We have seen RoSE grow into a linked collection of references to documents and people of various eras, and we know intuitively that each of these entities, living or not, has a story, has significance. Or more precisely, we can study and ascribe meaning to them separately and in clusters. Doing so might also be called a form of storytelling, or making sense of the data by tracing a certain vantage point or line of inquiry. The very acts of telling and seeing have deep ties to narratology. However, if narrative can be a macrocosmic model for organizing thought about disparate data points, it is also worth considering whether narrative information can be encoded as metadata. Such an endeavor could conceivably facilitate the use of RoSE to grasp the development of discourses and movements in the disciplines now represented in the system. That is, adding contextual and structural information to the markup of the data in a way that it could be leveraged algorithmically may help users do more inductive reasoning–to go from known connections in a mass of detail to an unforeseen overarching pattern with its own provocations.
Indeed, narrative is a kind of pattern, providing pre-established structures or cognitive schemas that can be overlaid on many kinds of information, though usually the information in fiction. Admittedly, applying narratological ideas to non-fictional areas risks a kind of willful projection, an unwitting Rorschach test. Even so, it may be worth suspending this skepticism to find out what conceptual gains are to be had when conceiving of narrative as metadata.
Information structure and narratology
To provide an initial way of framing this discussion, some thoughts from the domain of library and information science (Gilliland, 2008):
In general, all information objects, regardless of the physical or intellectual form they take, have three features–content, context, and structure–all of which can and should be reflected through metadata. – Content relates to what the object contains or is about and is intrinsic to an information object. – Context indicates the who, what, why, where, and how aspects associated with the object’s creation and is extrinsic to an information object. – Structure relates to the formal set of associations within or among individual information objects and can be intrinsic or extrinsic or both. (p. 2)
The role of structure has been growing as computer-processing capabilities become increasingly powerful and sophisticated. Information communities are aware that the more highly structured an information object is, the more that structure can be exploited for searching, manipulation, and interrelating with other information objects. Capturing, documenting, and enforcing that structure, however, can only occur if supported by specific types of metadata… But there is more to metadata than description and resource discovery. A more inclusive conceptualization of metadata is needed as we consider the range of activities that may be incorporated into digital information systems. (p. 6)
From these remarks, the role of structure is of special interest to RoSE. Documents and people themselves can provide narrative structure; they can also be fit into such structures in relationship to each other. Thus, narrative seems to be both an intrinsic and extrinsic structure, making it a potential lens for varying the level of granularity that data is presented at. Expanding the conceptualization of metadata to include narrative and narrative structures requires a better understanding of established theory in this area. “The basic distinction of the theory of narrative, then, is between plot and presentation, story and discourse,” according to Culler (1997, p. 86), who went on to make another observation:
Confronted with a text (a term that includes films and other representations), the reader makes sense of it by identifying the story and then seeing the text as one particular presentation of that story; by identifying ‘what happens’, we are able to think of the rest of the verbal material as the way of portraying what takes place. Then we can ask what type of presentation has been chosen and what difference that makes.
Thus, the data represented in RoSE can be thought of as a story, while RoSE is something that allows discourse to emerge through links, visualizations, and other usages or portrayals. The utility of this formulation is in its disclosure of the limitations of RoSE, or of any attempt to apprehend “what happens” or what exists–understanding is constrained by the point of view taken. In turn, variables of time, distance and speed, and scope of consciousness are all factors influencing the perspective on events in a story (Culler, 1997). In narratological terms, this is the focalization of a story (“who sees?”) as opposed to the narration (“who speaks?”). “Functionally, focalization is a means of selecting and restricting narrative information, of seeing events and states of affairs from somebody’s point of view, of foregrounding the focalizing agent, and of creating an empathetical or ironical view on the focalizer” (Jahn, 2005). Admittedly, this explanation is rather narrow and does not include definitions that emphasize the thoughts and emotions of the focalizer, since those qualities do not lend themselves to the present purposes. Indeed, the human-centric nature of narrative (Fludernik, 2009, p. 6) is one argument against applying a narrative lens to RoSE, which does not intend to favor people over documents.
