Eric Leifer is presenting a theory that attempts to explain how roles come into existence -- how it is that people sort each other into roles and positions.  He posits a notion of 'local action' as the mechanism through which interaction partners define each other's position.  Local action allows actors to buy information from each other as to the probability that a role will be conferred by the alter. In so doing, actors help lessen the risk of taking a role before that role is freely granted.  The theory provides us with insights into the day-to-day interaction of peers, the historical dependency of the 'meaning' of any set of actions, and the strategic options open to actors.  The paper starts with a discussion of the requirements of role behavior, then moves on to discuss the vulnerability of role behavior, posits the central points to local action, places it in the context of current & classic work on reciprocity, and ends with a discussion of power and multi-level interactions.

Roles
     Norms - the kinds of behaviors we engage in – follow from roles.  A teacher knows how to act AS A TEACHER, a student AS A STUDENT, a soldier AS A SOLDIER, etc.  Each of these roles could be filled by the same person, it's the context and the expected script that shapes the specific actions that role occupants take.
     However, all these examples are of already decided roles.  Once a person is sorted into a particular position, then behavior follows a (relatively) predictable pattern. When a role, especially one that confers status, is sought, how do people go about claiming, and getting, the role? That is the main question of this paper: what is the salient ideal for acquiring as opposed to occupying roles.
     Contests for roles are most interesting when the relative positions of actors are equal, and because larger forces tend to pre-sort people of similar ability and background into similar spaces, this is where most of the day to day struggle for roles happens.  In addition, in cases where actors of unequal ability are in conflict, the outcome of the contest is easy to predict. Thus, "The face to face struggle for a coveted role is the most intense when screening mechanisms have already run their course."

Leifer's Assumptions
1) Actors seek status and gain status by acquiring status-conferring roles.
2) Legitimate status difference is possible with in a setting and there is no basis for establishing roles other than through the actor's own efforts.  <this is the equality constraint: that the kind of interactions where local action is probable are ones where actors can determine through their own efforts the roles to be had>.

Leifer's action ideal is called LOCAL ACTION.

     "Local action allows actors to avoid claiming a (global) role until there is evidence it will be conferred. Ironically, when both actors use local action, status aspirations are perpetually suppressed. "

The Vulnerability of Unilateral Role Behavior.
     Role behavioral depends on its setting for interpretation.
     Example: A smiles at B.  This smile means something different if its from a strange man on the street corner wearing a trench coat as opposed to a well-dressed business man walking past in the hall.  IF ROLES ARE NOT ASSIGNED, ACTIONS MAY APPEAR CONSISTENT WITH DIVERGENT ROLE SETTINGS.  The same act may be interpretable in many frames.

     Roles require role compliments.  A person cannot be a leader without a follower, a lover without a mate, a benefactor without a recipient or a student without a teacher. If the interaction partner refuses to play the 'compliment role' (you play student to my playing teacher), then the claim one makes to the role is thwarted.

Thus, the claimant of a coveted role is dependent on someone accepting a complementary role.

The power that A's role behavior confers to B depends on:
     1) A's level of Dependence on B for a status claim
     2) The alternative responses open to B.

This then makes up the first fundamental principle of roles
     All roles need a compliment.

     The second principle follows from the fact that before the role setting is established, an act can be linked to more than one role.  The absence of a role, means that B can respond to A in almost any 'acceptable' fashion – because what counts as 'acceptable' is exactly what is up for grabs: the definition of relative role status.

     Thus there is a 'peculiar symmetry' where the same actions that confers status can also take it away, depending on the responses they elicit.  How B decides to react to A changes the meaning of whatever behavior A started with. [This is VERY similar to what Arendt was speaking about when she said that throwing an action into the world had unpredictable outcomes].

     Given that an act is dependent on the response it elicits for meaning, what ideal can we posit for acting that helps to establish roles?

Local Action.
     Because roles are taken when the other person decides to 'give' them, we need to know when to lay claim to a role. Leifer's solution is LOCAL ACTION: Behavior that is deliberately under-determined, such that many interpretations for any reaction are left open.  This allows each participant to buy information -- we can see how someone else acts, and thus make an inference to how they MIGHT act, if we lay claim to a particular role.  Actions gain their significance from the role possibilities they leave open.

     "Nonrole-specific action is needed that, ex ante, leaves open a range of roles and, ex post, does not prove inconsistent with any role that might be claimed later."

     Thus, local action marks a deferring of status aspirations: one doesn't act to maximize a particular outcome, instead, one acts to get information.  This means that any theory which attempts to describe actor's behavior in terms of such a maximizing proposition will fail.  See the footnote 3 on page 869.

     The outcome of people engaged in local action is balanced reciprocity.  Each actor throws out an action volley, and sees what the other person is going to do.  This mechanism, of information searching, provides a clean, clear reason for the often observed 'generalized norm of reciprocity.'

