Define Me: A Cognitive and Computational Approach to Critical Digital Identity Representation in Social Networking Applications

Invoking a theoretical framework situated at the intersection
of humanistic accounts of social identity
construction, cognition linguistics research, and digital
media technologies, the Advanced Identity Representation
(AIR) Project develops theory and technology
for users to represent complex, dynamic social identities
in digital media such as virtual worlds and social
networking sites. Here, we primarily present DefineMe
– Chimera, a social networking application that uses a
dynamic system of categorization and allows users to
define each other through metaphorical projection. DefineMe
is grounded in an interdisciplinary approach that
articulates the shared socio-cognitive substrates beneath
user representations ranging from user created profiles
on social networking sites to avatars in virtual worlds.
Secondarily, we present Identity Share, a social networking
project developed using the DefineMe database
structure that allows users to define identity categories,
share profiles, and anonymously follow each other’s web
searching paths. The result of the projects is an early articulation
of a spectrum of new user identity representations
with foci upon group membership, utilization/creation
of boundary infrastructures (Bowker & Star 1999;
Lave & Wenger 1991), along with cognitive models of
metonomy, metaphor, and visual imagery. (Hutchins
1996; Lakoff 1987)
1. Introduction
The Advanced Identity Representation (AIR) Project
is the name given to the research endeavor in Fox Harrell’s
Imagination, Computation, and Expression (ICE)
Lab/Studio investigating technology and theory to enable
digital experiences that engage a richer range of social
identity experiences than those found now in social
networking, gaming, and virtual worlds software. We
present DefineMe – Chimera and Identity Share as early
steps toward this end. DefineMe – Chimera is a social
networking application with a novel database system
and a front-end Facebook web application. Users can
label each other using self-defined predicates expressing
their metaphorical similarities to various animals.
These descriptions are used as a basis to construct and
reconfigure categories on-the-fly as the database grows
and to present chimera-like avatar characters to represent
the user as composites of various iconic animal graphics.
Though this project develops a whimsical, metaphorical
model of user representation, the theoretical and technical
underpinnings address issues such as coconstruction
of identity categories between individuals, marginalization
and centrality within identity categories, and the
imaginative nature of identity in race, ethnicity, and gender
critical contexts. As a second step toward enabling a new genre of digital
media identity experiences, we present Identity Share, a
critical web-based application that offers a balance between
allowing users to author profiles with both selfdefined
and normative social categories, at the same
time as allowing users to specify the relative importance
of particular categories. Identity Share allows users to
anonymously give others permission to follow their web
searches, view their wishlists with various websites, and
leave comments on their experiences of “(web)walking
in another’s shoes.” The goal is not to connect users as friends, but rather to allow users to have the uncanny experience
of viewing and sharing aspects of each others’
needs, values, and desires. Together, DefineMe: Chimera
and Identity Share exemplify early prototypes of the direction
that AIR Project systems may take in tackling
social identity phenomena in the future.
2. Theoretical Framework
The AIR Project draws on a hybrid approach to issues
of categorization and classification. Central influential
theories important to the AIR Framework are described
below.
2.1 Cognitive Categorization, Metaphor, and Blending
The AIR approach has some of its roots in grounded in
cognitive science theory (Lakoff 1987) which asserts
that categorization is a matter of both human experience
and imagination. George Lakoff’s work in this area over
two decades ago is well known and influential, yet to
our knowledge it is a thread that has been underdeveloped
with respect to issues of social identity construction
in the critical modes robustly developed in cultural
studies with the humanities, especially this approach has
not been significantly applied to cases of digital identity
representation (an exciting exception being the work
of Otto Santa Ana on metaphorical bias in Brown Tide
Rising (Santa Ana 2002)). Cognitive science research reveals
categories as being (1) based on “the same neural
and cognitive mechanisms that allow us to perceive and
move around” (Lakoff & Johnson 1999), (2) distributed
across members of a social group, external artifacts, and
even time (Hutchins 1996, 2000), and (3) always situated
in particular social and cultural contexts (Lave &
Wenger 1991). Important for the purposes here, Lakoff
describes a conceptual metaphor-based theory of how
imaginative extensions of “prototype effects” result in
several phenomena of social identity categorization that
are useful for the AIR Project (Lakoff 1987). These phenomena
include representatives (prototypes) or “best example”
members of categories, stereotypes that indicate
normal, but often misleading, category expectations, and
more. Conceptual blending theory builds upon Gilles
Fauconnier’s mental spaces theory (Fauconnier 1985),
elaborates insights from metaphor theory (Fauconnier
2006), and attempts to account for a wider range of semantic
phenomena.
