Mimesis means imitation, repetition. Practical paradigms are usually practiced mimetically, they are a matter of imitation. 

Fractals emerge through recurrence of earlier sequences and positions. A new fractal performs, through recursion, a sort of mimesis of previous patterns, but not, because imitation of previous patterns is a purpose, but because the schematized procedure is caused in varying dimensions and under the condition of further dimensions, through a formulated and schematized procedure of a higher, more abstract order.

Actually, the same is true for paradigmatization, which has been described by Thomas Samuel Kuhn. We imitate the behaviour of fellow humans, not because we are concerned about the particular behaviour, but because we desire to profit from successful routines, and because imitation is usually understood to be utile. The commonness of mimetical utility is normally igored, the sciences are interested more in the empirically provable original pattern, which is the origin of all imitations, the primary practice, as a theory.

Absurd enough, a scientist, who would i.e. honestly state, that he is making use of a methodically ideal means of production or experimenting, only because imitation is generally sensible, incredible. This scientist would be considered blind towards the immediate sense and utility of a genial approach in particular.

Criminals have it much easier in this regard. The psychically ill and perverted criminal, whose crimes appear like a devilish aping of prior crimes to the surving humanity, has fewer problems to admit, that the principal imitation is stimulating to him or her. The less, because paradigmatization of what is considered a tradition or exemplary practice always appears to be morally justifying. The acting under the influence of an emotional affect is one refuge of criminals at courts because of this.

The three concepts mimesis, fractal, and recursion lie close together. Recursion means the return to a previous pattern or step, or point. When certain antisocial concepts, like antisemitism, motivate humans across more than 1.000 years, then surely because so many people during the course of their more or less conscious deliberations, where to begin with such exuberant criminal energy, to realize the sacrificial suggestion concretely, always return to the same paradigmatic, here: antisemitic points.
 
The fractal emerges, because in the course of recursion, and mimetical procedures, necessarily a change is happening, that moves the whole construct, which can be called a tradition too, to another place, and into another dimension at the same time, which can appear as a simple scaling of previous patterns on the same plane of dimensions. The additional scalar dimension appears as being shrouded within the former ones.

On a greater plane, like i.e. that of a nation, a paradigmatical crime and pattern not only becomes obvious in more lucid detail, but it gains additional structures, which are caused by dimensions of size, and mechanisms, which are inherent to mimesis or a certain formular. 

Antisemitic crime and violence did not only happen until 1945, and then not any longer, but both continued steadily after the Holocaust, only in much smaller dimension, and at many, single, seemingly disjointed places, through many perpetrators, not connected to each other personally, or by association.

Jurisprudence, penal justice, and criminology are sceptical regarding theories of practical paradigms, and mimetical behaviours, like theories dealing with a contemporary spirit, or zeitgeist, implying that certain patterns of behaviour are caused by it, because by their adaptation, a relativization of the guilt of criminals could easily be brought about.

It is difficult, not to understand the fractal pattern in human behaviours, and their repetition, the return to certain approximation- or starting points, like the apish nature of human who rather apes one time too often, then one time to seldom, but also imitates to learn practically, as a dynamic underlying coercion of some sort, although this coercion need not be.

Humans are no computers, and what is inescapable for a machine, like the repetition of a calculation and procedural sequence, that is escapable for humans. Ultimate responsibility for own actions and mimetical logic, like in a vendetta, and in the ius talionis, do not exclude each other, actually.

The mirror penalties were not exerted, because the kings, punishing by mirroring, could not have acted differently, than by imiation. They did not suffer from some weird contradictory condition, like the rationalizing, and at the same time affective coercion to repeat another persons crime on the perpetrator. 

Also, the scientifical paradigm is being chosen by researchers and practitioners in application of free volition, normally, but not necessarily because it is evidently the best method, to approach a matter, or to produce a certain product. 
 

Examples for fractally recurring violence, of collective proportions are quite plausibly :

 * First and Second World War
 * Russian and german persecution of Jews and their annihilation, from 1900 to 1945
 * Bosnian War and putinist aggression
 * The War in Afghanistan until 1989, and the Djihad of Osama Bin Laden since 1996
 * The War in Korea from 1950 to 1953, and the Vietnam War from 1955 to 1975
 * The imperial aggression of China, especially in Tibet, and East Turkestan

While the above examples are phenomena of aggression of a very big, and more or less unite dimension, so were these collective aggressions and wars preceded by single, and only seemingly isolated, or altogether discontinuous deeds of violence. The enmity between the Han Chinese and the Uyghurs, i.e., is older than the USA. That means, that nationally or racistically determinable conflicts, like the history of slavery in and from Africa, the hatred towards Jews in Europe, and the struggle between russian-slavic, and german peoples, across eurasia, which all three lasted more than a 1.000 years, simply can not be reduced to single events.

Another example is the slowly, but statistically evidently decreasing criminality of repeat offenders on along a gradient of age. Young repeat offenders are the most criminal, old ones tend to repeat their personally typical crimes less often.

Another example is the sociological concept of the socially focal, or combustion point. At certain geographically determinable places, aggressions and crime are much more prevalent, than at others, which are more remote, and there are, similar to fractal patters, traceable on maps, certain centers of criminal or political aggression, and, by adding a coordinate of temporal progression, also describable in formula. But because of this, crime far away from these centers is not absent, it only recurs on a smaller scale, often according to the same, culturally distributed paradigms, like the typical "drive-by-shootings" across all of America, or the criminality erupting at block parties in the USA in particular.  

But the concept of "context" is, although correct as such, unnessecarily lump, like the more plastical, but simplifying comparison with natural catastrophes, as in the differentiation between pre-, main, or post-earth quake events. Human aggressivity and aggression is more complex and more typical, and it has social and political causes, and cultural conditions.  

The probability of the recurrence and the dimension of size and complexity, and even the detailed form of fractal patterns, lies within mathematic formulas. By this, fractals can be predicted, which means, that fractals are lacking the peculiar property of being unpredictable, which emerges through human volitions and purposes. 

To this degree, fractals are not applicable in their purest sense to describe the aggressive behaviour of humans, but the addition of chaotic or random conditions as fractal parameters would be rather unproblematic, that means, it is no actual deficiency of the mathematical theory, that there are no fractal models yet, to describe human aggression. 

Especially the projects to create artificial reality and world simulations, by using supercomputers will presumably, sooner or later, be conducted by making use of fractal formulas, to model virtualized phenomena of collective aggressivity according to a variety of types.

For psychology and neurosciences, a theory of fractally shaped human aggression would at least be a concrete condition, according to which the fairishly elaborated, and inductively only limitedly elaboratable, and much less interpersonally and sociologically generalizable concept of "disposition" could receive a theoretically more adequate form.

Such a formula would, in the end, be a special case, of a more general theory, of the fractally shaped human volition, to approximate the changing of stock values at stock markets, or the change in the behaviour of consumers, or citizens participating in political polls.

Benoit Mandelbrot and many other scientists remarked the misbehaviour of stock market dynamics, because they do not seem to fit any pattern, not even of a fractal kind. Anyways, if human aggression does, then the seeming patternlessness of the more chaotic stock market dynamics must come about only by a specific ignorance regarding the detailed conditions, which lead to a deviation from patterns.

If the effects of collective conditions, like boom and inflation, have shaping impacts on stock market dynamics, than this does of course hint at the common frailty of stock market dynamics to certain, powerful events. It must be the differentiated multitude of more or less collective, and more or less circumstantial events, which leads to the stock market behaviour.