-the-century dispute between Robert Moritz and Thomas Mendenhall shows how early researchers in rhetoric cast their inquiries in scientific terms to gain the legitimacy of scientific findings. Further, the little-known debate can be read as a precursor to the many debates followed over whether and how quantitative methods can resolve questions in rhetorical inquiry, and more fundamentally, what vision of language underwrites the assumptions of such methods. The debate I want to focus on began with Mendenhall, who first argued for his version of composition analysis in the March 1887 issue of Science.' In this article, as well as in his papers of 1901 and 1904, he seeks to prove his hypothesis (based on a remark by Augustus DeMorgan) each author has a of composition. The curve is based on the frequency with which an author uses words of different lengths, is, one-letter words, two-letter words, and so on. The number of words of each length, when tabulated and graphed, shows the characteristic curve of author. Mendenhall was a noted nineteenth century scientist who published in a wide variety of areas, including geology, geography and science education. He was a president of Worcester Polytechnic Institute and served for a time as an editor of Science. Thus a scientific approach to composition was for him merely an extension of methods he used in his daily work. Indeed, he originated his theory based on an analogy to spectral analysis, a method of determining the elements of a given physical substance. Just as each element gives forth a group of waves of definite length, and appearing in certain definite proportions (Mendenhall 1887, 238), so each person's style is marked by numerical constants--the frequency of words of each length. This analogy with spectral analysis reveals not only Mendenhall's procedure but also his notion of the inevitability of style. The style of people, like of elements, is determined by their nature and not their mode of existence; is, their texts are determined by inevitable displays of their fixed personality and not their rhetorical choices. The merits of this approach, according to Mendenhall, are that it offers a means of investigating and displaying the mere mechanism of composition, and it is purely mechanical in its application. (Mendenhall 1887, 245). Mechanism makes impossible human failings of choice and error: by focusing on a meaningless aspect of composition, the researcher can avoid the effects of deliberate changes by the author; by using a mechanistic data gathering method,