Alright, I couldn't help but weight in to the revived topic.
I assume we're talking about levelling fill for a structure. Large rocks can punch holes in infrastructure, though I almost always see either pea gravel around pipes, and for slabs or pavement, it's manufactured fill.
I see two ways to look at this from a testing perspective. According to ASTM D1557 Modified Proctor, a sample containing 0-5% over the 3/4" sieve does not include that oversize in the testing. That's right- it's just not accounted for. Knowing this, as an inspector, when I take the proctor report into the field, I look for compaction on the high side of 95%, as opposed to, say, 94.5% for low-oversize content structural fill. First, the field inspector has a certain level of competence when testing for compaction beyond what the gauge reads. If I'm testing bank run, and I get 4 tests at 91% and one test at 99%, I'm going to assume I've just tested over a large rock, as it is just plain obvious. The more large rocks you have in your fill, the harder it will be to make this call, and if you have a lot, it is impossible. Secondly, I think this rule is just like all other rules; to prevent the extreme cases of large rock content from going uncontrolled. It is a case where you have to make a judgement call. Though I think here, 3-4" rock may be a de minimus problem compared with others you might experience during the course of a project.
Hope this helped.