Primoridialism, Instrumentalism, and Constructivism

In our last two examples, agents in the model represented trees and abstract "live" cells.  But it's easy in agent-based modeling to start thinking of your population of agents as people, rather than abstract entities.  A very common way to view the world is that people are agents in a social environment, meaning that they have the agency to change their behavior according to that world.


So our task at hand here to to define a person using PS-I lingo.  There are three prevailing schools of thought, all three which have their pros and cons, but one of which we will focus on in a new model.  One way to tackle the problem is known as primordialism.  Also called essentialism, this school believes that the ego (or identity) is an unchangeable, singular, and endogenous attribute, endowed to people with certain nations, religions, cultures, or ethnicities.  This is often the default mode of understanding identity for many people.  Germans do this, Italians act like this, and Russians only see things like this.  The credo of primordialism might be that I do what I want because I am what I am.  The attractions to primordialism might be that it creates a well-organized world, allows for great predictive power after figuring out who is who, and it accounts for much of the stability in the world.  The problem is that it cannot explain variation among a population, how an agent might change it's identity through conversion or assimilation, and how a population might act differently in different circumstances.  Some notable primordialists include Robert Kaplan, Fouad Ajami, and Samuel Huntington.  You might imagine a PS-I model where each agent has one unchangeable identity.  What might be the catalyst of change in the model like this?

An alternative to primordialism represents the other extreme school of thought regarding identity, and is called instrumentalism or circumstantialism.  Instrumentalism believes that identity is a fluid and constantly changing phenomenon, and views ego as a preference.  Basically, anyone can be anything.  The credo for instrumentalism might be I do what I do because I know what I want.  This is an attractive way of understanding identity because it is consistent with liberalism and consumerism, it easily explains how conversion and assimilation work, and can allow for nations to act in different ways when encountering different circumstances.  The downside is that it has difficultly accounting for stability in the world, fails to explain why everyone doesn't always have the most favorable identities, and why people often die for their nation, ethnicity, or religion.  Some notable instrumentalists include Edward Miguel and Dan Posner.  An instrumentalist PS-I model might be one where every agent chooses whichever identity it is most surrounded by, which would be the most favorable at the moment for a given agent.  Every agent would have the ability to change its identity on a whim, and as frequently as necessary.  This would create either a very stable landscape (where everyone has one identity) or a very fluid landscape (where the most prevalent identity is changing every few time steps).

The third school of thought, which has been implemented in PS-I, is constructivism.  Constructivism believes that everyone has a repertoire of available identities, one of which is activated while the others are subscribed.  An agent most prefers to keep its activated identity, but if it must change, it would prefer to choose from its repertoire.  If there are no favorable identities in its repertoire, then only under extreme circumstances would it choose an identity from outside of its repertoire.  This is a combination of primordialism and instrumentalism in which although identities are sticky (as in primordialism), there is substantial fluidity in certain cases (as in instrumentalism).  The credo for constructivism might be I do what I do because I am what I am AND I am what I am because I became what was wanted.  Constructivism is attractive because it can explain institutionalization, and why institutions change, explain how assimilation and conversion occur, and can explain how elites have the ability to present "noble lies" to the people (in the words of Plato).

I will be sure to elaborate on constructivist political theory in later blog posts, but at the moment, I would like to present a more complicated model than the previous two posts, which you can download here.  Load up the model (simpleA.mdl) in PS-I and click Open Agent Viewer as well as New Field Viewer in the View menu. As you move your cursor over different agents, you'll notice that each agent has an activated identity (with * around it) in addition to a repertoire of five other identities.  When an agent counts the number of different identities surrounding it (including itself), if the weight of its activated identity is not the highest weight, it will start to consider either rotating its activated identity with an identity currently in its repertoire, substitute an identity in its repertoire with an identity outside of its repertoire, or substitute and activate an identity from outside of its repertoire.  All three of these choices must have a trigger, and the triggers in this model are 2, 5, and 7 ,respectively.

Try running this model to see how it works.  You'll notice that at some point, the model stops running because all agents have changed their identity using their triggers and there are no changes left to make.  In the main window, you can change several parameters: number_of_identities denotes how many identities to start the model with, Sight_range is how far one agent will look to see their neighbors, Influence_level is how influential each agent is on other agents, Rotation_trigger is the trigger that causes an agent to rotate its activated identity for an identity in its repertoire, Substitution_trigger is the trigger that causes an agent to substitute an identity in its repertoire for an identity outside of its repertoire, and Activation_substitution_trigger is the trigger that causes an agent to both subscribe and activate an identity from outside of its repertoire.  Please try changing these parameters and see how it effects the behavior of the model.  As always, you can open this model using notepad to see the code.  It's only about 75 lines long, so I recommend taking a look at it to get an idea what's going on under the hood.

Much of this post was adapted from a lecture by Ian Lustick.