My inquiry is a desire to understand why something works.. or doesn't.
Spokes work when they're at sufficiently high tension without residual stress.
Spokes fail due to fatigue in too few cycles because of high average stress (not all of the elbow was taken past its elastic limit when the spoke was formed) solved by stress relieving and/or high variation (as from spokes going slack and bending back and forth like you'd do to break a paper clip when they have too little tension) which you avoid by keeping everything uniformly tight with corrected spoke lines.
Thinner spokes will keep the rim supported horizontally over bigger bumps than thicker ones because they have more stretch at a given tension so the rim needs to bend farther before they go slack.
1.8mm elbows are tough enough - Jobst Brandt has over 200,000 miles on a set of 1.8/1.6mm spokes ridden all over including dirt roads and cobblestones and he's not a small guy.
However, with large hub holes (to make insertion easy) and/or not enough of an obtuse angle on the end of the spoke you get high stress where the spoke bends until it's supported and leaves stress which cannot be removed thus leading to premature spoke failures on that wheel until larger spokes are used.
Buying just a single 1.8/1.6 spoke is also hard - few shops stock them because QBP doesn't sell them.
2.0/1.5 spokes work too, but some people have issues with windup and tension + too much torsional stress can combine to break them when building. Torsional rigidity is 23% less than a 1.8/1.6 gauge spoke 52% less than a 2.0/1.8. I put tape flags on representative spokes (drive side, non-drive-side) to see how much wind-up I need to undo and haven't had problems when I lubricate the threads with anti-seize even when re-using 12-14 year old alloy nipples.
1.8/1.5 have less windup to deal with because the smaller thread diameter imparts less torque on the spoke when it's being tensioned.