My understanding of their vision is to use lessons from well structured hypertext where some of the navigation ordinarily handled as links is instead managed by physical transit. Also, the texts at locations may employ 'good writing' about place, engaging writers and historians. We brainstormed a bit in the workshop and my mind hasn't stopped since. Quite apart from that project and their goals, this post collects some extended thoughts and a proposal for a new Digital Humanities-centric project.
Let’s assume that the delivery to ‘readers’ is by cell phones, tablets or laptops, and the interface is via a dedicated app. That is, we aren’t limited by browser technology. Suppose we were in Boston and were able to leverage the Boston Maps Project, among other resources.
This was written quickly and is a real brainstorming ramble. It would need a lot of work to be presentable, but I wanted to get some of these ideas out there for your consideration. Feedback is welcome. Email me at ted g at alum dot mit dot edu (without the spaces).
The Basic Vision
The big picture: users can enter the app or web page from anywhere and get ‘the next part’ of one of the stories, depending on person, place, history (of the person and place) and possibly global states. (An example layered story structure is given near the end, and you might want to skip to there.)
If they go to a hub location, they can get bits of their story-in-progress or a related story, anchored to place in some way. Individual nodes and hubs (collections of nodes) are sprinkled through the city. Not all of the nodes, whether in a hub or not, will be advertised; these must be discovered opportunistically.
Some nodes will not be accessible unless others have been visited. The overarching story and story world is available to everyone, but nested stories, versions of stories and perspectives into a story may be unique to the person, how they got there and who they know. In this sense the discovery is something of a game.
Many hubs/nodes will be at locations where students will be in the normal course of their lives. A hub is typically no larger than can be traversed in a couple hours, but on a visit, some nodes may be available and others may since have been born. After collecting enough of a story, users can write new stories, extending or branching from what exists; some of these additions need to be spatially anchored. (Some tags on content are required, adapted from the fan fiction community.)
Some of these stories will be historically informed by events that have spatial association informed by the Boston Maps Project.
Some of the stories will be by professional writers and could require a micropayment at branch points.
The anchor audience is students. Cooperating research groups are invited to create narratives and narrative discovery projects for special communities; a cohort of interest are recent immigrants who lack a sense of place.
On Scale and Narrative Component as Agent
The original vision from the UK guys, as I understood is to provide a closed experience that can be consumed in an hour or two. That limits the scale to a couple acres or less. We will scale from that vision in two ways, creating many hubs, and many narratives. These are related one to another, for example you cannot advance in key elements of the story unless you physically go there, or hear from someone who did. But that second-hand knowledge may not match the story fragment you will get yourself, because you will bring a unique navigational history.
The narratives are designed to be often interwoven, with a primary connector being place. Another is reader desire.
This will be huge in scale, with dozens of primary stories, three story levels and a large number of variations. The scale will be open in allowing users at a certain level of discovery to write new stories or branch existing story lines.
The experimental challenge in this dimension is managing coherence and ‘unputdownability’ as the compositional elements of candidate narratives change. We experiment with two management techniques to handle engaging coherence:
- one of these is the governance afforded by requiring most (but not all) of the story navigation and discovery to be by physical visit. A challenge is to set a small number (possibly three) balance points between difficulty and pacing of narrative discovery with the reward. The three balance points might span from one end: easy with some reward, and more challenging with more reward at the other end in terms of acquiring place (and possibly some deduction). (‘Narrative reward’ is likely to always provide some key information desired from the history so far, but to incompletely fill out the newly governed situation in tantalizing way.) Thus, while extreme fantasy, magical worlds and metaphor are hosted, anchors in the physical place with associated discoverable history moderate confusion.
- the second technique to mitigate incoherence is to host mature infrastructure from our other work in the virtual enterprise self-organization and the recent focus on the ‘narratives’ that organize bodily systems. The extra mechanics are used to assign and support agency in the narrative soup. Some of these agents are explicit in the authoring, and some of these explicit in the story, like characters and key objects. Others are inferred or implicit and still others are invisible. These latter types require close integration of story and dynamic and why we require that main story lines are professionally written in house.
These two mitigating strategies when combined allow coding in functional reactive networks that manage governance. From the storytelling perspective, the effect is that as new pieces of information are added, what you thought you understood about older pieces of information and how they are connected might change. The change ‘makes sense’ and in most cases more sense than what you had before, but might open new questions.