For RoSE, focalization allows the imagining of new user scenarios: assuming the perspective of a certain author, say, and partially recreating her or his worldview at a particular moment by filtering out works not yet published, or somehow having prevalent influences loom at the periphery. Various views or groupings could be saved for sharing and revisiting. It might then be said that sense-making through RoSE can be a matter of choosing the eyes through which to see a succession of data. Of course, this is still problematic since it is impossible to think of documents or groups as having a perspective in the same sense as an individual. Perhaps a document “sees” the entities that it references or that it is referenced by.
In any case, the metaphor of vision has certainly shaped RoSE (visualizations as means of presenting the data), and an awareness of focalization in narrative helps delineate the things constituting a viewpoint. The temporal variable has already been recognized through our use of timelines in the visualizations, but it may be worth returning to questions of distance/speed and scope later on.
Much more could be said about the origination of specific narratological ideas (Fludernik, 2009), but coverage here is necessarily abbreviated. From narratology, then, there is a rich vocabulary to be gained about under-problematized aspects of our data and the parsing of information about people and documents. If these features can be harnessed through metadata, they may offer advances not only in making user-driven discoveries but system-driven recommendations, as has been gestured toward in the project prospectus–”Then it comes clear: what you know and still do not know, and who or what next to research, to read.”
Narrative in relation to knowledge and ontology
In further considering the prospects of narrative approaches, it must be asked: “is narrative a fundamental form of knowledge (giving knowledge of the world through its sense-making) or is it a rhetorical structure that distorts as much as it reveals?” (Culler, 1997, p. 92). The constraints on narrative in terms of focalization have already been discussed. As long as such limitations are duly noted at the outset, there should be few qualms about narrative’s coloring of experience and data. Indeed, oral narratives/poetics are a traditional and effective form of knowledge transfer in many cultures (Bauman, 2004).
Some comments about documents in particular may contribute insight into the area of knowledge and ontology. Bibliographic catalogs have long struggled with these issues in the context of documents and other information objects, since metadata standards have not been able to capture semantic complexity. By performing a variety of functions, written documents allow people to establish or mirror complex social relations (Smith, 2005). Although catalogers debate the nature and classification of documents for aiding information seekers, this fundamental observation has rarely figured into their discussions. However, acknowledging the functional (and narrative) capabilities of documents may lead to better systems for retrieval and discovery, in that users of such systems are often interested not just in the title or author of a document, but what the document and others like it might be said to accomplish (such as storytelling).
Much of the scholarship on bibliographic ontology has been concerned with subject-predicate statements to be evaluated for their truth or accuracy–involving questions of provenance and notions of the original work, copies, editions and so on. While logically sound, this modeling is much removed from people’s everyday experience of documents and other sources of information. That is, a functional turn is needed to more fully account for relationships between these sources, which may also be social or interpersonal, as RoSE rightly recognizes. Ignoring the context in which informational entities are accessed, and to what ends, prevents us from seeing the narrative power of data.
On the subject of more sophisticated uses of catalogs, LeBoeuf (2005) has this to add:
In combination with other ontologies, and thanks to tools borrowed from the Semantic Web techniques, the incredibly rich but largely unexploited information dug into our catalogs could allow for such deductive reasoning or “inferences” as: If Person X was female and spent most of her lifetime in Victorian England, then the works written by that person can be regarded as a testimony of women’s condition and state of mind during the Victorian Era; or: If Person Y was friends with Person Z, and Z’s correspondence has been published, perhaps I will find information about Y in Z’s published correspondence… (pp. 5-6)
This sounds strikingly similar to a capability that might eventually be expected of RoSE; though not mentioned, it is evident that detailed structural metadata would be necessary to achieve this sort of reasoning. In other words, “metadata produced by annotating multimedia with respect to a shared ontological vocabulary will allow for searching and navigation by concept” (Tuffield, 2005, p. 2). Although RoSE has not stressed multimedia documents, neither has it precluded their inclusion in a semantically navigable network of some future stage.
Attempting to fit some documents, people, and groups into certain narrative structures may be helpful in crafting analogous paths or focalizations of the data despite variations in content. That is, using parts of these structures as metadata invites attention to parallel epistemic developments among the three period groups RoSE divided into for data entry.
