The answer to what will either lead a man kill or not to kill will often depend on the background of the person giving the answer. Psychologist, sociologists and political scientists will tend to give different answers based on their point of view (the influence of the individual, society or power).
I came across several first hand accounts of of fighting during the First World War as part of a a paper for a course some years back that reveal the complexity of the issue. In one instance, a British officer is at an OP making a sketch of the German positions. He can observe a German soldier carrying out personnal hygiene and points him out to the sentry. The sentry remarks that he had been watching him but had not shot. It could be argued that the soldier had not shot out of keeping him position secure, but the answen given by the sentry was that the enemy was "just a lousy German." By lousy he meant that the German was lice infested and was going through the routine of killing the little critters. He was not a threat and he was also identifiable as a fellow soldier living in terrible conditions. Would we make the same decision today?
I came across some other accounts of British soldiers killing Germans with spades during trench fighting including some who were trying to surrender. This tells me that the circumstances of the situation have a lot to do with the decison to kill or not kill (no great discovery but worth pointing out). Did operant conditioning lead the soldiers to kill enemy soldiers or was something else at work?
Both situations above show radically different behaviours by soldiers who came from the same society and training system (a sweeping generalization perhaps). Neither outcome seemed to depend on authority. Distance seemed to have an inverse effect from the one often given that increased range from the opponent will increase the likelihood of "killing." The threat posed by the enemy is perhaps a factor, although I need to be careful here. In the first case the German soldier poses no threat to the sentry. The Germans killed trying to surrender in the trench did not pose a threat either, although there are cases of people trying to surrender turning on their captors. What is significant, though, is that these killings happened during a hand to hand melee. The attackers had little choice but to kill once they entered the trench system and encountered resistance. Once they started killing or being killed it was hard to automatically turn this off (which might explain the slaughters during routs throughout history).
Trying to explain behaviour as either "natural" or "societal" can lead to some long but often fruitless nature vs nurture debates. I think that it is very hard to determine what natural behaviour is since we are social beings. Perhaps Infanteer is on to something as "society" definately does not want people killing each other but an individual can certainly benefit (sometimes simply by surviving). Society and authority can have an influence one way or another. Societal influence will tend to reduce the probability of killing (but that may be changing). Training, group pressures and authority can increase or decrease the probability, but I think that ultimately the situation will dictate. I think that the issue will defy neat solutions.
Cheers,
2B