Narrative
in
Multi User Learning Environments
- Exploring a Perspective -
by
Stefan Berreth
EMMA thesis IMM 1997/98
- European Media MA in Interactive Multimedia -
Hogeschool voor de Kunsten Utrecht
Utrecht, The Netherlands
Faculty for Art, Media and Technology
Europe, 18.8.98
1 Prologue.................................................. 2
1.1 Personal Motivation.................................... 2
1.2 Acknowledgements....................................... 2
2 Introduction.............................................. 2
2.1 Objective.............................................. 2
2.2 Problem Outline........................................ 2
2.3 What this work will not provide........................ 2
3 Narrative, Teaching, Cybertext and Drama.................. 2
3.1 What is narrative?..................................... 2
3.2 Narrative and Teaching................................. 2
3.3 Cybertext.............................................. 2
3.4 Drama in Computer Human Interaction.................... 2
4 Learning models........................................... 2
4.1 Group Based Learning Forms............................. 2
4.2 Knowledge Building..................................... 2
5 Suggested Design Pattern.................................. 2
5.1 Phases in Educational Models........................... 2
5.2 Assumed Setup.......................................... 2
5.3 Phases of the Design Pattern........................... 2
6 Suggested Design Pattern and the Examined Theories........ 2
6.1 Pedagogical Content Knowledge.......................... 2
6.2 Cybertext.............................................. 2
6.3 Drama.................................................. 2
7 Epilogue.................................................. 2
8 Glossary.................................................. 2
9 References................................................ 2
1 Prologue
1.1 Personal Motivation
My personal motivation for writing this thesis was a personal feeling of only finding unsatisfactory concepts and patterns for designing network based multi user learning environments available around.
The need for learning applications, whether based on multi user communication or not, is huge nowadays as I have personally experienced in several occasions in the recent years.
So for example when I was working at the Institute for Didactics of Physics and Teacher Training (IFPL) at the Technical University Berlin on Interactive Screen Experiments the potential and first raw lines of what principles such a multi user learning environment would have to adhere was gaining shape. A subsequent incident involved me in the conceptualization of an offer for building an educational multimedia application for the employees of the Volkswagen AG world wide for teaching in house procedures of quality assurance and the participation therein. This experience made clear that such a challenge could not be met by simply enlarging the existing and often unsuccessful concepts of single user learning applications to support parallel sessions and then label it multi user learning.
The final igniting spark for writing this thesis came with my attendance at a presentation of the plans for the The Network University right after we had finished a prototype for a shared workspace web application for the Society for Old and New Media in Amsterdam called PublicSky during the MA year at KMT HKU Hilversum. It appeared clear to me that the mere availability of technology, financial and human resources and the motivation to set up a distant learning class based on the web and access to academic knowledge and resources wasnt enough to build a flexible, scalable, robust and universal learning environment for a class thats members at large wouldnt have real world contact with another during the course or ultimately during their whole life. What is needed is a sufficiently constraining and at the same time generic design pattern that supports the core processes in successful multi user learning environments.
Attempting to sketch and motivate such a pattern within an MA level thesis within the very limited time at hand in this particular course turned out to become a StarTreckish challenge To boldly go where no one had gone before with the additional spicing of having no time and forum where initial ideas and concepts would have been debated and proven to be not missing the point entirely. This especially lacked when formulating ideas that were reading existing concepts in uncommon, very general ways to in turn derive a valid and suitable formulation for the goal in question. Such a forum could have been an environment to detect and correct mistakes and faulty assumptions quite easily along the way. However, such a forum was not available and thus those faults are still in there.
Thus, taking this work as it is, an untested attempt to merge existing approaches from literature theory, computer human interaction and pedagogy into a design pattern for network based multi user learning environments, it is an invitation to both, a discussion and refinement of the concepts suggested here as well as the one of an actual thorough implementation along the lines of the design pattern sketched here.
Stefan Berreth - sb@well.com
Hilversum 18.8.98
1.2 Acknowledgements
I thank Louis and Liane for being much more than just patient.
2 Introduction
2.1 Objective
The objective of this thesis is to analysis the nature of narrative found in teaching situations and explore how these concepts can merge with modern understandings of narrative in new media and drama in human computer interaction. Eventually a design pattern for crafting network based multi user educational environments will be sketched and in turn related to the theories it was initially derived from.
2.2 Problem Outline
New medias most interesting feature seems to be trivial
at first glance but is crucial to the understanding of what has developed so
far in this field: There are options and potentials in these media for how
communication can take place in our culture that are - you name it -
"new" to our culture. The structures that potentially can emerge in
the communication between humans as individuals or groups and between humans
and computers in new media are unparalleled in older media and thus in our
experience as a communicating humankind. This entails that the understanding of
how to design new media applications that make use of the richness of
potentials in expression, content, behavior and communication structure and
- -dynamics is still very young and clearly underdeveloped compared to the one we
have in dealing with traditional media our culture is used to already for
some longer period of time. This applies as well to consumers as to generators
of media products.