This is a key element in the research: to discover what the tipping points are between complexity that adds confusion and greater scope that adds engaging (re)clarity.
On Space and Form
The importance of place is an essential part of the experience.
- place has history. Insofar as we have it, that real world history is available for the fictional narrative experience and for non-fictional historical narratives that (ideally) employ the unputdownable dynamics.
- place has effort. Some story nodes will be just happened upon by readers, but most require an investment. The app can optionally (and anonymously) calculate the cost in getting there, assuming that the reason for travel is narrative continuation, and that the cost is distance-related (as derived from cloud map service APIs). I think we will allow reader-users to arrange for surrogate visitors.
- place can have social context in encountering people. This is addressed elsewhere, as the social component of the system is not the primary focus and is expected to be handled by collaborators.
- place can have shape and form. That is discussed here.
Features of a place matter in normal experience, physical features like open space plus the form of enclosing or engaged shapes. These clearly have some psychological effect on visitors to a space and are available to be leveraged for the narrative experience. We especially consider psychological characteristics of the enclosing space if one is indoors in an interesting space. Some architects call it form, as opposed to style and simple features (like high ceilings). An intense study some years ago into the effects of different enclosing forms with a dozen of so significant built spaces has results that we can test and extend.
The science on this so far is poor and depends on the situation theoretic tools we are cooking up. But a good intuitive writer can evoke, capture or exploit this physical form. Some good architects do this, creating a ‘hypernarrative’ by the succession of forms encountered as someone, for example moves through a house carrying complementary narratives of self and purpose.
We have four interests in this general area.
- Augmented Reality. A collaborator handles this. Surely there are a number of fascinating dynamics to study, but exploring it is not a core priority for us.
- Space as a Character. At least one of the ‘main thread’ stories has the city embodied as a character; simultaneously key features of the city have identity as characters. This loosely follows the lead of Finnegans Wake and how it handles Dublin. But the anchors are more in how the physical fabric emerged than in Joyce’s word play. This emergence is rooted in the Maps Project, but extended through dynamic models of the space as it has ‘matured’ or is growing to some state following urges. The connection with architectural and historical studies is obvious, as is the play with augmented reality. The key quality we exploit is the potential role of place as (at least pantheist) agent in some stories, including the history of how a city and space ‘want’ to evolve.
- Kutachi. There are two primary syntactic elements in these narratives: words and places. (Cinema is covered immediately below). We have a wonderful set of tools to work with words. Some of these employ extended TEI (Text Encoding Initiative) markup which we use. There is not an equivalent for form, though we have some promising foundations for feature annotation from the CAD world. A core effort in the project is to develop a katachi/kutachi-informed TEI equivalent.
- redframer. This is our project to crowdsource deep narrative dynamics as discovered in movies. A tight connection between the projects is at the level of dynamics, governance and engagement. But we can have a wholly different connection. Boston (as our launch city) has interesting histories, but locations also have presence in fictional films, and there is no reason these ’stories’ cannot be in the toolkit to be integrated.
On Text as Agent
There are two related ideas here.
The conventional notion of hypertext has relatively small, static packages of information. The dynamism in the system is how, whether and in what order the reader discovers them. An essential effect is that they accrue over time to produce the effect of long form narrative, different for each experience.
The contract here is that the authors create small experiences that are creatively wired up. But what if the process of navigation changed the fundamental nature of the ‘text’ at a node based on what was experienced before? That is, in conventional hypertext, everyone brings a different past experience, but the node they arrive to has (roughly) the same artistic artifact. What if that artifact itself was composed uniquely for each history? We accomplish this by two techniques.
One is simpler: allow the artifacts some agency in moving around the space. In other words, instead of gluing each text artifact to a node to be discovered, make each text artifact hungry to find someone to experience it.
The Links Have Agency
The second technique comes from functional reactive programming. Some background is in order.
A conventional approach to hypertext tools, indeed anything broadly in this domain, is to envision an artifact and associated experience and then build it with the appropriate tools. A creator, for instance will have something in mind and then go to programmers to make it; perhaps he/she will build it using some programming tools divorced from the creative process, usually constraining it. Or they will go to prefab tools created by programmers, like Word or Storyspace.
But creative programmers often think of themselves as storytellers, and the program not as an engineered product but as art itself. Set that aside for a moment.
