The way stones combine creates a line of thrust. As long as this stays within the wall it stands up.
As the forces are downwards all walls will almost inevitably settle to some degree, although this might be minuscule.
All the basic principles of building from setting stones flat, maintaining batter, running axis of stone into wall, tight face, good hearting etc., play some role in maintaining the previous 3 points, and maintaining equilibrium. Essentially good building reduces the potential of stones to move during settlement and keeps the forces in line so to speak.
A wall can settle principally in 2 ways.
-Sinking into the ground
-Movement of stones to close gaps. (Unless every stone fits perfectly with every other stone there in very least potential for movement)
Assuming a well built wall with stones sitting well on those below them, the greatest potential for settlement is the ground. The amount of settlement will depend on the weight of the wall and the relative softness of the ground compared to the wall (ie if the ground is as hard as the wall - bedrock- the wall can only settle in this way if the stone itself is compressed which if the wall was 3 or 4 metres high might be as much as 1mm, to be crushed under its own weight a wall would need to be well over a mile high).
Sitting a wall straight on turf will kill the turf and its root matter, this will decay and the wall will settle.
So ideally all walls should be sat on bedrock, but as with everything else in walling we have to compromise, so the answer as to how deep is until a suitable substrate is reached. There is no definitive answer as to what is a suitable substrate or depth.
Common sense and a lot of room for argument.
It does however mean be wary of made up ground unless it is highly mechanically compacted. I like the idea of setting walls on concrete, I`ve done it a couple of times. Heresy, but if the footings can`t move and you do a good job with the building little can go wrong,. Even if you do a bad job with the building there is much less that can go wrong than would otherwise be the case.
On another site (made up ground, clay) we used 15cm of MOT type 1 on Terram. Mechanically compacted. 7 years on there`s no obvious signs of movement. An original wall on the site dismantled and rebuilt (the footings had tipped) with resetting of footings and increased height has collapsed in 2 places. The footings were set directly on the `disturbed` clay....
Foundation trenches as drainage channels. Walls across slopes can `dam` water within footings. As soil becomes wetter it softens and there is more potential for settlement. Often the reason sections (in dips) across slopes collapse rather than due to any perceived effect of the slope itself.
If the footing are buried is this a trench/drainage channel. If it contains hearting and hence air it probably is. This could be an argument for filling all the internices with soil (not necessarily neglecting good hearting in the process) to prevent/reduce water penetration. If water gets in and stays in the soil will soften and the wall will settle, and settlement is the enemy of wallers! This is more likely to have a serious effect than any perceived wrong doing in the use of soil.
If stone `A` to the right of stone `B` settles sideways by a mm more than `B` then the wall to the right of `A` or below `A` might be tighter. However the wall to the left of `A` will be looser. This is `differential settlement` which is what ultimately causes collapses (other than those created by outside forces, cars, bulls, kamikaze sheep, ramblers). It seems likely that Mathematically in this instance the best we can hope for is a wall as tight overall whatever that means (haven`t worked out an exact mathematical proof yet). If it is as tight but some of it is tighter, some of it must necessarily be looser. If it is looser it is likely (perhaps not certainly- depending on interaction of all other factors of stability, some of which might have feasibly improved NEEDS THOUGHT) to be weaker, if a wall is only as strong as its weakest point it is now overall weaker.
Settlement out from the face (bulge) intuitively this must be looser, exactly how this works needs further thinking about. Is tightness somehow a function of force and volume? I.e. a wall has a certain force in newtons under its own weight, if its compressed into a smaller area it has the same force, but the construction must necessarily be somehow compressed, which given the relative durability/strength of stone must be due to less air, which is less gaps, which means tighter.
Does this mean that a wall that collapses in on itself is tighter? Probably depending on exactly how we define tightness. However in this instance its possibly irrelevant as to do this it must be poorly constructed. (It also has other implications with regard to lateral forces and interaction of face stones such as separation/reduction of contact). Essentially if a wall is built with no hearting it could conceivably tighten into a pile of stone on the ground! You can reach a point where the term `tightness` actually has no relevance.
Commonly A stone tipping, one settling more than the one next to it, creating a lateral force. Simplified when the lateral force exceeds force holding stone in place (vertical force and friction) it becomes displaced (or partially displaced).
What causes it.... Poor stone use/placement. Shape of stone. Differential ground conditions (ie soft/hard spots. - if we dig up a footing that has tipped how do we level the trench. Soil probably bad - can we compact it, hearting maybe.... always been against hearting in footings but this depends exactly how size affects settlement FURTHER THOUGHT REQUIRED.
Internal structure of a wall.... theoretically need to distribute everything equally to ensure whole wall settles in same way, if 2 adjacent bits of wall settle differently then potential collapse. Extreme example of piece of wall with lots of throughs alongside piece no throughs, bit without throughs collapses but at least partly due to fact that piece lots of throughs has barely moved.
Even if it is going to settle more how could you ever work out how much?
Masonry does not need to learn about stone to stone contact, the mortar IS stone and hence masonry has 100% stone contact. Hence the way the loads are distributed. Rather wallers need to understand why stone to stone contact is so important
Not glue, eg stones traced on edge can work in masonry wall not because they are glued in, rather because the forces are distributed in a predictable manner vertically. Hence lateral force is less of a concern. (Does not mean they are not a potential weakness compared to other stones in masonry once something goes wrong). (Food for thought here with the shaped polygonal walls of Delphi weren`t many of the stones here vertical. So tight contact holding them in place through friction or load distribution??mmmm.) Further reading Gordon "Structures"