Interestingly what happens at first when a new medium emerges is the process of
remaking traditional medial forms within the new medium. As well as in the
first years of motion picture, when the camera mainly was placed somewhere
(like a traditional spectator in the audience) to look at the play on the
stage, nowadays most of the things that are made in new media is hardly
extending the mental framework found in traditional media like print, TV,
letters or telephony. Not before the makers of film started to understand the
richness and the potential for expressions that are at the core of working with
a camera they were able to use it to generate convincing works that we are used
to today - working with essential techniques such as cuts, camera movement,
field of view and sound, yielding an self-sufficient way of how to communicate
through film. In parallel what happens within new media nowadays - as well on
the side of the makers as for the recipients - is by large a remake of what is
known from the old media already. Not that this was an unnecessary part in the
whole process of a culture exploring an emerging new medium. But not until one
has gained understanding and experiences of what lies at the expressive core of
a new medium and what it means for the potential to communicate through it one
will be in state to develop the language that will naturally be inherent to
timeless works in the new medium - the mediums own form.
The process of developing such a new form in turn is always self-referring
iterated process of (re-)invention and (re-)evaluation of attempts to make this
new form explicit in a work. Not that this process ever actually comes to a
halt in any living medium whatsoever - but the initial exploration certainly
the most interesting and challenging part of the lifecycle of a medium - at
least for the designer and researcher in this field. It is this phase of the
development that is happening right now in this field.
In a special area of new media, the field of educational applications, becomes an increasingly important issue in the emerging knowledge- and information society.
The focus in this work is narrowed to systems that are intended to serve as multi user learning environments.
A common problem that arises when traditional communication channels are transposed to new media and digital applications is the fact that the structures that emerge in multi user communication are difficult to predict. In this context multi user learning environments are especially interesting to study concerning the potential communication structures: In a community formed by multiple users that share a common goal a self organization concerning the structure of occurring communication emerges that is located in the triangle between personal one-to-one, hierarchical broadcasting and scattered many-to-many communication.
As long as the potential of possible communicative multi
user structures are constrained to match the traditional structures like a one
to one situation on a telephone (like personal email) or the one-to-many
situation in broadcasting and newspapers (like standard websites), the action
and reaction patterns between the communication instances are familiar to
designers. This is why up till now most often new media products are designed
that follow these traditional patterns.
However this restricted approach cannot be applied when the communication
structures become more open in such a way that a great deal of channels and mechanisms
are provided at once that allow for the formation of communities within the
tool or medium itself.
In these cases the design of such a framework allows much
less of the traditional notion of controlling the flow of information, the flow
of action or the flow of experience. This doesnt mean there is nothing left to
design anymore, though. The design of tools and media for open, complex, in
principle unpredictable systems for and consisting of communicating agents
(humans and computers) doesnt mean that just everything will occur
arbitrarily. However the methods for the design of such systems, the mental
framework one has to employ in order to understand what one is doing when
designing such a system, are different from these of a traditional author,
engineer or designer. Its rather like developing the architecture of a house
or a city or like writing the constitution for a people in a country or setting
up rules for whatever community you might have in mind. As a designer then one
is not in charge of the end product anymore, but rather becomes the architect
of a system which is not completed unless it is used by a often open and
dynamic - group of people that form a community within constraints. The process
of being used constantly reestablish the completeness of the system itself
and during the lifecycle of a system the constraints (or rules) themselves even
might become altered by the users themselves.
In order to find design patterns for such systems - and herein this is work
will be focusing on learning environments again - one has to develop models
that lead to an understanding of the principles by which human-human
interaction and computer-human interaction actually takes place in learning
communities and modern medial forms exhibiting cybertextual structures. From
these insights one has to derive methods for building multi-user frameworks
such that they are successful in the sense that they actually let grow
communities from the group of people using them while serving the overall goal
of high quality learning of a given subject matter.
One approach that can be taken here is taking the perspective of narrative
theory, ergodic literature and drama in computer human interaction and to
examine how it works out when they are merged.
2.3 What this work will not provide
This work will not provide a sharp and comprehensive discourse in terms of the own measures of either fields touched here like pedagogy, narrative theory or drama. It will merely take core concepts from these fields that suite this sketch of ideas and concepts of a design pattern for building successful multi user learning environments.
So die-hards and geeks from each of these fields will most
certainly miss rigidness and comprehensiveness in treatment of their home
field. Avoiding this problem would have meant not to write this thesis at all,
which was not an option.
3 Narrative, Teaching, Cybertext and Drama
Introduction
In this chapter, first the notion of narrative will be briefly explored.
Then, based on the works of Sigrun Gudmundstrittor [Gud], light will be shed on the role of narrative in teaching whereas teaching should be understood as being communicating knowledge in general and not just the loaded notion of a teachers uttering information to be memorized by passively receiving students.
Following up, the common notion of narrative will be broadened to what Espen J. Aarseth [Aar] in his work describes as cybertext by outlining his framework that provides strong and appropriate tools for the analysis of the textual structures recently becoming popular again in the new digital media that are not comprehensible with the classical notion of narrative in literature theory.
