Stanford Prison Experiments
The Stanford Prison experiments show the enormous power of social context on the way people behave.
In the experiment, randomly selected people, who were screened for psychological issues, were either prison guards or prisoners.
Simply giving people that distinction, and dressing them in either "guard" or "prisoner, cause them to adopt a completely different persona of either extreme sadism or of helplessness.
The connection here is that, if you put people in a situation where the context would suggest a different behaviour than the one they normally experience, people quickly fill that role, even if it is a role of abuse. When riots start, or when there is a breakdown of social order, people will act in ways that are entirely contrary to their usual behaviour.
Ashe Conformity Experiments
In these experiments, Ashe had one subject in a room with 9 people who were only acting like subjects. They were all shown 3 lines of clearly different length, and asked to identify which was the longest. Importantly, the subject was told this was part of an eye test, so their ability to see the longest line was actually reflective of their perception.
In the experiment, the fake subjects all said the shortest line was longest, and iirc, this caused a huge majority of people to also say the short line was longest. There are two relevant things here. The first is the control of individuals that group dynamics have. Even though the long line was clearly longer, people would give answers they knew to be false in order to seem like part of the group. More interestingly, the subject thought the study was about perceptual abilities. In this regard, not only did they give a wrong answer, they gave a wrong answer that reflected poorly on them. They were willing to let the experimenter think they had poor eyesight so that they gave the same answer as everyone else. Further, this wasn't a group that identified with poor eyesight, or really anything. What I mean is that, the group the subjects conformed to wasn't an actual group in any way, there would be no sense of loyalty or friendship to persuade this person to give the wrong answer. It is as though the subject would give knowingly false answers that reflected poorly upon themselves for no other reason than to belong to a group of strangers who aren't actually affiliated in any way.
The obvious parallel to riots from this is that, as individuals, we have an extreme bias toward acting and thinking as others do. When social order starts to break down and the context of behaviour changes, it spreads very quickly by people feeling the need to participate in something that looks like a group.
diffusion of responsibility
numerous studies have shown that, the larger a group is, the less likely any individual in that group is going to take responsibility for actions. So like, if you witness an assault, and you are alone, you are far more likely to do something to stop it, than if you are in a large group of people. The theory is that, because there are more people there who could do something about it, you feel less pressure to be the one who does.
The connection here is obvious, as if there are numerous people rioting, it is less of your responsibility to do something about it
crowd modelling dynamics
while it is true that modeling the behaviour of individuals is extremely difficult, modeling crowds is not. This is because, as groups, people respond more to how the group behaves than how they would as individuals (see above).
In the case of riots, this indicates that people aren't acting as individual "political" actors, but rather as part of a group, their behaviour defined largely by the local contextual dynamics.
Group Think
Because of all of the above reasons, it is very difficult for dissenting views to appear in a group. Thus, once riots have started, it is almost certain that they aren't going to stop because someone thinks it is bad.
Anonymity and Responsibility
Another part of acting as a group is the loss of identity. People feel less of an individual when they are part of a group. When you feel this anonymity, you are able to act in ways that you normally would not, because of social repercussions. The internet is a perfect example of this.
So, being part of a group that is rioting makes one feel as though they aren't responsible for the actions they are performing.
other stuff
most action is not motivated by "top-down" sorts of "I feel this way, so I will do this thing". Rather, it is motivated by "bottom-up" sorts of logic, "I did this thing, so I must have a reason for it". This comes from tones of evidence, most notably that of work done on split brain paitents by Gazanaga. Further, when you look at the work of Libet, you find that our motivations for behaviour come after we are already ready to perform the action, meaning that, consciously, we aren't in control, but rather, our conscious awareness reflects what our subconscious has already prepared to do, not the other way around.
thus, these riots almost certainly reflect group dynamics and psychology in a situation of decontextualized social order where people feel less responsibility to act in a socially constructive manner, and, as they get caught up into it, basically feed off of general social psychological mechanisms. Any political aspect to it would be on the part of the observer, trying to define people's behaviour as "this" or "that", rather than trying to understand the underlying behaviour of people in general. If you want, we can extend this to other riots, like those in LA or those after the death of Indra Ghandi, and we will see that these very mundane aspects of human psychology are a far superior explanation for riots than is any appeal to class struggle or unifying political analysis.
[sociology is such a joke]