One programming paradigm uses transformative functions that are persistently chained in a network and the way the program works is by changing the connections or links in this network. Think of the network as a collection of nodes and connections among them. The connections rewrite what goes through them. If something way back in the network changes, then every channel rewrites what went through it and all text at every node is liable to change, perhaps radically.
As the user of the program ‘reads’ a result at the end of a fabric of channels, he/she perturbs the channels so that the result adapts not only at that node, but all previous and future instances of it and what it will feed. In programming terms, this means that every information channel adapts. In hypermedia terms, it means that everyone interacting with the text in all the places, modify text not only there but in subtle ways in many other places on an reader by reader basis. An existing collection of formal theories, such as channel theory, come to bear here.
Our experience with reactive fabrics for distributed assembly of story is useful in this case.
On Adding Locations and Characters
Now focus a bit on the narratives themselves. These will have characters and be discovered through visits to places/spaces. A logical speculation is what would such a system look like if human actors (in the real world at node locations) were involved in the experience. There are some theatrical experiments along these lines, but we think that would significantly constrain the experience, so the notion of a ‘fourth presence’ is not considered feasible in this context, though some collaborator may experiment with us. The experiment would have to be constrained by our necessity of global coherence. Without this, governing forces cannot be modeled.
(The first presence is the reader who in some instances may appear as a character; the second presence are the written characters in the narrative, some of whom acknowledge you and ‘follow’ you, perhaps in backstories or parallel but temporarily hidden stories. The third presence is the personality of the authoring system that combines the composition software and its authors with the professional writers and the voice they present.)
But that still allows for the direct intercession of the visitor in the story and the unexpected appearance of place-centric characters who may directly address the reader. And it still allows for strongly anticipated inter-personal experiences among visitors to the places, each of whom will be presented with a different narrative and who (in relatively rare cases, be expect) might share with others what they have seen. When this occurs, the inherited link history of those visitors can be expected to be imposed on your history in some way.
There is no reason that this cannot involve cosplay from inherited, external narrative. For example it is desirable from a project perspective to have a narrative cluster that (by some means) is integrated but separate and limited in time. For example, a fan conference where stories can be set in the host city.
And of course there is the ‘regular way’ of introducing (and terminating) characters, by writing them in by authorized contributors.
Adding Locations and an Open API
Creating new nodes by adding physical locations and associated story branches should be just as easy, including temporary and frangible nodes. The latter may be nodes that can be broken or altered by visitors, deliberately or by concerted action. This is a good place to restate the priorities of the project.
- the main goal for the sponsors/creators is to explore new annotative and presentation modes, using space, place and narrative construction and purpose.
- as engagement is essential, there are crowdsourcing and scaling experiments. These are used to determine the effectiveness of different modeling, construction and interaction models, but are an end only in the context of coherent presentation of material.
These two require some discipline in focus and underlying infrastructure. At the same time, it is such a promising, exciting environment that others will want to join, bend and subvert it. Some useful collaborative research projects are indicated here; there are likely to be others.
At the same time, there is no reason not to publish an open API so that creatives and startups can play with the general notion and content without the restrictive focus of a purposeful research project.
A goal of the project is to devise and stress annotative and enrichment techniques for the digital humanities, while moving the focus from library science to also support creative engagement. Traditionally this has been done by annotative and indexing technologies, for example TEI, was developed to annotate existing text documents. TEI cooperates well with infrastructure developed to support XML, and it makes sense to think about:
- what is the ‘markup’ for meaning, intent, effect within the confines of next generation digital humanities?
- what is the equivalent for spatial and cinematic material, knowing that existing standards and methods of geolocation have to be subsumed?
- what do we do to include modeling of engagement and effect?
There’s no question that this has to be approached at a fundamental, formal level.
We rely on four novelties to make this work. All of them can be characterized as ontological.
We have been through enough projects to mistrust the approach of the so-called Semantic Web, though this is the default for many projects. That approach was settled before more careful formal tools arrived. It is document-oriented where we want to be process and transformation-oriented. It is geared toward tenable specialized domains and simple reasoning where we want something more general and capable of handling complex reasoning. (We have a strategy to deal with non-tenable reasoning cases.)
Therefore, we follow the approach taken by the biomedical research establishment in building a modern ontology core in the Basic Formal Ontology project. While this was incubated to share biomedical information, and the design is not particularly friendly to what we do, it is well thought out and worth investing in. We’ll add a situation occurrence type that can instance as an understanding of the story so far.