Then the essentials of the ideas of Brenda Laurel [Lau]
will be lined out who has successfully laid the relation between drama and user
task oriented computer human interaction. This will be extended and criticized
where necessary to let them apply to multi user applications and particularly
multi user learning environments - later on.
3.1 What is narrative?
Lets first have a look at several aspects the term narrative carries and how it is understood within this work.
General
Narrative in general refers to the structure, knowledge and skill required to tell a story. A story in turn characteristically has a beginning a middle and an ending and is the expression of a succession of events and their relation among each other. The relational structures amongst the events and settings are usually called plots.
Dictionary definition
When looking up the term narrative in a referential source we find, for example in Oxfords Dictionary, the following definition for narrative:
"A spoken or written account of connected events in order of happening."
This definition alone is unfortunately an unsatisfactory one as it obviously doesnt comprise a large amount of works that are commonly regarded as narrative too - works where the order in which events are revealed to the reader isnt the order of the happening of those events. Furthermore it emphasizes the matter whether it is spoken or written, which is obviously an inappropriate restriction to the notion of the term narrative. Narratives that both are neither spoken nor written and/or reveal the events in a different order than they occurred are far from being uncommon. A popular and recent example here is Quentin Tarantinos movie Pulp Fiction (1994) where the story is being told by being acted out. One of the key features of the film is a weird non-chronological order of scenes of what actually happened. And it clearly is a narrative.
Telling vs. Story
Generally, most people agree on the intuitive and vague definition of narrative being telling a story. By this approach ultimately the problem is split into two distinct parts: the way of telling - how - and the message told - the story.
This notion lies within the lines of how structural theorists like to see the term narrative to be understood. Mainly they regard the narrative as being the end product of a story that experiences a presentation, i.e. that is told in whatever form. What comprises the events, characters, settings and their relations is regarded as being the story, whereas the discourse is the way the story is expressed, i.e. whether its written down, spoken, drawn or painted, or acted out in drama, motion picture, mime or dance (Chatman [Cha] 1978 in [Gud]).
Other schools, like the Russian Formalists, have laid the distinction along lines comparable to those of the structuralists as seeing story as half fable or fabula, which also contains the order in which events actually happen, and the sjuzet, which is the order in which these events are revealed to an audience (see Phillips [Phi]).
In this sense not just intuitively - a story that gets told becomes a narrative.
Narrative and values
Another aspect or narrative is that the presentation of events (the telling) always occurs in a way that cannot avoid to assign values to the events and their relations, i.e. to assign moral. The values that are assigned to the entities within a story are either expressed explicitly or implicitly. This assignment sometimes simply occurs through the circumstance that the story telling person regards an event as worth being told or not. To quote Gudmundsdottir [Gud]
Values and narratives are inexorably intertwined. Together they have one fundamental principle in common, a principle that is basic to the narrative nature of pedagogical content knowledge. This basic principle is that narratives help us interpret the world. Values and narratives are interpretative tools that constitute a practical, but also highly selective, perspective with which we look at the world around us. (italics by me)
The next section will examine the role of narrative in
teaching situations.
3.2 Narrative and Teaching
An area where narrative very naturally plays an important role is in the communication of knowledge.
Pedagogical content knowledge
Sigrun Gudmundsdottir [Gud] elaborates this point as it can be observed that teachers rearrange their knowledge about the world and the subject matters they teach into a narrative form referred to as pedagogical content knowledge[1]. By pedagogical content knowledge is meant the special way in which teachers own knowledge about a subjects matter is arranged, memorized and prepared to be told. This transformation is, following the results of the research in this field found in Gudmundsdottirs work [Gud], a very natural thing to human culture in general and thus applies not only to teachers. However, teachers exhibit this structural feature of communication very strongly, as their profession is to teach i.e. to communicate explanations, and thus they lend themselves to being subject to systematical observations.
By transforming the knowledge into pedagogical content knowledge, facts get organized as events with a structural relation and assigned value. They are arranged into a form that enables to tell someone else in a coherent way that something happened or how something is. This process of course is not limited to any particular pedagogical concept as one might probably think at first glance. It is, as pointed out already, rather a universal and natural way for humans to prepare knowledge that is intended to be communicated to someone else.
Teller listener relation
The teller listener relation for that matter does of course not have to lie within the lines of the old fashioned notion of a teacher delivering information to passively listening students.
The process of telling pedagogical content knowledge of course occurs equally among students themselves. Then, however, the distribution of the roles between teller and listener are not static anymore but rather change dynamically depending on how the communication is organized or self organizing.
Students working in a group for example, have to communicate their ideas and already assembled knowledge about the subject matter to each other. When they work with media to research on a given topic like texts books, multimedia, films or audio tapes, the teller-listener relation is found between the author of the used material and the student, who in turn has to prepare the insights gained in a suitable way to communicate what she has understood (and what not) to her fellow students. This preparing in a suitable way is the transformation of the knowledge gained into pedagogical content knowledge that is narratable.
