- Muses and Measures: Empirical Research Methods for the Humanities
Willie van Peer and his colleagues have produced an important, innovative book that ought to be widely studied by humanities scholars. Its central thesis is that humanists need to think as hard about their methodological assumptions as they do about their theoretical and hermeneutical ones. For Van Peer et al., the problem with standard humanities methods is that they are "often merely speculative, in the sense that there are very few checks on the assertions made," and therefore "they do not yield particularly reliable forms of information" (p. 7). While this is in itself a pedestrian observation, the authors' proposals for addressing this limitation are not. They boldly argue that literary scholars can do more than borrow ideas, findings, and vocabularies from the sciences—they can borrow the methods too.
Muses and Measures is a call for humanities scholars to actually do science. If the authors have their way, ordinary humanities jargon will expand to include wundt curves, likert scales, within-subject and between-subject experimental designs, normal distributions, descriptive and inferential statistics, t-tests, p-values, and analysis of variance (ANOVA). This prospect will strike most humanities scholars as nightmarish – the latest in a depressingly long string of foolish, doomed attempts to cram humanistic inquiry into science's alluring but ill-fitting mold. For two reasons, however, readers should resist any reflex of fear or disgust until they actually read the book.
First, there is a fundamental difference between what Van Peer et al. are proposing and previous attempts, like psychoanalysis or structuralism, to establish sciences of the literary. While previous schools have imported concepts, vocabularies, and an aura of rigor from scientific fields, they have almost never sought to apply the methods of the sciences. These days, almost everyone thinks that barriers between disciplines are permeable (or illusory), and almost everyone celebrates work that links knowledge from disparate disciplines. But humanities scholars have almost never seriously tried to move the methods of the [End Page 393] sciences across the disciplinary divide. The novelty of Muses and Measures—its quiet radicalism—is in how it shows that those methods can be judiciously and fruitfully applied in the humanities.
The second reason that humanities scholars should resist the impulse to reject Muses and Measures is that its authors are reasonable people: they are not so wild-eyed or silly as to think that every important humanities question can be reduced to numbers, fed into computers, and solved by clever algorithms. Neither are they trying to foment a conquest of art/scholar culture by science culture. Rather, like C. P. Snow, they are seeking to establish a Third Culture on the fertile ground between the two cultures of the humanities and sciences. The authors respect the indispensable role of traditional methods of close reading and careful reasoning—their goal is to add a new set of tools to the standard repertoire of humanistic study, not to throw out the old ones.
Muses and Measures may be a manifesto, but it is not written as one. The book is a how-to manual, intended to guide beginning students through the processes of empirical research, from "problem formation," through study design and data analysis, all the way up to the presentation of results in conference and journal papers. Along the way, the authors provide basic, chapter-length primers on the different aspects of empirical research, including diverse methods of data collection (observational, experimental, survey, content analytic, etc.), detailed information on how to work with that data using the statistical software program SPSS, and a short course of four chapters on basic statistics.
One fact amply demonstrated in Muses and Measures is that the digital revolution has put scientific methods within the practical reach of all but the most technologically clueless humanities scholars. Statistical programs like SPSS mean that researchers no longer need to tabulate numbers by hand. While researchers still need to understand statistical concepts and the reasons for using one...