ps-i: political science - identities ps-i originally developed to test constructivist theories of...
TRANSCRIPT
PS-I: An introduction
PS-I: Political Science - IdentitiesPS-I originally developed to test constructivist theories of identity
adoption and substitution “Identities” represent public information that is shared and evaluated
by individuals Identity “bias”: global information available to all individuals about
the relative attractiveness of each identity Identities can be designated as:
“permanent” (if an agent has the identity it cannot lose it) “unobtainable” (if an agent does not start with the identity it cannot obtain
it) The “meaning” of any identity is derived from modeling decisions
about how identities are distributed across a landscape For example, an identity that is distributed only to individuals in one region
of the landscape and (possibly) is permanent and unobtainable could be thought of as a regional/ethnic identity
The number of identities (“spectrum”) in the landscape is a modeling decision (up to 56)
An overview of “agents”Agents are the decision-making individuals at the
heart of the modelTypically, agents will have a “repertoire” of
identities: a subset of the spectrum of identitiesOne of these will be the “active” identity: the identity
the agent displays publicly to other agents (its “color”)
Typically, agents have a “range” of 1: the Moore Neighborhood
Typically, agents will have a “self-influence” of 1 (meaning the agent considers itself with equal weight to its neighbors)
Agent “decision-making”At each iteration, every agent does the
following (unless the agent is “immutable”):Determines the agents in its “neighborhood”
(excluding “inactive” agents)Counts the number of agents in its neighborhood
activated on each identity (each agent contributes the quantity equal to its influence; if no agents in the neighborhood are activated on a particular identity, its value is set to 0) and adds the current bias value for each identity; this is the “identity weight” for each identity for each agent
Agent “decision-making” Next, the following four determinations are made:
1. The discard candidate is the identity in the repertoire with the lowest identity weight (excluding permanent identities)
2. The swapout candidate is determined the same way excluding the activated identity3. The rotate candidate is the identity in the repertoire with the highest identity weight4. The acquire candidate is the identity not in the repertoire or unobtainable with the
highest identity weight Now, the following sequence is executed:
1. If the activated identity’s weight is at least as much as all other identity weights, the repertoire is left unchanged
2. Otherwise, if the rotate candidate’s identity weight is greater than (the activated identity’s weight + the level 2 parameter value), it becomes the activated identity
3. Otherwise, if the acquire candidate’s identity weight is greater than (the activated identity’s weight + the level 3 parameter value), the activated identity is discarded, the acquire candidate is acquired, and the agent becomes activated on the acquire candidate
4. Otherwise, if the acquire candidate’s identity weight is greater than (the swapout candidate’s identity weight + the level 1 parameter value), the swapout candidate is discarded and the acquire candidate is acquired
5. Otherwise, the repertoire is left unchanged
Customizing agent classesImmutable (it does not change but still influences its
neighbors)Inactive (it can change but does not influence its
neighbors)Change the influence level (the relative weight of the
agent in the identity weight calculation of its neighbors)Change the self-influence level (the relative weight an
agent gives itself in the identity weight calculation)Change the sight radius (the size of the local
neighborhood)Change the Level 1, 2, and/or 3 parameters (altering
the likelihood of changes to the repertoire)
Customizing landscapesSize and shape of landscapesRelative distribution (random or otherwise) of
classes of agentsBias value range(s), the likelihood
(“volatility”) of changes to bias values, and the nature of changes to bias values
Intentional seeding of identities