Incomplete stories
Especially when someones knowledge about an issue is significant incomplete one function of transforming this knowledge into a narrative form is to complete the otherwise incomplete picture. Gudmundsdottir [Gud] remarks in this context that
in most cases, transformation involves progressing from a incomplete story to one that is more compelling. We do this by establishing first the connectedness or coherence that moves the storyline along through time. Next comes the direction of the story, the goal or point of it all. With the goal established, events are selected, rejected, or transformed and take on a significance that they would not have other wise possessed. [] By using narrative form we assign meaning to events and invest them with coherence, integrity, fullness and closure.
It is during a conscious act of completing the picture where to most people it becomes conscious that transferring knowledge involves the generation of a pedagogical content knowledge of the issue. When, for example, some knowledge wants to be told to a person that for some reason doesnt share the same context with the telling person, this context has to be provided explicitly inside the story. Doing so demands the explicit telling of the relevant context to provide the appropriate background in front of which the idea to be communicated first gains the intended meaning.
It has to be mentioned however that to achieve these completions of the story, it happens more often than never that the pedagogical content knowledge is supplemented with derived or assumed ideas, i.e. element that are just plausible, but not necessarily founded on the actual facts[2]. Of course this reflects various presumptions about the world the telling person weaves into the story by this act.
More consciously this regularly happens in journalism where mental shortcuts are given to the receiver of a story that in rigid terms of the subjects matter are incorrect but provide the desired coherence of the story in the limited time or space of a journalistic contribution and thus enable better comprehension on the side of the receiver, and in a sense communicates better.
Universality
The notion of arranging pedagogical content knowledge in a narrative form as a fundamental structural feature of human communication has lead thus far that for example McEwan and Bull [MEB] (1991 in [Gud]) argue that
there cannot be a content knowledge without a pedagogic dimension because the understanding and communication of an idea is itself a pedagogic act. [] To understand a new idea is not merely to add to the existing stock; it is also to grasp hold of its heuristic power its power to teach. Explanations are not only of something. They are also always for someone.
This is the key insight that almost naturally points to
the role of narrative in modern pedagogical concepts such as knowledge building
or project based learning. It puts narrative in the role of something not to be
overcome as an obtrusive instrument of a old fashioned teacher role but rather
gets transformed into something different, a sort of narrative that now
consciously arranged and stimulated occurs in the learning process between
the students first hand experiences and the students communication about them
among each other, to the teacher or perhaps to people outside the course or
themselves at a later time.
3.3 Cybertext
Problems of traditional narrative theory
The following complaints have already a settled tradition among critiques: The great beauty and power of narrative is lost in the upcoming new media, children dont know how to read a book from the beginning to the end anymore, and alike is heard.
The attempts to find back true narrative structures in new media, especially in media that have a hypertextual structure at their basis, come to a sudden end when nonlinearity enters the stage. The author suddenly seems to loose control, plot, story and characters cannot be developed anymore the way it used to be or not at all; narrative vanishes and with it its great power to tell stories.
The way out
This chain of conclusion is by far shortcoming, as Espen J. Aarseth develops in his book Cybertext Perspectives on Ergodic Literature [Aar]. The impotence of literature theorists and -critics of the past to get grip on the emerging new forms of experiences and communication that is mediated through computer based media is rooted in the narrow notion of what constitutes literature at all. The notion of ergodic literature, that is literature that demands more than trivial action from the reader to reveal its textual body, is developed by Aarseth and a long history of ergodic texts is put in relation to the more recent emergence of computer based hypertexts and - -media. Without going into detail about the whys and whats of the historical development inside literature theory concerning this phenomenon (which is excellently treated in Aarseths work [Aar]) I will recapitulate the principal concepts of cybertext that are of concern here.
Replacing traditional plot
The obvious lack of coherent plot in hypermedia, the presumed out of control situation of the author as the one who should be in charge of the developments within the story, actually occurs, but it is replaced by other mechanisms that are more fundamental and universal than the classical narrative structure and its not a new phenomenon in new media, as ergodic texts have existed at least as long as linear writing[3].
Communication models
In the classical narrative communication model, the communication flows from the author via the narrator via the narratee to the reader who in turn identifies with the narratee to get grip on the story. In ergodic texts on the contrary the coherence between the fragments of text is disrupted at the points of communication between author and narrator as well as between narratee and reader. Within the fragments of text the reader encounters, the narrator communicates perfectly well with the narratee. However the author doesnt have full control over the situation anymore as to when and how this communication takes actually place. Complementary the reader repeatedly gets disrupted in her efforts to get a notion of the underlying whole, the story. The identification of the reader with the narratee gets sabotaged by some unidentified instance that is hidden in the intrinsic nature of the structure of the hypermedium itself.
Something interesting happens at this point from the part of the reader to regain some coherence. She starts a strategic counter attack (Aarseth) by becoming a metareader, reading her own way of reading, to escape the limited role as a reader who cannot identify with the narratee anyhow reliably anymore. It is, as Aarseth puts it, an effort to regain a sense of readership.
