In lieu of an abstract, here is a brief excerpt of the content:

  • Empirical Questions Deserve Empirical Answers
  • Colin Martindale

What is wrong with the current state of humanistic literary studies? On the theoretical level, we find various types of postmodernism, none of which makes much sense. On the other hand, there are approaches such as Marxism, Feminism, and the New Historicism. One can at least understand the contentions of such theorists, but these contentions are generally quite implausible. If poetry were an effective means of maintaining the status quo or of oppressing anyone, one wonders why this was never noticed by politicians or generals. Now that the fact has been “discovered,” one wonders why it is not put to use by the power elite. What has led literary theory to stray onto such strange grounds? I believe that the faulty methods used by many literary theorists offer us at least part of an answer.

Humanistic critics have no reasonable way of deciding among alternative answers if theorists happen to ask a question that is answerable—i.e., if they propose an empirical investigation. Their only method is one of producing examples and trying to persuade the reader. 1 Can a scientific approach provide something better? Certainly it can. First, the scientist does not pose unanswerable questions. Second, science provides methods for resolving disputes, for answering questions that are answerable. These methods are the scientific apparatus of counting, measuring, experimenting, and so on. Humanistic inquiry is simply not as effective as scientific inquiry. If a question is meaningful, it is by definition an empirical question. Science is the only method to us available for answering such questions. Humanism has no decision-making methods, no objective methods of choosing amongst competing hypotheses. [End Page 347]

To paraphrase Plato, literary critics hypothesize many fine things but know nothing of how to test them. Science already has the methods. We merely need to apply them. To give an example from my own research, Raymond and a number of other literary scholars tried to say that across the course of the nineteenth century, the content of French poetry became more primary process (Freud), dedifferentiated (Werner), associative (Wundt), or mythic (Cassirer) in its content. 2 Unless one can translate Raymond’s comments into these theoretical terms, it is difficult to understand what in the world he is talking about, let alone to test his contention. Because literary critics usually make up their psychology as they go along, it is difficult to understand what they say. Because they have only a proof-by-example method, it is difficult to believe what they say. If we do make the translation and develop a metric to measure the degree to which poetic content is primary process, we can test Raymond’s hypothesis and find out if it is correct. 3 We can not only find out if it is correct. We can fit curves to the data, compute rates of change, relate rates of change to other variables, and so on. One cannot fit curves to what Lotman aptly called “literary chit-chat.” 4 Without such an objective measure of primary process content, one would no doubt assert that twentieth-century poetry is even more mythic in its content, because the poets themselves have said so. 5 In fact, the metric confirms Plato’s warning that poets lie.

Some literary theorists, of course, do things that are at least close to science. Numbers would be of great help to such theorists if only as mnemonic devices. Given that people cannot hold more than about seven things in mind at once, the task of sorting out and comparing hundreds of texts is very difficult unless some counting is done. 6 Humanists can count, but they seem not to realize the value of this ability. Lévi-Strauss and the structuralists were doing what might be called zero-degree science. 7 It is reasonable to segment texts and look for regularities. It is not reasonable to look for regularities before one has determined the reliability of the segmentations: that is, before one has determined whether people agree as to how to segment texts. To compute a reliability coefficient, one has to use numbers, and one has to know the formula. Humanists don’t...

Additional Information

Print ISSN
pp. 347-361
Launched on MUSE
Open Access
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.