2.2 Sociology of Classification Infrastructures
The AIR Project is influenced by accounts of classification
from sociology and science studies. In Sorting
Things Out (Bowker & Star 1999), Geoffrey Bowker
and Susan Leigh Star make the case that classification
systems are necessary for information exchange and
communication. The social challenges regarding classification
systems arise from cases where tension exists
between contexts, for examples, when one’s self-conception
differs from prevalent social stereotypes. Important
tools for bridging between communities are “boundary
objects,” defined by Bowker and Star as objects that “inhabit
several communities and also satisfy the informational
requirements of each.” The AIR Project develops
what Bowker and Star term “boundary infrastructures.”
These are defined as “stable regimes managing multiple
boundary objects, allowing the necessary information
to be accessed by multiple communities.” Also crucial
from Bowker and Star, is the concept of “torque,” the
condition where biographies are “twisted in classification
systems” to arrive at painful lived experiences. One
poignant example Bowker and Star present is the schism
between societal and self-perception and the disruptive
movement between or misapplication of categories, especially
for people labeled as “black” or “colored,” in
apartheid South Africa. The gap between self (or local
community-based) definition of an individual’s place in
a classification system and hegemonically imposed definition
of classifications, and disarming the negative results
often arising from such phenomena, is a central to
the critique performed by the systems highlighted in this
paper. As opposed to computational identity applications
that are based on standard, static classification systems,
the dynamically configurable, imaginatively grounded
AIR Project identity systems are boundary infrastructures
that allow users to customize their user profiles and
preferences for particular communities.
3. The AIR Framework
Based upon the cognitive and infrastructural approach
above, and previous work in imaginative computational
discourse and identity construction (Harrell 2007,
2008a), a brief summary of key aspects of the AIR Project’s
new constructs for implementing and analyzing
computational identity follows.
3.1 Shared Technical
Underpinnings of Identity Applications
A technical infrastructure-oriented means to compare
computational identities is the first pillar the approach
developed in this project. Various computational identity
applications such as social networking sites, avatar creation
systems for virtual worlds, and games are implemented
using a limited and often overlapping set of techniques.
Fig. 2 describes, at a high level, the components
that comprise the majority of widely used computational
identity technologies (Harrell 2008b).
The six components in Fig. 2 that commonly form the
basis for avatar/character/profile construction can enable dynamic and contingent models of social identity in
digital environments as described in (Gee 2003). Understanding
the reciprocities and overlaps between the technical
means by which users stage their identities across
digital media forms can enable more powerful customizability
and cross-community communication facility in
social identity systems. 3.2 Cognitively Grounded Model of Computational
Identity
The cognitively grounded model of computational identity
of the AIR Project is summarized in Fig. 3. It forms
the second pillar of our approach. Our digital identity models serve to critique infrastructures
of social classification that can, unfortunately, often
serve to reify naïve models of identity, and which do no
capture the dynamic, constructive, and enacted identity
phenomena encountered in everyday experience. This
model is an analytical construct used to help to understand
the interplay between underlying infrastructures,
such as the technical underpinning described above, and
the subjective interpretation of digital identities. The
utility of the model is that we can identify where schisms
exist between a technical structure (e.g., a data structure
specifying a player character as a string called a “priest”
and an associated “heal” procedure that allows addition
to an integer called “hit points) and a real world idealized
cognitive model as encoded in a classification infrastructure
such as “occupation description” (e.g., the description
of a priest perhaps as either someone versed in a
metaphysical body of knowledge or as merely the facilitator
of a particular type of institution). We can then construct
new infrastructures, using techniques such as suggested
in the AIR Project, that more closely align these
structures and models in order to construct the hybrid
of computationally afforded identities and real-world
identities that James Gee calls the “projected identity” as
shown in cognitively grounded AIR model (e.g., a player
taking on the role of a priest in a computer role-playing
game and trying to be helpful and supportive to her or
his friends). The key here is that our understanding of
both computational structures and the ways that users interpret
them is based in imaginative cognition processes
such as conceptual categorization and blending.