The Aporia Epiphany Pair
The active involvement of the reader to come to an understanding what is inside the textual body by taking the metareader position to actively puzzle and make sense, i.e. struggle with the aporia to reach epiphany and consciously observing and reevaluating his own strategy in doing so, forms the core of the attribute ergodic in this context. Actually, cybertexts can be understood as comprising as well the two subelements of traditional representation description and narration[4] - - as well as the additional element of ergodics [Aar 94-95]. In this picture narrative is still a representation of the succession of events with their structural relation between these and as such every experience in a hypermedium can be communicated in the form of a traditional narrative afterwards, but the very happening of experiencing and exploring the textual body (the act of reading) of a cybertext is something rather related to a game where the user has to participate nontrivially in order to make not necessarily all events happen to her. In a game there also occurs a succession of events with a strong relation amongst each other through the active engagement of one or several instances (like humans or machines) but the game in itself it is not a narrative however it can be narrated in various ways [Aar p.94].
The latter is vividly demonstrated in any radio life program featuring a soccer match for example.
At the core of ergodics lies thus the activity of the reader (or player or participant) of a cybertext to find the link out from an incomplete set of textual fragments. On a very structural level the cybertext prohibits the reader to get behind the complete picture instantly, as it presents the fragments of text (or another medium) in a sequence that by large is dependent on the game between the choices of the reader and the implemented inherent structure of the textual body through the author.
This holding back of fragments that are needed for the reader to get a to him satisfying impression of the issue communicated, is the aporia. It is very natural to the basic structure of cybertext, as it only could be prevented by delivering the text in a codex literature equivalent alternatively, i.e. at once. The very moments in which this aporia resolves we encounter epiphany, which corresponds in a certain sense to the notion of closure in traditional narrative. However, Aarseth points out, as there can also be identified an aporia in traditional literature the kind of aporias in codex literature and cybertext differ:
In contrast to the aporias experienced in
codex literature, where we are not able to make sense of a particular part even
though we have access to the whole text, the hypertext aporia prevents us from
making sense of the whole because we may not have access to a particular part.
Aporia here becomes a trope, a absent pice du rsistance rather than the usual
rsistance of the (absent) meaning of a difficult passage.
Complementary to this trope stands another; the epiphany. This is the sudden
revelation that replaces the aporia, a seeming detail with an unexpected,
salvaging effect: the link out. The hypertext epiphany, unlike James Joyces
sudden spiritual manifestation (Abrams 1981, 54), is immanent: a planned
construct rather than a unplanned contingency. Together this pair of master
tropes constitutes the dynamics of hypertext discourse: the dialectic between
searching and finding typical for games in general. The aporia epiphany pair us
thus not a narrative structure but constitutes a more fundamental layer of
human experience, from which narratives are spun. [Aar, 91-92].
As described here, aporia and epiphany are elements that are key features to the nature of hypertexts already. So there must be more to cybertexts than the aporia-epiphany pair, or else the invention of the term was obsolete. However, it isnt
The Text-Machine
When employing the notion of text as a text-machine, like Aarseth does [Aar p.21], the perspective opens up that enables to put in chart a vast field of modern textual works that wouldnt fit into classical literature concepts comparably natural and meaningful. It is not trying to render the traditional categorization of text obsolete, but rather broadening the conceptual power of the analysis tool without letting it loose its sharpness.
The notion of text as text-machine basically shifts the perspective from text
being the collection of signs in the middle of the chain author-text-reader
(or: sender-message-receiver) to a situation where the medium by which the
message is carried gets on a par with the verbal signs and the texts operator.
This enables to emphasizes the nontrivial contribution the reader (as an
example for the operator) makes to the issue of what he perceives at all.
Aarseth visualizes this by letting the three elements verbal signs, medium and
operator form a triangle (the text-machine) that represents the area, in which
the cybernetic intercourse between the three part(icipant)s happens [Aar
p21,22].
This concept of text serves as the necessary background for the exploration of the dynamics that occurs in the use of cybertexts.
Cybertext Typology
Aarseth performs an extensive analysis a on a diverse collection of texts on the basis of the in the following described classification according to his self developed typology model [Aar p65-75]. Without elaborating on the analysis itself I will here give a brief explanation of what this typology model is about as it provides a solid framework for the classification and understanding of all kinds of texts in our focus and thus will be useful to understand what will be developed later.
First, according to Aarseth, and in agreement with the notion of text as a text-machine outlined above, it is stated that
A text, then, is any object with the primary function to relay verbal information. Two observations follow from this definition (1) a text cannot operate independently of some material medium, and this influences its behavior, and (2) text is not equal to the information it transmits. Information is here understood as a string of signs, which may (but doesnt not have to) make sense to a given observer. [Aar p62]
It is further useful to distinguish between the textons of a text and the scriptons that are generated. The textons are the strings as they exist in the text, whereas the scriptons are the strings that appear to a reader. Obviously these are not the same as the example of error messages that an operating system shows. Here most people dont know (luckily) what the body of potential errors messages of their operating system (the textons) comprises. However, largely depending on their own actions, sequences of signs (the scriptons) - often repeatedly become displayed on the screen, forming a unique succession of signs represented.
