Most of us may have known the fact that Iranian people are unpredictable. What I will talk about here is not unpredictability in general but unpredictability in favoring political leaders. Although everyone may be fully familiar with this issue in Iran but I want to clarify it with more examples. We all remember days in which Khomeini was being praised extravagantly, days when no one could dare to criticize either him or anyone (anything) related to the Islamic regime. Also those days that Khamenei was elected twice as an elected President of Iran. The same thing repeated several years later, for the new president, Rafsanjani. It seems that this scenario is being replicated again for Khatami. The aim of these few lines is not to blame anyone or any idea but to try to explain only a possible reason for this phenomenon.
I want to use a statistical realm called Bayesian statistic to shed light on some hidden angles of this behavior. Bayesian statistics is all about updating old beliefs when new facts are observed. It assumes that for each variable (unknown or what is not completely defined) some prior belief can be assigned (Prior Knowledge) which then can be updated to a new belief (Posterior Knowledge) after observing some facts related to the unknown variable. The unknown variable in this case can be the personality: e.g. evil-ness or angel-ness of a political leader. Assume that everyone in our mind has a place on the axis of evil-angel. The axis has two extremes: complete evil and complete angel, as shown in the figure below.
We are not usually sure about the value of things, so we can express our uncertainty as a distribution function (a little knowledge of statistic is needed here). The surer we are about a value, the tighter its distribution function that means we do not allow the value change freely in its domain.
Now consider we have a prior knowledge about a leaderís personality which is placed in our evil-angel axis in our mind as bellow:
In its best case, this prior knowledge usually comes from the leader's history, but in the societies that people stereotype other people, it can have other origins such as whether the leader is "Mullah" or "Sayyed" or whether s/he smiles, s/he is beautiful, has a calm and kind face, or other stories/rumors. When the leader comes to power, people update their knowledge about her/him. Let's assume that the leader does a number of good and bad things during her/his time in power, as shown by the points on the evil-angel axis in the figure below. These points represent degrees of evil-ness and angel-ness, which can be related to their performer (the leader).
These new pieces of data if used fully to update the prior knowledge about the leader would result in the new curve plotted next to prior knowledge curve. However, what happens in reality is that people usually stick to their prior belief and try to keep it unchanged. The only way to do that and maintain consistency in their mind is to down the weight of the points that do not conform to our prior knowledge and at the same time give more weight to the points that support their prior beliefs. Blaming other people for what the leader has done or assuming that the leader has acted under duress are but a few examples of sticking to prior belief. These modifications will result in a new graph which is depicted here:
This process happens every time a new fact is observed, however, the more facts contrary to the prior belief are observed, the more difficult it is to modify the weights and to keep the same prior beliefs. At one point this process will get reversed, which means that the level of inconsistency between the leaderís functions and our prior belief is raised to a point which can not be remedied by the down-weighting process anymore. At this point people are mentally exhausted because they haven't been successful in justifying the leader's actions. Tired of this justification process they finally eliminate their prior knowledge.
In this situation what would be the new belief about the personality of the leader? Let's look at the observed facts so far on the evil-angel axis (see figure below). You may see some shifting in the leader actions from the right to the left side of the axis which means a drift to the evil side. This can be because the people have not given appropriate feedback to the leader and the leader gradually has deteriorated.
The above curve represents the true representation of the observed behaviors, without any perception and processing bias. However dealing with a personality in the gray zone needs much more effort. According to a theory, the degree of information transmission is a function of one over uncertainty. Therefore the more uncertain we are about a system, the more difficult it is to work with that system. A personality in the gray zone as is the case in the above graph cannot be modeled as easily as a personality in the black or white zone. Keeping in mind that the people are already tired and may feel belittled because of justifying their leader's actions for many years and being betrayed in return, one can predict that people would not spend any more energy for saving their leader face. After eliminating previous belief about the leader the new belief will be made as shown below, using some modifications in the weight of each deed to outweigh the negative sides.
In its new form, the curve is much easier for interpretations and dealing with uncertainty. Every thing is again predictable and easy to judge. No need to bother each other's brains by trying to find out the facts behind each face or by attempting to find the best way to give feedback to the leaders. This is the end for the current leader, but who is going to be the next?
[P.S.] The word "leader" can be replaced with some other things such as friend, an idea, a religion etc.