In an oft-cited lecture on ‘The Application of Thought to Textual Criticism’, A. E. Housman remarked:
A textual critic engaged upon his business is not at all like Newton investigating the motions of the planets: he is much more like a dog hunting for fleas. If a dog hunted for fleas on mathematical principles, basing his researches on statistics of area and population, he would never catch a flea except by accident. They require to be treated as individuals; and every problem which presents itself to the textual critic must be regarded as possibly unique.Footnote 1
While the practice may not have been widespread in the 1920s when Housman was writing, biological researchers now routinely track the movement and behaviour of fleas using mathematical models.Footnote 2 Despite our best efforts, dogs remain frustratingly unable to perform even rudimentary statistics.
In fact, the groundwork for such investigations had been in place since 1907, when the Dutch physicists Paul and Tatiana Ehrenfest introduced a statistical model to describe the jumps of a given population of fleas between two dogs,Footnote 3 a model praised as ‘probably one of the most instructive models in the whole of Physics’Footnote 4 and ‘mentioned in almost every textbook of probability, stochastic processes and statistical physics’.Footnote 5
Although he may not have been familiar with the statistical physics literature of his day, Housman was no luddite; he was simply refuting the idea that the work of the textual critic was essentially systematic and a matter of following abstract rules. Given its materials, textual criticism, according to Housman, must be a series of separate individual responses to unrelated local problems. In stylistics, as in textual criticism, each problem has aspects unique to itself. Yet each one also has enough aspects in common with others (we believe) to make comparison and aggregation feasible. Acknowledging these two truths, our own research practice is to combine the traditional methods of the literary historian (some of which, like close reading and analytical bibliography, overlap with methods Housman would recognise) with new modes of analysis (e.g. text mining and data visualisation, computational stylistics, multivariate statistical analysis, and algorithmic criticism), and to do so critically, aiming to apply the logic and common sense to which Housman appeals.
This combination of qualitative and quantitative methods allows us to shift between microscopic and macroscopic modes of inquiry – to zoom in on an individual flea in isolation, to zoom out to observe relationships between groupings of fleas, or to zoom out further to appreciate the larger ecology of the dog – to explore the fine details, as well as their contexts.
Clifford Geertz describes this ‘characteristic intellectual movement’ as ‘a continuous dialectical tacking between the most local of local detail and the most global of global structures in such a way as to bring them into simultaneous view’. For Geertz, this ‘inward conceptual rhythm’,
Hopping back and forth between the whole conceived through the parts that actualize it and the parts conceived through the whole that motivates them, we seek to turn them, by a sort of intellectual perpetual motion, into explications of one another.Footnote 6
We are aware that our work is a long way short of offering a unified theory of dramatic language, or of early modern drama – that we have remained closer, perhaps, to the flea-biting dog than to Newton describing planetary motion. Nevertheless, we have tried to combine the local and the general in the way Geertz describes. We have also tried to work between the qualitative and quantitative poles of method. In this spirit, we like to think that our findings about particular areas – about when verse and prose contrast in dramatic style, and when not; about the separate stylistic identity of dramatic characters; about the distribution of props by author, genre, and company; about collective stylistic changes over time; about company style; and about the way 1660s plays map onto Elizabethan, Jacobean, and Caroline drama – may be of interest to those who will never conduct a t-test or a PCA. Better yet, our findings may even prompt some readers to try an experiment of their own to test our claims. Our hope is that the chapters of this book serve to illustrate the possibilities for a future mobile criticism, in which mainstream and computational methods will be able to test and invigorate each other, lead to better answers to some specific literary questions, and uncover hitherto hidden aspects of the working of literary language.