4. DefineMe – Chimera: A Critical
Identity Construction Social Networking
Application
The first system constructed using the AIR theoretical
framework is a Facebook application entitled DefineMe,
the first version of which is called Chimera. Specifically,
we implement aspects of Lakoff’s metonymic idealized
cognitive models for categorization to allow users to
co-construct their own and others’ avatars as boundary
objects. (Lakoff 1987) The premise behind DefineMe is
to allow users to define each others’ avatars using both
commonplace and abstract metaphors. Users can append
metadata to other peoples’ profiles to drive dynamic
generation of avatar images. The initial content domain
consists of animal metaphors that can be mixed-andmatched
algorithmically. Animal metaphors are potent
entrenched metaphors for human personality. (Turner
1996) (e.g., sneaky weasels or docile sheep), however
this animal metaphor-based version is only an initial experiment.
The model extends to more directly socially
engaged categories such as social scenes, fashions, or
movements.
The DefineMe database is designed to be lightweight,
dynamic, and extensible, while implementing categorical
relationships between members. When comparing
profiles, DefineMe is designed to match lexical items and logical relations directly, or it can compare the structures
of profiles following insights from the analogical structure-
mapping engine (SME) developed by Ken Forbus
et. al. (Forbus 2001; Gentner 1983) The DefineMe – Chimera
application reported on here focuses on creating
metaphorical projections as described above. The DefineMe
database relies on tags to create additional descriptors
for each category or member. For instance, one user
could describe her friend as a ‘lion’ (which would be the
member) because she is ‘strong’ (which is the tag). Another
user could add an additional tag, stating that she is
a ‘lion’ because she is ‘carnivorous.’ These tags can comprise
vertical parent-child links (e.g. a ‘lion’ is-an ‘animal’)
or horizontal implicit links (e.g. in another user’s
profile a ‘lion’ is-an ‘Ethiopian symbol,’ yet the system
may still create a category linked by the concept ‘lion’).
The fact that users define other users has the potential to
both entertain and agitate, regardless it allows for critical
inquiry into the phenomena echoing real-world labeling. Following the work of Eleanor Rosch cited in (Lakoff
1987), the tagging system can also be used to define aspects
of categories themselves. For instance, a ‘robin’
tag can be added to the category, ‘birds,’ to define the
prototype of that category. In this way, members caare not seen as drawbacks of the systems. Each system is
considered to be a culturally situated critical intervention,
rather than a usability oriented productive application. In
bringing to light more nuanced and imaginative identity
phenomena, such as potential ostracism, prejudicial exoticizing
of the other, or unflattering labeling, we hope to
also provide the potential to disempower such phenomena
through dialogic engagement. These systems can
be considered cultural productions, or digital media art
projects, in the sense that they are provocative cultural
interventions situated in an environment increasingly
encroached upon by hegemonically enforced, often corporate,
models of user identity. As such, the systems succeed
only to the degree that they engage users as evocative
systems, inspire critical thought, and are construed
as adequate for capturing personalities using archetypical
avatars or conjure the sensation of experiencing the
web through another’s eyes. Beyond this, however, we
see the systems as prototypes that suggest directions that
could enhance the expressive and empowering potentials
of productive, utilitarian, or commercial systems such
as computer games and popular social networking sites
with features such as self-definition of categories and deployment
of imaginative metaphor. Despite our provocative and critical interventionist
stance, the systems are engineered to mitigate against
abuse, and certainly distress of users is not our goal.
Looking at the two systems consecutively, mitigating
factors designed into the systems are as follows:
DefineMe: Chimera Design Factors
1. Users are only allowed to tag their Facebook
“friends” who have added the application.
2. Users have access to a limited database of animaltypes.
3. Users must “opt-in” in order to receive a generated
avatar.