There has been extensive inventorying of narrative models for those interested in technological applications (Kwiat, 2008), with the models loosely categorized through a grammatical conceptualization. The metaphor of grammar, like that of vision, has been a subtle but persistent presence in RoSE–the subject-verb-object specification of relationships between entities. Moving beyond grammatical modes (Kwiat, 2008, p. 69 ff.) is a leap that should be studied further. But for now, a preliminary effort is made to show in one model how distinct sets of data might be tagged so as to encourage the use of narrative structures as metadata. Branigan (1992) cited Todorov’s assertion that basic narrative is a five-part causal transformation of a situation (p. 4):
1. a state of equilibrium at the outset
2. a disruption of the equilibrium by some action
3. a recognition that there has been some disruption
4. an attempt to repair the disruption
5. a reinstatement of the initial equilibrium
Parts of this model or inversely related; thus, it can also be symbolized as A, B, -A, -B, A (Branigan, 1992, p. 5).
The sources of sample documents and people for applying this model will be the American Revolutionary War and post-structuralist literary criticism.
Although this rough draft of a narrative markup has avoided being human-centric in that the documents included have consequence, it is clear that a different kind of bias has surfaced, in the form of linear temporal progression. Another flaw is the inability to account for events, which are of course dynamic combinations of people, groups, and performances that are not always precipitated or recorded by documents. Far from creating a coherent narrative of a war or a literary movement, this exercise has shown the need for domain experts to contribute their understanding to the markup process.
It is also difficult to formalize this sort of schema given the subjectivity in labeling entities (with concern about the pejorative connotation of “labeling” applying here as well). For example, the Boston Tea Party might have been a reparative effort for the American colonists, a reaction to perceived injustices, but a disruption of equilibrium in the eyes of the British. This is again a matter of focalization. There are also instances in which the distinction between the third and fourth stages (recognition and repair) is unclear, or in which there is no obvious entity to be marked. The equilibrium model appears unsuited to describing the development of post-structuralism since there is no return to the same initial state (the irony of applying a structural markup to post-structuralism has not gone unnoticed).
In dealing with these complications, though, it may also be useful to keep in mind notions of Hegelian dialectic structure–a movement or proposition may act as the current norm (thesis) for rebellion against (antithesis), but at one time it too was a negotiated result (synthesis). The deficiency of one model of narrative structure in covering different sets of data highlights the need to investigate alternative models, just as alternative visualizations for RoSE were discussed.
Evaluation of Opportunities/Limitations for the Transliteracies Topic:
As developers, we might anticipate that RoSE could someday reverse engineer a more successful version of this markup, identifying what characteristics are salient among all entities marked as a given narrative/structural element and then suggesting other entities (and storylines, or research questions) based on those similarities. Then again, that expectation may be merely a wishful coda to the story of RoSE itself.
Bauman, R. (2004). A world of others’ words: Cross-cultural perspectives on intertextuality. Malden, MA: Wiley-Blackwell.
Branigan, E. (1992). Narrative comprehension and film. New York: Routledge.
Culler, J. (1997). Literary theory: A very short introduction. Oxford: Oxford University Press.
Flundernik, M. (2009). An introduction to narratology. New York: Routledge.
Gilliland, A. (2008). In M. Baca (Ed.), Introduction to metadata (1-19). Los Angeles: Getty Research Institute.
Jahn, M. (2005). Narratology: A guide to the theory of narrative. English Department, University of Cologne. Available online at: http://www.uni-koeln.de/~ame02/pppn.htm.
Kwiat, J. (2008). From Aristotle to Gabriel: A summary of the narratology literature for story technologies. Technical Report KMI-08-01, Knowledge Media Institute, The Open University, UK. Available online at: http://kmi.open.ac.uk/publications/pdf/kmi-08-01.pdf.
Le Boeuf, Patrick. (2005). FRBR: Hype or cure-all? Introduction. Cataloging & Classification Quarterly 39(3/4): 1-13.
Smith, B. (2005). How to do things with paper: The ontology of documents and the technologies of identification. Paper presented at Ontolog Forum, Oct. 13, 2005. Available online at
Tuffield, M. (2005). Narrative as a form of knowledge transfer, narrative theory and semantics. Progress report submitted to doctoral committee at the University of Southampton (Intelligence, Agents, Multimedia Group). Available online at: http://eprints.aktors.org/423/01/AKT_DTA.pdf.