Turjan said:
I see this first and foremost as a game. That's why I usually avoid situations that pose very tricky moral questions. This means that I construct principally two situations:
1. If I present a full enemy society, like a functioning orc cave complex, that doesn't only contain combatants but also elderly, children and other noncombatants, those will most probably surrender. A proper capitulation is something PCs can easily deal with. At least my players can

.
Note that the assumed "full enemy society" assumes that orcs are basically humans with fewer feats and different stat-mods (and different color skin). It's quite conceivable that someone could conceive of a full enemy society that does not present the elderly and where children are not really non-combatants. Herriman the Wise's storyhour, the Happenings of Lucifus Cray, Warlock, features an orc society where, at least elderly orcs unlikely to be able to fend for themselves would not be present and where orc children are unlikely to be strictly non-combatant. Likewise, Nemmerle's storyhour, Out of the Frying Pan featured a trip through an orc cave complex where the children were not presented as non-combatant.
Heck, even real life anthropology offers a few examples of societies without either group. Some groups of inuit, for instance would be without non-combatant elderly if adventurers began attacking them, and the Ik described in Collin Turnbull's the Mountain People had children who would not be non-combatant. Now, it's true that the Ik at the time of the Mountain people could not be described as a functional society, but there are two things to consider:
1. Creatures who are "usually" chaotic evil may have a different definition of functional than we do. (It may be that this different definition accounts for why orcs are often described as primitive and/or bestial).
2. All societies in real life are not functional by that definition. You don't need a functional society in order to loot and pillage. And, anyway, a lot of the most interesting things from a game perspective happen when societies do break down.
(For that matter, the assumption that women were non-combatants in ancient societies needs some revision in the case of sieges as it was fairly common for them to join the men in pitching anything heavy over the walls at their attackers. If I recall my Old Testament/Tanakh correctly, one of the ancient near-eastern leaders was killed when a woman tossed a millstone down on his group of soldiers assaulting the gate of a fortified town).