Based on the concept we are considering regarding the inhibitor of intraspecific violence (the Lorenz mechanism (LM), as we call it in view of the corresponding theory of Konrad Lorenz, as well as the violence inhibition mechanism (VIM) according to James Blair’s model), we find that a human’s ability to commit violence against another human is a deviation consisting of a malfunction of the innate LM/VIM. Thus, violent offenders are people suffering from a genetic and neurophysiological pathology [1]. This concept can be countered by the argument that the level of violence in human society, from the very emergence of humans to the present day, has decreased at rates completely incomparable to the speed at which traits are fixed in a population as a result of natural biological evolution. Additionally, different levels of violence are observed under different social orders. I will not dispute the validity of such assertions; however, by providing two arguments, I will prove that they do not refute the LM/VIM concept.
The first thing to understand is that a person with a weak or dysfunctional version of the LM/VIM does not necessarily have to commit violence. A good upbringing, the absence of circumstances that provoke violence, and the presence of circumstances that deter it, while not eliminating the risk completely, reduce it to a certain extent. Combine this with the fact that violence is committed by a minority of people. Military evidence shows that only up to 2% of people experience no resistance to killing (and war is a highly provoking circumstance for violence). An analysis of crime in the Russian Federation determined that in the worst possible scenario (and in reality, it may be better), on average, only one new violent offender committing a violent crime no less severe than battery is reported per 695 people per year.
If environmental circumstances change such that, for example, instead of one person per 695 people, two people begin to commit violence per year, then the level of violence in society will increase by at least twofold, and most likely much more, since those who have previously committed violence may begin to do so more frequently. The consequences would be significant. However, violence would still originate from an overwhelming minority of people, so observing such changes does not refute the LM/VIM concept.
Let us now turn to the evolutionary argument. Under natural conditions, biological evolution is indeed a very slow process. However, it can be accelerated through artificial selection. Cultural influence can also lead to such selection, as it artificially alters the social and physical environment in which genetic selection occurs. This is called dual inheritance or gene-culture coevolution.
A classic example is the development of lactose tolerance, which resulted from the development of dairy farming that occurred only approximately 5,000 years ago. Another example: between 800 and 1700, the culture of the Ashkenazi Jewish ethnic group practiced limiting employment to professions requiring a high level of intelligence, which created selective pressure in favor of people possessing it. Combined with the genetic isolation of the ethnic group, this presumably led to high IQ scores, as well as a number of hereditary genetic diseases among its representatives. Furthermore, as a result of territorial expansion in the second half of the last millennium, mixed races appeared. And even the consequences of urbanization and industrialization, which began 200-300 years ago, are already leading to measurable genetic changes [2].
Similarly, cultural influence can lead to the strengthening of the LM/VIM. The growing cultural unacceptability of violence creates selective pressure in favor of less violent people, while violent offenders are filtered out more and more. Culture does indeed influence the level of violence in society, but this does not contradict, and may even fit into, the LM/VIM concept.

Sources: