How many discrete units does it take to make a database? For plot there needs to be at least two distinct, but related events.  E. M. Forester famously identified the difference between plot and narrative with the following sentences: “The king died, and the queen died,” is a narrative; “The king died, and the queen died of grief,” is a plot. One set is relational, the other is not. Can a database narrative be made of only two relational elements? We associate databases as potentially unlimited sets, always changing and expanding. 

The word “database” connotes a vast repository of stuff and “narrative” an ordered selection. According to Manovich’s distinction, a narrative hides its set of source material (paradigm) and makes visible the selection (syntagm). Alternately, a database foregrounds the paradigmatic set and hides the syntagmatic or narrative ordering.  But, linear narratives can be endlessly inclusive and digressive and a database can be a limited aggregation of data. 

“story is generally organized through absence. Put another way, absence is presence. That seems very much at odds with computer data. But think of the problem this way: absence is a kind of aperture.”

Missing data in a story—the background of the characters, the off-screen or offstage action, the unspoken thoughts, the hidden influences on the plot, the cultural contexts of the telling—are essential for story to be present in the mind. Absence is indeed an aperture. We experience story in the brain as a neural network—a field of semantic and sensory effects, of which plot may be one unifying element. The storyteller starts with charged mental data — voices, images, sensations, abstractions—and then arranges selections of data into a material presentation (a narration) so that it can be delivered to an attentive mind. The material form of a story is always the compressed (zipped) version of a living network designed for activation inside other living networks. While plot provides the scaffolding, it is ultimately our cognitive and emotional investment—the associations, desires, visualizations, decodings and fast searches—that transforms a mere series of selected details into a story structure that is more than the sum of its parts.  

Samuel Beckett’s nearly plotless novel “Molloy” begins with more missing than useful data.  And yet the lack of information engages the reader on many levels about setting, character and plot. 

“I am in my mother’s room. It’s I who live there now. I don’t how I got there. Perhaps in an ambulance, certainly a vehicle of some kind. I was helped. I’d never have got there alone. There’s this man who comes every week…”

The novel proceeds only to hint at causes—an off-stage plot. The voice of Molloy, one might say, is an interface to the character’s database-consciousness and the missing data, erased for reasons the reader may never understand, can “fill” the consciousness of a reader with similar self-editing capabilities.

The missing data of a database dictates the what, how and why of its existence. A database is, typically, organized around the relational aspects of its data. That is, the selection process and tagging system of a dataset is foregrounded so that information retrieval can happen quickly and efficiently. Information from a bank account mixed up with a database of addresses would complicate and disrupt search and retrieval.  A set that has no apparent relations between items or whose relational categories are not clearly available, raises questions. What makes these items a set? Who made this set and why? How is one supposed to use this set? A collection or archive can be based on proximity (the contents of a drawer) or on culturally defined categories (antique European furniture) or on a collector’s own tastes and whims (cabinets of curiosity).  “Data monumentalizes exclusion.”   Like any linear narrative, the selection and identification of the relational sets of a database tells its own story; what is included and excluded is always a culturally and politically informed process. A census, for example, might segment human beings into categories—race, religion or sexual-orientation—that may not translate across cultural narratives of identity. 

Like the ellipses and ellisions in modernist and postmodern fiction, a database fiction might benefit from a certain amount of confusion about categories, selection processes and even the reliability of navigation tools. A syntagmatic rather than paradigmatic approach to data selection might raise questions and form schemas about its author. Data that is excluded might call attention to and form queries about its narrative importance. The limitations in the amount of data might make exploration more accessible and focus attention on the qualities and gaps within of a finite set. Like Haiku poetry where a constricted list of seemingly unrelated images and sensations can have lyrical resonance, the aggregation of a database can open gaps between things and resonate on narrative levels.