Another important structural element of a text is the so called traversal function the mechanism by which the scriptons are presented to the user of the text [Aar p62]. A texts traversal function can be classified by the following variables that can take the described values[5]:
1. Dynamics: In a strict Text the scriptons are constant; in a dynamic text the contents of scriptons may change while the number of textons remains fixed (intratextonic dynamics, or IDT), or the number (and content) of textons may vary as well (textonic dynamics, or TDT)
2. Determinability: a text is determinate if the adjacent scriptons of every scripton are always the same; if not, the text is indeterminate.
3. Transiency: If the mere passing of the users time causes scriptons to appear, the text is transient; if not it is intransient.
4. Perspective: If the text requires the user to play a strategic role as a character in the world described by the text, then the texts perspective is personal; if not, it is impersonal.
5. Access: If all scriptons of the text are readily available to the user at all times, the text is random access (typically the codex); if not, the access is controlled.
6. Linking: A text may be organized by explicit links for the user to follow, conditional links that can only be followed if certain conditions are met, or by none of these (no links).
7 User functions: Besides the interpretative function of the user,
which is present in all texts, the use of some texts may be described in terms
of additional functions: the explorative function, in which the user
must decide which path to take, and the configurative function, in which
scriptons are in part chosen or created by the user. If textons or traversal
functions can be (permanently) added to the text, the user function is textonic.
If all the decisions a reader takes about a text concern its meaning, then
there is only one user function involved, here called interpretation.
[Aar p62-64]
In order to get a better understanding of the landscape modeled for cybertexts
it useful to give a figure putting the last variable (user functions) into the
context of what was outlined so far [Aar p64]:
This figure gives an idea of what the information feedback loop between text and user looks like. The interpretative user function is the one in which information flows from the text to the user whereas the other user functions represent the information flow from the user to the text. Consequently one can talk of a cybertext when at least one of the three user functions pointing towards the text can be identified. The interpretative user function as a matter of the fact present in any text.
Discourse planes
With that in mind Aarseth goes into the examination of another structure that emerges when examining cybertexts regarding the relation between revelation of events and user progression.
In traditional texts the events reveal themselves to the reader in the course of progression with the possible notion of this process forming a discourse plane. In hypertexts and cybertexts however the discourse plane divides into distinct planes of which there are the event plane, a progression plane and a negotiation plane.
In narrative, the discourse consists of the event plane, where the narration of the events takes place, and also what I call the progression plane, which is the unfolding of the events as they appear to the reader. Here [in narrative] these two planes are identical, as the readers progression follows the event line. In an exploratory ergodic text such as hypertext, the progression plane is divorced from the event plane, since the reader must actively explore and nontrivially make sense of the event plane. In adventure games, the relation between events and progression is defined by a third plane of discourse: a negotiation plane [Aar p. 125]
With this knowledge about cybertext structure in mind the development of multi user learning environments is equipped with a perspective on the layers that are present in the structures of the digital media they are modeled in. Knowing about these universal structures provides one with a framework that helps both interpreting and analyzing as well as synthesizing multi user learning systems.
3.4 Drama in Computer Human Interaction
Brenda Laurel
For the goal of designing a multi user learning environment framework another perspective has to be taken into account apart from the structural analysis provided by Aarseths cybertexts, one that puts the users experiences central.
In her work Computers as Theatre [Lau], Brenda Laurel develops a theory of analyzing and designing human computer interaction that follows principles derived from dramatic theory in theatre[6].
While her work is intended to serve as a model for analyzing and designing interfaces to single user tasks, I believe some of her notions are even valid in more stretched contexts than she allowed at the time of the writing of her book.
Whole action
One of the central concepts for successful human computer interaction is the notion of a coherently represented whole action, characterized by a proper beginning, middle and ending. Their explicit and coherent representation (enactment) leads to catharsis for the user when such a whole action ends successfully. Here success; does not mean the success of the task a user is trying to complete according to his or her personal measures of success of the task but whether the end of the action is properly fitting into a whole.
A whole action is made up out of incidents that are causally and structurally related to another. A good beginning of a whole action is accomplished by laying out effectively the potential for further action and setting up promising lines for further scenarios with the first incidents whereas a good ending is provided by a completion of the action represented and by providing an emotional and informational closure for the user.
A whole action also has to has to be perceivable as a whole. This implies that it has to be perceivable as a whole in terms of memory (represented complexity) and time span. It has to be not to long and not to short and have the appropriate magnitude for the time span it comprises. Laurel from the latter derives that a complete whole action must not last longer than about four hours as this is the maximum amount of time one can pay uninterrupted attention to some process (or its enactment).
Multithreading whole actions
While I principally agree on the conclusion concerning the limited time span of a whole action when regarding uninterrupted attention, the necessity of providing actions as whole actions must be extended to actions that operate on larger time scales. This is naturally done by allowing actions to be interrupted. This must not be misunderstood with an attempt to allow for actions that provide no whole action experience because one regards the comprising action as the one which provides for the well orchestrated ending. It is rather an attempt of modeling the notion of whole actions that allows for non uninterrupted representations thereof.