4. Users can “opt-out” at any time.
5. Users’ database entries can be edited by moderators.
6. Users have access to only a limited format for tagging
each other.
7. Users can delete entries on their profiles that others
have created.
Together, we believe that these factors strongly help to
avoid the system’s potential to be applied in an overly
negative manner. It is a contract between friends to sign
up for potential compliments, teasing, and, we believe,
self and social insight. Ultimately, DefineMe – Chimera
is intended to present users with a controlled experience
of torqued identity. The fractured identity of a monstrous
chimerical representation is then, an accurate reflection
of the limitations of applying modular and discrete classifications
to a real world biography.
Regarding Identity Share, mitigating design factors implemented
include the following:
Identity Share Design Factors
1. Users can create their own self-classifications.
2. Users can select which classifications are important
to them.
3. Users can avoid or utilize normative categories such
as gender or occupation.
4. Users can allow or disallow the system’s tracking of
their web visitation paths at will.
5. Users’ real world identities are kept anonymous. n belong
to multiple groups, but individuals can represent the
prototypical members of groups. In this early version,
each user is seen as a member of each assigned animal
category as well. This membership allows the system to
use an individual as a prototype stand-in for the category.
For instance, rather than just labeling a friend as a lion,
one could state that your friend, Emily, is like your friend
Bobby because she is brave. The system can then take
all of Bobby’s attributes and apply them to Emily’s chimera.
This relatively lightweight structure avoids some
of the pre-defined categorization built into many social
networking infrastructures, and has the potential to explore
some of the more nuanced identity phenomena
mentioned in the theoretical framework above.
5. Identity Share: A Critical Identity
Construction Social Networking
Application
Identity Share, a social networking site for “non-friends,”
and Daniel Upton’s MS thesis project in Digital Media,
is also developed under the umbrella of the AIR project.
The system allows for social networking by providing
users with facilities to construct profiles, follow and
comment upon other users, and perform game-like tasks
that encourage users to consider exploring both like and
different profiles of others. Identity Share offers a novel
means of self-representation based upon open-ended categories
and tags. Standard profile models that typically
include normative categories such as name, age, gender,
location, and race are replaced with a customizable list
that exists as a database, growing as more categories are
added. Database consistency is maintained by giving users
typeahead
functionality when adding custom categories
and by presenting existing categories in order from
most common to least common. The database structure is
based upon the same layout used by DefineMe – Chimera.
Users can select which categories are most important
to them by indicating that they are primary to the user
using checkboxes. By allowing for primary selection
of categories, we consider the system to implementing
centrality phenomena from the cognitive science theory
above, i.e., “the idea that some members of a category
may be ‘better examples’ of that category than others,”
to a users profile. (Lakoff 1987) This means that a user’s
profile, as a collection of categories that define a user,
is no longer viewed as just a set of static characteristics
that are true about this user, but rather as a complex set of
characteristics where some may be “truer” or more definitional
to the user’s self-conception. To take this even
further, in a future implementation Identity Share could
offer a ranking system for each category, thereby not
only providing centrality, but a centrality gradience, “the
idea that members (or sub-categories) which are clearly
within the category boundaries may still be more or less
central.” (Lakoff 1987) This offers a new dynamic to social
network profiling that doesn’t currently exist on the
popular social networking sites.
6. Ethical and Humanistic Implications
When social stakes are low, many people are inclined to
reveal their baser selves. Indeed, in a project such as DefineMe
– Chimera the potential for using the system to
ridicule is quite apparent. Likewise, the ability to anonymously
follow users’ web usage experiences in Identity
Share offers a potential that may seem to verge on the
voyeuristic. Yet, these potentially disempowering uses 6. Users’ perceived affordances to communicate with
one another are highly restricted.
7. Users have full control to delete any of their data in
the system.