Quantum Interaction Categories
We want to reason about narrative using emotional intelligence and culturally shaped worlds. Logic is not well suited to this; a well known problem. We implement a reasoning system based on advances in so-called geometric logics devised to better handle quantum mechanics but which have been applied to other domains. This allows us to reason about where someone has been, how they have interacted with the story and what to next present and where. It absolutely requires the sort of ontological formality noted above and linked to text using QI tags extended from TEI.
This is the theory that logicians devised to handle the problem of context affecting the meaning of information. We consider a persons’ makeup as a situation; a story-so-far as a situation and a place as one as well.
Everything on the interpretation side depends on the creative contract between author and reader. This includes what is normally considered library science and search strategies. You cannot just ‘tag’ elements of a story without considering this use. So we have to model how things hang together and why. It is the only way we can (mostly) guarantee that the effectively infinite variety of stories folks will cocreate will be coherent and unputdownable.
All four of these have ontological implications and from a project perspective most of the work will appear to be work in that domain.
The system already has some narrative folding built in. There are
- the narratives each visitor brings to the experience are a layer ‘on top’ of the story; these will be deliberately engaged.
- we have the ‘real world’ narratives involved with the physical transit among places.
- because we involve larger scale environments, transit events that happen in the real world come to bear.
- and of course we have the presented narratives.
Regarding those narratives, this project supports more explicit ‘folding’ than usual in the written word, but common in movies, especially movies appreciated by young viewers. Folding is the imposition of several stories one on the other with a focus on how the reader is involved in the experience. An example is detective fiction where the detective exists in a folded layer between the world of the murder and the world of the reader. That detective is partly in our world as we navigate into an understanding of the narrative’s causal dynamics, and is partly in the world of the presented events, directly interacting with characters who are trying to ‘write out’ his discovery. There are many, many such folding dynamics that are emerging and evolving in film. It is one of the goals of the project to discover and better model them such that contributing writers can imply them for effect.
The effects we are looking for are high unputdownability and maintained coherence in opportunistically fabricated narrative.
Authors will include reader/writers motivated to extend the story. We provide for these writers, emphasizing any contribution they make at a node, even if it a revisit for the purpose of registering some writing. That is, it is conceivable that a reader will come to a node, have an experience, go away and formulate what they might contribute and even write it, then come back to the node in a day or two. When they come back, they should be able to reset the experience to their previous visit to enter what they have brought, assuming that the experience may have changed in the interim.
It is also conceivable and welcome that these writers will introduce new places into the manifold story so that the narrative possibilities grow in way related to (but not exactly like the existing concept of) viral growth. It is also conceivable — and will be encouraged — to have these contributions include anchors to places that the public cannot visit, for example a reader/writer’s home or (in the case of immigrants) their native places. In this case, readers should be able to virtually visit those places and the narrative nodes associated with them.
Places in the system are simply GPS locators of public, safe locations so can be added promiscuously.
The example story structures tumbled into my mind during the hypertext conference because most of them are things we thought up in the redframer context. In that context, the location element allowed reader/writers to introduce places in films that may or may not align with physical, visitable places.
The story fabric has three layers. A device has one character (Johnny/Jack/Jake) as a primary character in all layers, based on Jack from Jack and the Beanstalk.
The overarching story is Book of Giants. Giants are literally gods, inhabitants and directors of dreams. Some are embodied in place in the way Dublin has life in Finnegans Wake. That is, a giant can be incarnated as a place or geographical feature. The giants are known to each other and their stories overlap. There are many such characters, often collaborating and themselves jostling for influence not over each other but the urges they collectively influence. But all giants collapse at some point into two characters, Big Jake and his mysterious sexual partner.
He is master of the three fundamental urges that he invokes by chanting their names (Fear Fire Fore) and commanding them to ‘Form.’ They then weave a linear braid. Ordinary humans would see this braid as an endless woven vine; in the generative story it is a beanstalk from the three urge seeds. Jake is obsessed with maintaining order in the world and much of this is trying to indirectly govern his younger selves Jack and Johnny as they compete with each other to control the future Jake exists in.
Jake can only do this by changing the governing laws of the world they live in; we will see this as changing the rules of the narratives. The nature of the braid of urges — its form — determines this governance.
