There's a little something I'm working on for my own pleasure (erm, IYSWIM) which I might spill a few details of for demonstrative purposes for an alternate POV... I suppose (thinking about it, right now) it's a bit Rock, Paper, Scissors, which seems to suggest that it works as a concept. (Oh good, I had been rather worried that it wouldn't...)
Civilisations/entities are defined by three main compass-points, being warlike, expansionist and/or secretive. A warrior race has an advantage over an expansionist one (at least at first meeting), the expansionist race out-competes the secretive and insular type, while this third group can inveigle themselves under the radar of the warlike species. (However, there are also blends of qualities, but it does make them less effective in a shared quality than a 'pure' one, and the outcome would depend on the balances of each quality, and the nature of the encounter.)
Who's Good and Bad? Well, any civilisations themselves are 'Good' (sections of the society may tend towards somewhere else on the continuity, of course) and 'Bad' is the power-type that considers this society to be a walk-over. But the walk-overable power-type... Well, probably not even worth a passing consideration (despite the reciprical concerns).
So, how might we put that in DF terms? Well, current evil zones are ones that should just walk over standard dwarves that settle. In various ways, players have devised methods to fight (or shield themselves from) this onslaught. Doesn't tend to stop the
natural attacks from happening. Meanwhile, Neutral/Good zones (as defined by DF) are equally "my place", for the player dwarves, even though no natural home form them is in DF's Good-defined areas.
But what are the Elves thinking? Dwarfs digging up their land, felling their trees! They'd probably find ('Neutral') bare-rock mountainside inhospitable in the extreme, but might they have some kind of "turn undead" ability for the zombie wildlife? And let's say for argument that Goblin civs would find the natural flora and fauna of the so-called-'Evil' areas to be much like a nature reserve and (normal player defences notwithstanding) Neutral mountains just a bit less full of undead stuff, but imagine their problems when trying to live among Unicorn herds!
It's not an exact equivalence (I suppose that taking 'Evil' as 'Warlike', that would make Neutral as Expansionist and Good as the third aspect, but I can see alternate mappings, too), and I'm taking definite liberties in that last paragraph, but perhaps imagine that sphere-alignment is like RPS (RPSLizardSpock, or even like
RPS101) so that for any two spheres, one is dominant so that a (to take an RPS101 element) a Castle-type sphere race is good to go into a Rain-sphere area (while a Rain-type race has a disadvantage in a Castle-sphere area) due to the "Castle providing shelter from rain", while anything Water-based (Water, itself being dominated over its neighbouring Rain item) has a natural ability to overcome a Castle-based element, in whatever capacity, due to a "Water floods/erodes Castle" relationship.
Though given the contrivance of some of the RPS101 dominance/subservience reasons, I'm not entirely sure that there'd be an easy route towards arranging DF's own compliment of spheres into perfectly balanced order around the circle-of-competition. OTOH, if it's taken merely as an
edge of one over another (and, as we know, dwarves can populate Hell, never mind significantly subdue an Evil biome) it needn't even be
perfectly balanced, anyway. And in the case of multi-sphered alignments there might be positive
and negative modifiers for any particular entity or region interacting with another entity or region.
I was replying here immediately after (though not necessarily
in reference to) GreatWyrm. But, in the usual nature of such synchronicity, I can see points made by Dorfimedes and Manveru Taurener, in-betweentimes, that meld to what I'm saying. Especially
the idea of gathering together the known spheres. Anyhow...