An introduction to an article accepted to New Literary History in summer of 2017. Cite as Dennis Yi Tenen, “Towards a Computational Archaeology of Fictional Space,” forthcoming in New Literary History.

Space is a hard thing to pin down. It identifies dimensional continuity and a topography, that is, a relationship between objects. It is also itself an object: a limit-defining quantity even in its most abstract sense. “O God, I could be bounded in a nutshell and count myself a king of infinite space, were it that I have bad dreams,” Hamlet says of his ambition and his dreams.¹ A human palm can be a part of the body or a map. A mirror is a piece of furniture and a frame for reflection. Under extreme magnification, the head of a pin appears a vast and mountainous terrain, home to angels and bacterial detritus. The characterization of diegetic—let us call it also virtual and fictional—space presents further difficulties. A stretch of land in fiction measures also a stretch of the imagination. These units do not always have names or explicit boundaries. Vladimir and Estragon wait for Godot: “A country road. A tree.”² Two vectors are enough to situate the world. A road gives us the X and a tree the Y axis: an infinity in a nutshell.³

In this paper I propose to reconsider theories of diegetic space which rely on explicit framing (i.e. “two people walk into a room” or “in Spain”). Rather than looking for maps, I define space in terms of grammatical categories denoting objects. The emphasis on objects leads to a method for literary archaeology, informed by cognitive theory and anthropology. If the universe is made of atoms, a fictional world is also made up of atomic relationships that form basic, stable configurations, or, what I call, narratological primitives. I construct several such basic spatial buildings blocks here—diegetic density and narrative clutter. Their application to a well-explored body of Victorian novels challenges several long standing historical intuitions related to the development of material culture in the nineteenth century.

A theoretic reconfiguration grounds a descriptive method: a model by which fictional space can be, if not fixed, then approximated. Further, following Lisa Samuels, Jerome McGann, Johanna Drucker, and others, I am interested in literary modeling and visualization as kinds of a transformative reading practice. The formal, computational methods I present here are not meant to prove anything. They are first and foremost exploratory and experimental tools. They occasion opportunities for close reading or “reading against the grain” and not just reading at scale. My methods are diagnostic, in that they identify areas of interest and unusual trends that require closer critical attention.

An undercurrent and a starting place of this essay is therefore also a critique of a certain mode of quantitative literary analysis, which essentially advances a number of complex methodological procedures to arbitrary effects. Formalists old and new are perpetually and cyclically in danger of falling into the trap set by Stanley Fish in the 1970s. Fish cautioned that a relation can always be found between any number of low-level, formal features of a text and a given high-level account of its meaning. For example, the use of past participles or gerunds may rise and fall with the vagaries of literary style. Or it may be due to changing archive collection practices over time: sample bias. As Fish puts it: “there are always formal patterns” and “a relation will always be found.”⁴ This is astute. Methods will always produce results. The difficulty lies in filtering meaning from the noise. A whole subset of statistical methods—causal analysis—is dedicated to this problem. It involves, perhaps unsurprisingly, explanatory frameworks that chart a partway between dependent and independent variables. Methods require theory. A tautological formula cannot in itself produce meaning: one must have an a priori idea about what is meaningful. Correlations become more convincing with the interpretation of causes. Analytics, in other words, are meaningless without poetics.

To get past Fish, we must first understand the difference between prediction and explanation. Karl Popper wrote that the aim of theoretical science is to find “explanatory theories […] which describe certain structural properties of the world.”⁵ A theoretical interest in explanation is non-instrumental, insofar as it is “irreducible to the practical technological interest in the deduction of predictions.”⁶ A diagram of a storm weather system holds explanatory and not just a predictive power. In fact, it does not describe any specific weather systems at all, past or future. Rather, a diagram teaches us something about the relationship between causes and effects: high pressure, condensation, wind, and rain. To understand how something works—to form a theory—further entails the possibility of effecting systemic change. To understand disease pathology is also to imagine ways of inhibiting it. In this way, explanatory models reach beyond the apparent phenomenon: from what is to what might be. Models contain the remainder of the real, which is poiesis itself, creativity.⁷ What is the point of theory? one might ask. There is none, Popper answers. Theory is a storehouse of potential applications, knowledge for knowledge’s sake.

Even a simple forecast places ideational constructs in relation to the empirical world. Predictive power and explanatory power reciprocate one another. To predict is also to imagine, albeit in a more directed, fully potentiated way. Pattern recognition suffices for this purpose. One can forecast the sun will rise each day based on past experience without knowing anything about planetary mechanics. Patterns allow us to extrapolate from known contexts to similar unknown ones. A weather model (of earth) will produce accurate predictions provided that the planet remains more or less the same.

Not all theories are predictive, however. The interpretation of historical events involves an account of singular causes and effects that do not repeat in the same configuration. The interpretation of literary texts also hinges on deeply contextualized, affective, or embodied dynamics. Predictive power may or may not be necessary to understand how a text works, where explanatory power is. To know how texts work, in the echoing words of Percy Lubbock, Boris Eichenbaum, and Susan Sontag, is the essence of poetics.

I offer these reflections on method on route to an argument about cultural analytics—the application of computational methods to the study of literature and culture—generally, and, more specifically, in an attempt to outline some strategies for explanatory modeling of space in literary texts. Some research questions, I maintain, are amenable to predictive analysis. Others require explanatory power. The two complimentary approaches involve different logics and modes of argumentation. It is important however, to differentiate clearly between the two. The worst kind of error is one where predictive results are taken for explanatory ones: the sun will continue to rise because it has been rising regularly. The “because” is unwarranted. We must not mistake mere extrapolation for an account of deep causes and effects. The understanding of culture, to paraphrase Clifford Geertz, cannot be limited to pattern recognition. It is rather an interpretive effort, “in search of meaning.”⁸ As such, it requires other attributes besides statistical significance or convenience, such as simplicity, novelty, and persuasiveness. An explanatory, interpretive model does not just extrapolate, it produces insight.