Jimbo, could you please comment on the following attempt to understand your point?
We often need to know the relative importance of information in the context of an article. An important use is to e.g. establish attribution (who hold this POV; are they the majority, a minority, a tiny minority), or relevance (is it relevant to many people; is this between the subject and a limited number of others; how widely has it been reported/interpreted/analyzed?) To discover this, we will virtually always need reliable, secondary sources: the mere *existence* of a primary source whose actual number of visitors is unknown (such as court records) does not tell us anything about its importance, while both the existence of e.g. books or mainstream media reports on the information and the interpretations/analysis/etc they provide, may tell us what we need to know.
Information without sources that provide a context, constitutes original research - not to the degree that it is untrue, but to the degree that we do not *know* how (un)important it is and are relying on our personal insights instead. We should err on the side of caution by not including it in the encyclopedia.
Thanks,
AvB