I'll take a shot at this.
I have to admit, at the outset, that I'm an electrical engineer, not mechanical, but I'm accustomed to working outside the electrical area. So, some of you guys with specialized knowledge (like John, in particular) may want to comment on this. I'm fine with that.
Anyway, I recently undertook a study to try to understand why bolts loosen. It's all pretty interesting, especially for a geek like me who is easily amused. (My electronics specialties are noise and distortion; please try to contain your excitement.) Most of our understanding of bolt loosening comes from a classic 1968 paper by a guy named Junker, who showed that cyclic lateral forces were the culprit, and designed a test (now called a Junker test) to evaluate it. The video referenced above gives an example; you can find more on line.
Lateral vibration loosens the bolt, initially perhaps only microscopically, and that reduces the stress in it. Then, with less stress, the vibration moves it more easily, reducing the axial stress further, and so on until it loosens completely. This process can be prevented a number of ways, including increasing the elasticity of the joint (e.g., with a spring washer or stretch bolt). Then, as the bolt begins to loosen, there is less decrease in the axial stress, so the loosening process is impeded. Not necessarily completely prevented, though; it depends a lot on the vibration environment, the type of locking device, and probably other factors. Of course, there are other tricks, like nyloc nuts, which increase the thread friction independent of stress, and actually look pretty good in Junker tests. Belleville washers, which combine a serrated surface with a domed shape, thus a spring effect, also seem pretty good.
I found results of several Junker tests for split washers. They showed a minor improvement over nothing, at best, and in some tests, they even looked worse. Unfortunately, I found only raw tests, with no interpretation or underlying analysis, so it's hard to say what the difference was. I suspect it depends on the design of the washer and the material of the bolt and substrate that it's threaded into.
Another point: Junker's work post-dates the design of our cars, so the understanding of bolt locking, when the cars were designed, was not as good as it is today. For that reason we probably should not view the locking devices used in the original designs as the only, or even best, way to go.