The earliest soldiers were probably not a separate social class or caste from their brethren so much as a people practicing a different mode of tribal survival.
[more theory about early - 100,000 BC etc. - manís social, economical, and political arrangements - research what is known already]
It is quite likely that mankinds development (I am thinking of the period between our "evolution", ending with recognizable, fully human anthropoids, and our "history", beginning somewhere around the end of the last ice age) passed through many stages where the environment was a much greater threat to our survival than was the interaction with other groups of humans. A band of these humans might live together in a quite complex social group, using what they knew of agriculture, hunting, and tool use to pry a degree of material comfort and even security from the world around them.
The reliance on predictable yields of crops, use of shelter, and migratory patterns of game species would tend towards the existence of equally predictable sub-populations of humans who have succeeded in accumulating stores of foodstuffs, clothing materials, and tools. These sub-populations, whether an isolated tribe of 100 or a loosely interacting region of several thousands population would be a ripe target for any group that had discovered the rewards of aggression and plunder. The accumulation of several years agricultural and "industrial" output could be stolen with only a few days or weeks of "work." In time, perhaps, with appropriately developed techniques, these resources could be controlled and managed for the aggressors indefinite enjoyment by using the former owners as slaves.
Whether there is truly a genetic basis for this aggressive behavior I will leave unproved. I suspect there is a small genetic element, spread throughout almost all humans by now and perhaps always, and a fairly large degree of environmental feedback, encouraging this sort of activity. Once the trait makes its way into the gene and meme pools of humans, however, its general successes will tend to keep it there.
Whether in its early days there literally existed tribes with so little contact outside their locale that they were able to develop along pacifist lines and other tribes who mercilessly plundered the productive labors of others, or if the preference for violent solutions was early on scattered throughout the population, there would almost certainly exist groups with varying overall tendencies, some more peaceable, and some more violent, than others.
The physical separation of the groups that I imagine existed would reinforce both behaviors in isolation. As the groups come across each other, due to migration, decreasing food supply, or increasing population, their differing styles would not lead them to benefit mutually. The aggressive groups would tend to dominate the pacifist tribes and accumulate more of the available wealth, whether food stocks, access to better farm and hunting grounds, access to reproductive opportunity, and easily defended territory.
As human groups came in contact with one another more and more, necessitated some degrees of "political" specialization surely became useful, as leaders, productive workers and some form of defensive workers would begin to focus on one role. These divisions may also have been more apparent as time management, i.e., certain members would have farmed or hunted and when the need arose joined the ranks of the defensive task force.
[section on division of time-labor as useful outlet for otherwise inward-directed violent tendencies in general population - sending violent elements "outside" would be profitable, versus keeping them "home"]
Another way this mixture may have come about is the proactive attempt of the gentler tribes to avoid what might be viewed today as military dictatorships. It would quite behoove a passive, successful group of farmers, in order to lessen the destructive impact of raids by neighboring soldiering groups, to form a pact with one such group. This pact would integrate the soldiers with the farmers group, taking care of their material needs, perhaps even at a preferential rate, in exchange for their "protection". This protection racket would be first against the aggressive behavior of the specific soldier group in question, and second against the attempted raids of other groups.
While the aggressive behavior of the new members of the tribe, the soldiers, may pose some problems within the tribe - especially in times of peace - it could usually be turned outward against other tribes to yield both material profit and domestic tranquillity.
This arrangement seems to have survived the test of time and is with us still today. It is at best an uneasy truce, since the genes and memes are comingled. A certain small proportion of the population, which may be referred to as "pure soldiers", see the pragmatic use of force as the ideal route to material success. [section on business leaders as another manifestation of this phenomenon] At the opposite end of the spectrum, some small percentage are todays pacifists. The large group in the middle, while preferring to be left alone to their daily struggles and successes in making a living, also quite eagerly "rally around the flag" when aggressive behavior is engaged in by the leadership.
In recent centuries we see an attempt to get this system under control to an extent. [section on the internal conflict of business as a mode of aggression and its need of stability in social mores for success] In the United States, for example, the military is under the formal control of the civilian population. This is symbolized by the fact that the President, a civil servant and political leader, is the Commander in Chief of the armed forces. So theoretically at least, the aggressive elements of society, or at least their physical manifestation in the existence of huge military forces, must answer to a leader who serves at the will of the general population.
The precariousness of this arrangement becomes evident when we consider who is relied upon for information about other states potential behavior towards our own, the possible range of solutions when these are not in our best interests, and the probable outcomes of military engagements to reconcile these interests. To a large degree, the sources of these data are members of the professional soldier class, those who have made a career thriving under the assumptions and hierarchy of the military organisation, and have studied military successes and failures not with a view to reducing armed conflict but to increasing the probabilities of victory in such conflict.
These people see their route to personal fulfillment and professional advancement lying on the battlefield (typically as officers, not cannon fodder of course!), and are thus prejudiced in favor of these modes of conflict resolution. They may also tend to see the other states (the "enemy") in terms of their own value system, seeing aggressive intent where there is none, threat where none exists, and projecting their concept of "defense" into the entire sphere of global politics.
Keeping them at bay can be difficult, as they are able to muster the support of that great middle ground of the population who share the [genetic/memetic] tendency towards resolution of conflict and disagreement by violence or threat of it. They claim the territory when a conflict becomes difficult to manage by peaceful, diplomatic methods. The pacifist is listened to when all is well, but when the ugly face of war looms, she is not considered to be an "expert" any more and the populationís ear is turned to those who are - the soldiers.
© Huw Powell