In real life (which includes work on a computer) the actions one commonly performs are rarely finished uninterrupted. Yet in practice there is no problem with multithreading ones actions - starting an email, being interrupted by a phone call, looking something up on the web, taking a note, talking to a colleague, drinking coffee, writing a letter, resuming editing the original email which actually happens constantly and usually is no problem for us[7], as long as the nature of the actions are such that they can gracefully be put to sleep and called awake again when one turns ones attention to them again.
On a different level that encompasses the action that is to be "put to sleep gracefully" the process of doing so has to form a proper action itself and has to provide a proper closure that allows one to take the focus away from the action to be interrupted. Thus the actions on different levels depend on each other to develop coherently. This illustrates that it is very often inappropriate to regard actions as isolated, monolithic or single layered entities.
This is paralleled by the continuity of presence of objects and other entities in time and space we can turn our attention to or away from.
For example one can be sure to find the book one is reading back the way it was put away when coming back after the walk to the kitchen to get a fresh cup of tea. In computer human interaction this coherence and continuity in representation has to be crafted consciously by the designer of the system. Thus to have a comparable coherent way of actions happening in parallel in a digital environment it has to provide a consistent and universal way enable and communicate the change of focus between different contexts that carry the different actions. An essential element for achieving such an experience of reliable and pleasant task switching is the clear representation of task context by the agent who is communication it.
The sense of agency is important to communicate the notion of for example an application being still there but sleeping or possibly fulfilling some (repetitive) task silently in the background like for example checking email.
The key difference (in this respect) between computer task experience and a (traditional) theatre play is that in general one cannot multithread real theatre plays in a comparable way as it is usually done with actions on a computer. This is mainly due to the fact that drama works as such - - be it for enactment on stage or in another medium is traditionally intended to experienced in a single attention span. But even this is obviously broken up by for example TV formats that span several editions of lets say a serial or series.
Naturally a play experienced continuously in full length is a much more beautiful experience than a play that is cut into pieces at the will of a spectator and reassembled in his or her mind only.
Obviously this discussion is not an attempt to state that chopped up actions that are still consistent within their eigentime yield the exact same experience as whole actions that are performed in a continuous manner. When there is a smooth way for the user to put a certain action on idle to resume it later, i.e. to interrupt its performance for a given time, then the partial reassembled pieces put together will have to fulfill the same structural requirements as whole actions that are enacted uninterruptedly in order to provide a comparable satisfying experience.
Agents
What has implicitly been assumed in these considerations is the presence of well formulated and well represented agents that carry out actions. Agents in this sense are the parallel to characters in drama. And as in drama, the notion of character is not restricted to be represented by humans at all. Although in drama characters are often bound to persons, they can be represented by any entity that by virtue of their goal(s) interferes with the overall plot by their own actions. The Dramatica New Theory of Story [Dra] handles the notion of a character very universal, namely as
a set of dramatic functions that must be portrayed in order to make the complete argument of a story. Several functions may be grouped together and assigned to a person, place, or thing who will represent the character in the story. [Dra ch.4]
Thus it is not uncommon to regard the concept of character sufficiently detached from the notion of a person, which parallels the notion that Laurel applies to the concept of an agent. According to her an agent is one who initiates action [Lau p4] which principally might be any distinguishable entity in the context of user(s) and a system. For the notion of an agent on the computer side, the term is concretized by her in the following way:
A computer-based agent is defined as a bundle of functionality that performs some task for a person, either in real time or asynchronously. [Lau p46]
Put like this the notion of what kind of entity a digital media based agent can be appears to be too. I will examine the notion of more general entities that can function as agents in our sense in the following.
More general Agents
Today it is well imaginable that without giving up the idea of user centeredness, that tasks are performed not for or by the one and only individual on one single user application but rather that these processes happen in a dialogue with more abstract agents that can be as fuzzy as for example a network community. This considerations are not uttered here to reject the notion of what agents are formulated by Brenda Laurel but rather to explore what such agentness possibly can mean in the context of networked people communicating through multi user applications with another and jointly with the application itself.
Whatever kind of entity the agency is attributed, it is essential to have a clear picture of the these that are involved as agents in the enactment of the user experience.
What characters do in theatre play, namely putting otherwise unrelated incidents in causal relation to each other, is performed by agents in human computer interaction. By communicating and having a clear goal, the otherwise unrelated incidents earn their causal relationships (causal according to the natural laws of the agents involved) and are put in a coherent and - with the lifetime of the agent comprehensive context for the user. [Lau p76]
The challenge of providing the user with a satisfying whole action task experience that has to be multithreadable puts the need for a clear and reliable communication of the specific context even higher on the priority list than it would be without. Again, this is naturally archived mainly by modeling the representation on well defined agents.
The break in experience when a whole action is aborted half way is equally experienced with the discontinuously performed action as with an ungraceful ending (i.e. not ending) of a continuously enacted action. The major difference with actions that can be put to sleep is that there is the danger of that they might be forgotten there like with a book one forgot to continue reading, it is a source of major frustration when the attempt to reestablish the relations between the incidents experienced after resuming the action with those that were experienced before the interruption fails. Even stronger is the frustration when the whole context cannot be reestablished anymore, e.g. the book is permanently lost.
