First, I need to restate the rule to give it nuance. Original research isn't allowed on Wikipedia. Primary sources are acceptable for simple statements of fact, but editors can't make analysis of or synthesise what is contained in that material. That analysis or synthesis requires reliable secondary sources.
The epistemological basis for the "no original research" rule seems to be because Wikipedia is not a suitable vehicle for synthesis or analysis: it is written by anonymous editors with varying degrees of expertise in the subject area. They are not assumed to be competent to conduct such analysis. They are only assumed competent to find secondary sources that make such analysis and cite them.
The rule also has some pragmatism about it. If such analysis was allowed, it would have to be performed on Wikipedia with its reasoning clearly articulated and developed. Wikipedia is an encyclopedia, it shouldn't be cluttered by such reasoning.
The appropriate place for analysis and synthesis of this kind are peer reviewed journal articles or other secondary sources where the author's expertise or authority to speak to a point, and their reasoning, is made plain. A reader should, at least in theory, be able to check the source and verify as such.
It is also not seen as acceptable for editors, who lack accountability for what they post, to offer such analysis. A news report is expected to have some degree of accountability, to put the relevant statement into context and to consider competing views. A Linux kernel mailing list post does not do this.
That said, in my experience, a blind eye is turned to uncontroversial analysis.
Many of those things are the Wikipedia equivalent of legal fictions. They may not be true in any meaningful sense. News reports have very little accountability outside of cases where someone can sue the newspaper, and the "analysis" that people are assumed to be incompetent at and the "finding secondary sources" they are assumed to be competent actually require fairly similar levels of competency.
How is it a fiction that news reports come from identified individuals, while Wikipedia is edited by anonymous individuals with questionable credentials?
I'm not convinced of the characterisation as a legal fiction, either. There is a logical justification for this position, as stated, in detail, in my reply. A freely editable encyclopedia isn't the place for contentious analysis.
Where it fails is that media lies are not considered primary sources for this stuff, but instead secondary... When their follow the political spin editors want to present...