     The Problem of the norm of reciprocity is that it assumes that people are willing to take on one role or another, (i.e. giver or taker) without regard to why they would give that role up.  If there status to be gained from being a taker, then the wise actor capitalizes on the other's foolish giving.  At the same time, if the giver can make the taker look like a leech, perhaps status can be gained.  The problem then, is that there is nothing INTERNAL to a generalized norm of reciprocity, that would give rise to reciprocal action.  The NORM of reciprocity under theorizes roles.  Local action allows for re-action BECAUSE roles are undefined.

In class, we discussed the ex-post nature of meaning that follows from Leifer.  If actions and roles are constantly up for grabs, the story you tell about a series of events can never be absolutely determined, especially by the participants. (this should get you thinking, again, about Arendt).

     The second point was to be made about the relation between the relative skill of actors and the potential for fresh action.  Looking at the 2 x 2 table of hi and lo skilled actors, we saw that a contest between to actors of unequal skill yields pre-known outcomes.  Action between two equally matched POOR actors is largely an arena of LUCK, and if new unique action is invented, neither have the capital to see it.  Only in the space of equally skilled and (to use some stuff from Sewell) knowledgeable actors, can we expect to see true innovations.

Basic points to take from the reading:

• People seek status
• Status is only achieved by occupying a status giving role
• The problem is thus identifying how one gets a desired role
• Getting a role is complicated because:
• The status of a role depends entirely on what another person (the role compliment) does
• Any given role behavior can have multiple meanings
• Local Action is the role ideal that Leifer proposes to identify how roles are achieved.
• local action consists of doing things that have ambiguous meaning in order to get information from others.
• From Leifer's discussion of local action we see that:
• The meaning of an event is contingent on the events that come after it
• Status differentiation comes through a search process rooted in ambiguity
• reciprocal action can be explained without reference to a governing 'norm of reciprocity'.

After discussing Leifer, we went on to discuss rational choice theory in general and game theory in particular.

Rational Choice Theory refers to an explanation of human behavior that is rooted in how people make behavior choices.  The theory says that people decide to do things based on the returns any such action would give them.  They weigh the cost of each possible action against the benefit, and make the choice that leads to the largest benefit.  While seemingly simple, this theory of action can have dramatic consequences when many people are making similar choices based on similar criteria, and thus has been very useful in explaining social behavior.

One of the most active areas of research  in the general field of rational choice has been within game theory.  Game theory is an area of mathematics, that seeks to understand how people will behave based on a general rational calculus.  By making payoffs (the result of action) contingent on other actor's actions, we gain insight into the social outcomes of individual action.  One of the most useful examples of game theory is found in the prisoner's dilemma.  The prisoner's dilemma is a classic game, and is usually set up something like this:

    Scenario:  Two prisoners, each committed a murder and a robbery.  The sheriff can prove the robberies, but not the murders.  The sheriff makes the following proposition to each person:
 
    "If you care to confess to the murder, and your partner does not, you will go free.  Your partner will be charged and executed but you will go free; and, of course, vice versa.  If your partner confesses and you do not, then you will be executed and your partner will go free.  If you both confess, you will both be charged with murder, and as a reward for being helpful, will revive a ten year sentence.  If neither confess, we can't prove the murders and you will get two years each for the robber.  This offer is being made to both of you."

What will the prisoners do?

We can look at this situation with the following payoff matrix, with Jill's actions in the top row and Jack's actions in the left column:
 

Payoff Matrix	Jill remains silent	Jill confesses
Jack remains silent	Jack: Gets 2 years (3) Jill Gets 2 Years (3)	Jill is let go, (4) Jack is killed (1)
Jack confesses	Jack is let go (4) Jill killed (1)	Jack: 10 years (2) Jill: 10 years (2)

Now we can assume that each actor will rank these options, and here the ranking is pretty clear, and we can assign each points.  The highest ranked outcome is equal to 4 points, the next to 3, etc.

Going Free - highest payoff for the individual  (4 points)
Getting 2 years - Next highest payoff (3 points)
Getting 10 years - next highest (2 points)
Getting killed -- least favorite (1 point).

So what should happen?  Rational choice says each person will do that which gives the highest payoff, but the payoff depends on what the other person does.  Thus we examine the decision processes for each actor depending on what the other does.  Since the situation is the same for both players, let's look at it from Jill's point of view.  There are two possibilities: Jack will confess or Jack will stay silent.
 

• If Jill assumes that Jack will confess then it is in her interest to confess as well (she gets 10 years instead of a death sentence, a 2 point choice over a 1 point choice)
• If Jill assumes that Jack will stay silent then it is in her interest to confess, since by ratting him out she goes free (a 4 point choice over a 3 point choice).

Thus, regardless of the choice that Jack makes, it is in Jill's interest to confess.  Since Jack faces the exact same situation, he too will confess, and the expected outcome is that both players will confess, leading to a sub optimum outcome (they both get 10 years, each of their second-to-last favorite choice).

This type of situation occurs all the time is social life, especially when 'public goods' -- things that people share the benefits from but not the costs -- are involved (recall the tragedy of the commons example).  Many other types of games are possible (that, because of a different payoff matrix will result in cooperation or no dominant solution), and this field is rapidly expanding.