Problems with multithreadable actions
While a lot of people do not consciously design their actions as multithreadable actions as graceful multithreading is handled by the representation of the real world working environment they are very often performed in such a way, that is, well interrupted by fragments of other actions and still perceived as a whole process.
However, in digital media environments it is not always trivial to archive graceful multithreadability as in many cases an interruption of the process is not possible technically without becoming an unfriendly one that implies a rough abortion.
So for example (on a contemporary PC) one cannot interrupt the copying of a file to resume the process at a later time in general. One either has to request the system to abort the whole process completely (which is a proper ending in the sense of a whole action) or one has to wait until its finished. A system level disk error half way for example would constitute a rather ungraceful abortion of the action in question.
By suggesting this extended idea of what can still be regarded as whole action, I implicitly offend against Brenda Laurels requirement that a whole action must not extend to a longer time span than a user can concentrate on it uninterruptedly. From my perspective a whole action may rather extend over a longer period of time as long as the continuity and coherence in appearance and behavior is maintained and the process of suspending the performance of the action and resuming it later bears to proper beginnings and endings for whole actions on the level of the action of suspension itself.
I do recognize that the ideal situation of uninterrupted actions provides the most comprehensive experience in terms of drama, and to archive this maximum, Laurels restrictions certainly apply. However, when such a uninterruptedness cannot be provided without forcing the user to work in an unwanted (and unnatural) single thread, the alternative is to apply the remaining principles of a whole action to the extended concept of whole actions that allow for multithreading, outlined above.
It must be noted that this extended notion of whole actions that do not necessarily have to occur in monolithic blocks, nicely connect to the feature of cybertexts of having separated planes of progress and events as described in the earlier discussion of Aarseths work.
This link comes into focus when one regards a user, who performs a given (overall) task, who naturally switches focus between different participating agents in order to achieve the overall tasks goal. For a whole action to be present, all levels of tasks and contexts involved have to form a whole action within their respective reference frame. However their experienced reference frame does not necessarily have to be mappable to the flow of time in a continuous manner, i.e. may objectively be regarded as interrupted.
The multithreadedness of a cybertext story might thus be regarded in terms of human computer experience of whole actions if one allows them to exist in several concurrent threads.
The possibility to design the interaction between a user an agents that represent whole actions (overall - on the level of the user task - - as well as on the level of their agent context) in a way that allows for multithreadability, is a key prerequisite for designing interaction between agents and multiple users.
Note, that tracing back the requirements to features that rely rather on an clear and unambiguous communication of context than on an uninterrupted representation, for what are necessary conditions for whole actions to be enacted and experienced, yields a much higher scalability of actions over time than Brenda Laurels original notion. For this to be the case it is essential that the context wherein the events occur gets communicated unambiguously. This allows for interruption of the continuous representation of the context even if it extends the time span of uninterrupted attention.
Constraining actions
An important aspect to be examined in this context is the
fact that according to Laurel the dynamics of a whole
action shows a well formed transition from a range of possibilities for
incidents via a (more constrained) range of probable incidents to a final
necessity of incidents. This dynamics commonly has a graphical representation
of a flying wedge [Lau p70]
In the beginning of the action everything within
the possibilities of the system, thus the potential, is possible. From left to
right the potential for subsequent actions gets narrowed by the occurrence of
the first incidents, for example by simply representing the context of the
action, so that some incidents become more likely to occur than others, i.e.
they get probable, whereas towards the end the potential set of incidents
narrows that far that one (or very few) incidents of those probable becomes
necessary and must occur for the whole action to be complete.
Types of constraints
This process of narrowing the
possibilities of incidents is achieved in human computer interaction by
communicating constraints to the user on what will possibly happen next.
Constraints are according to Laurel - either of extrinsic or intrinsic
nature. Extrinsic constraints are those that are communicated explicitly outside
the direct context of the possible incidents to happen for example the
command syntax of a script language that is provided through a manual or an
online help whereas intrinsic constraints are communicated through the
behavior of the system for example that one cannot chose the copy command
in the edit menu when there is nothing selected that could be copied, although
the menu entries (thus the functionality) is still there.
There is another type of
constraint that narrows the expectations of the user about what is about to
happen in a given context of an action that is neither of the types given by
Laurel. The constraints that arise from the use of metaphors in the
representation of an action are not intrinsic to the action in question nor are
they extrinsic.
The use and working of metaphors
are equally powerful and difficult to control from the creators point of view
as they build on an assumed (or unintentionally created) relation between a
representation and an entity it refers to. The behavior and value of the
representation is then inferred from the entity the metaphor (according to the
user) refers to if the metaphor works. The constraints that emerge from this
inference then transcend the duality of intrinsic vs. extrinsic constraints.
Freytag Graph
Returning to the nature of the
flying wedge, i.e. the narrowing towards the point where no or very few choices
are left from what incidents will occur next has another structural
perspective, the one of dramatic complication that usually is captured by a
representation called Freytag Graph [Fre]