So this level of the worlds of stories goofs with the stories in the other worlds and is our explicit in-the-story opportunity to put the narrative managers as they manipulate the other two levels.
This level is written by professionals and is intended to be endless. No one ‘reader’ will be able to exhaust it by design; and most readers will encounter a different set of stories. The reader must visit a location to get more of the story. Some versions from different readers and sequences will conflict.
One story at this level involves Jack, a gifted magician who hates his gift. He ends up performing magic in all sorts of situations, including tussles with his younger self and characters (unbeknownst to him) influenced by his older self. A recurring storyline is his trying to trick Johnny (his younger pre-adolescent self) into not taking the magical items Jack gave him. Jack’s internal urges are hard to control, and he does his best magic just after sex. (The possibility exists that his apparently long suffering wife is the true mage.) Jack’s magical powers are a result of his older self encountering a younger self as a child and giving him three magical beans, one for each of the three primary urges. Jack is unhappy in his life, even suicidal, and is on a continuing quest to re-encounter his young self and undo this transfer. He repeatedly fails, each time inadvertently increasing the power given to the younger self.
The Jake stories all overlap, but the stories at this Jack level usually do not, as each are world-defining stories. The public can write stories at this level, but the effort is high, because you have to define the world as well as convey the story. Mature worlds can be inherited, though, so a Star Wars, Marvel Cinematic Universe, Wizarding World and so on can be instanced by writing a story.
The extra effort here is in setting the world ontology, done by describing the differences from an existing world, so as a practical matter the authors here are at least initially coordinated through project management. They are hypertext only and episodic and with more detail. That is, the stories grow by adding more characters, episodes for each character and more detail to each episode.
What differentiates the Jack story is its spatial nature and linking to the maps.
This level is an open world — many open worlds, each based on some foundation story. Users can create and extend these worlds readily. For now, we leave the suggestions for this level undefined. We expect that a few epic series may emerge.
An Example Focus Narrative
The project as described above is an omnibus framework for research in the digital humanities. While relatively new, that field has a tradition of acknowledging differentiated groups and including their needs. Already, there is an institutional definition of who these groups are and what issues are salient. The GiglaMesh project attempts to subsume many of these issues by allowing group-specific branching and narrative form.
Additionally, two experiments are a special focus.
Ethnomathematical Concept Abstraction
Usually, when ‘diversity’ is celebrated, it is a sort of inclusive tolerance or a celebration of cultural variety. We also find it intrinsically valuable for a different reason, additionally holding that the best experiences emerge from a staggering variety of perspectives. We go rather much deeper in this direction than is customary.
Ethnomathematics assumes that much of mathematics is not discovered but invented, and different cultures have invented novel abstractions to serve similar needs. As with all such abstractions, different approaches have different strengths and weaknesses. Having creative team members with different coherent abstraction systems allows a team to surround a problem. Western mathematics covers most of the world, but we still find strong pockets of other abstract reasoning systems. Many of these are from stressed parts of the world which are producing masses of refugees.
It is possible that in the US we have a chance to take advantage of these diverse strengths before globalized education all but eliminates them. It could be the last era where we can preserve these by recognizing, honoring and utilizing them.
Situated Refugee Sense of Place
The United States has always been a favored home for refugees and other immigrants. It is in our fabric, but many of these people come from areas where their people have resided for centuries. The sense of self, tied to cultural identification is necessarily rooted in places. When they arrive here, this is all gone. We do understand many issues associated with relocation and transition, but we have scant insight into how this sense of place and loss of it affects basic abilities. Yet we have about 45 million people in the US that are born outside the US. Adding their children doubles that number, nearly all concentrated in major cities. In fact, nine US cities are each home to over a million foreign born souls. We do know that this loss of place disproportionally affects women. The over a million LGBTIQ of these are a group of special interest, as are the larger class of people who were oppressed minorities in their home country.
There is no deeper or more accessible path to understanding cultural dynamics than through the engagement with narrative.
This subproject of GiglaMesh focuses on the needs of these immigrants by both providing a way to understand the role place can play and providing a vehicle to remediate loss.
We may call it the Walkabout Project, honoring the 40,000 year old Australian Aboriginal culture and their concept of solitary connection with the land by opportunistic meandering. We will connect with scholars that work in this area and work with (primarily NGO) organizations in world cities.