These factors were developed over the course of iterative
refinement of the project based on informal user feedback
(mainly via open-ended interviews). The greatest
challenge with the system was to allow for user generated
categories while also pruning sparsely used and
idiosyncratic database elements. A second challenge
regarding anonymity and privacy is addressed by careful
controls such as articulated above, and by providing
quite clear and prominent information on the nature of
the site. Quite contrary to being a site to allow people
to “spy” on others, it is an “opt-in” site oriented toward
users with a desire to share their personal styles, definitions,
and web behaviors with one another. Finally, it
is a system that is proposed as a balance between the
limited and discrete, yet highly modular and structured,
information structures provided by digital media and the
continuous and transient, yet not computationally amenable,
phenomena of identity as shared in the real world.
7. Conclusion
The AIR Project examines the humanistic implications
of emerging technologies by seriously considering the
cultural effects of user identity within current digital
media and the shared sociocognitive
foundations that
ground their construction. Following various accounts
describing the procedural nature of the computational
medium (Manovich 2001; Murray 1997), the AIR Project
looks at the underlying data structures and algorithms
and how they implement cultural identity effects, and
posits a technical framework for more deeply engaging
identity semantics of classification and categorization.
Technologies for implementing socially empowering or
expressively critical and transformative experiences are
necessary to create experiences to engage real identity
social phenomena that lie at the center of so many of our
political debates and rich critical fictions.
Acknowledgements
We would like to acknowledge early assistance from the
undergraduate researcher Richard Shemaka. Additionally,
we thank Carl DiSalvo for significant input on Identity
Share as Daniel Upton’s coadvisor
with Fox Harrell.
References
Bowker, G. C. & Star, S. L. (1999). Sorting Things Out:
Classification and Its Consequences, MIT Press, Cambridge,
MA.
Fauconnier, G. (1985). Mental Spaces: Aspects of Meaning
Construction in Natural Language, MIT Press/Bradford
Books, Cambridge.
---- (2006). ‘Rethinking Metaphor’, in Cambridge Handbook
of Metaphor and Thought, ed. R. Gibbs, Cambridge
University Press, Cambridge, U.K.
Forbus, K. D. (2001). ‘Exploring Analogy in the Large’,
in The Analogical Mind: Perspectives from Cognitive
Science, MIT Press, Cambridge, MA.
Gee, J. P. (2003). What Video Games Have To Teach Us
About Learning and Literacy, Palgrave Macmillan, New
York City.
Gentner, D. (1983). ‘Structure-mapping: A theoretical
framework for analogy’, Cognitive Science, vol. 7, no.
2, pp. 155-70.
Harrell, D. F. (2007). Theory and Technology for Computational
Narrative: An Approach to Generative and
Interactive Narrative with Bases in Algebraic Semiotics
and Cognitive Linguistics, Dissertation thesis, University
of California, San Diego.
---- (2008a). ‘Algebra of Identity: Skin of Wind, Skin of
Streams, Skin of Shadows, Skin of Vapor’, in Critical
Digital Studies, eds A. Kroker & M. Kroker, University
of Toronto Press, Toronto.
----(2008b). Digital Metaphors for Phantom Selves:
Computation, Mathematics, and Identity in Speculative
and Fantastic Fiction and Gaming paper presented
to The 29th International Conference on the Fantastic in
the Arts, Orlando, FL.
Hutchins, E. (1996). Cognition in the Wild, MIT Press,
Cambridge, MA.
---- (2000). Distributed Cognition, paper presented to
International Encyclopedia of the Social & Behavioral
Sciences, University of California, San Diego.
Lakoff, G. (1987). Women, Fire, and Dangerous Things:
What Categories Reveal about the Mind, University of
Chicago Press, Chicago, IL.
Lakoff, G. & Johnson, M. (1999). Philosophy in the
Flesh: The Embodied Mind and Its Challenge to Western Thought, MIT Press, Cambridge, MA.
Lave, J. & Wenger, E. (1991). Situated learning: Legitimate
periferal participation, Cambridge University
Press, Cambridge, U.K.
Manovich, L. (2001). The Language of New Media, MIT
Press, Cambridge, MA.
Murray, J. H. (1997). Hamlet on the Holodeck: The Future
of Narrative in Cyberspace, Free Press, New York.
Santa Ana, O. (2002). Brown Tide Rising: Metaphors of
Latinos in Contemporary American Public Discourse,
University of Texas Press, Austin.
Turner, M. (1996). The Literary Mind: The Origins of