I think the general unspeakable reason is pretty simple: once you give people power, some of them will misuse it. And since it takes more effort to correct a misuse of power than to commit the misuse in the first place, any institution that gives people power becomes more and more corrupt over time as misuses of power outweigh valid uses of power.
That said, it's pretty clear that there are more than a few Wikipedia admins who seem to have embraced deletionism for topics that aren't near and dear to them personally and/or which poke at whatever their particular political hot buttons are.
There is not a motivating need to limit WP's scope, and indeed that is why WMF forked off projects like Wiktionary and Wikidata to their own TLDs. The main problem with WP is that it is far easier to be wrong than to be right, and the effort required to be right is linear in the number of words written. In short, the number of editors per page required for acceptable quality does not roll off with large numbers of pages, but stays relatively high, at around 1-2 editors/page.
Worse, the number of moderators per editor does not roll off either. The number of bureaucrats required therefore keeps growing, logarithmically but steadily, and the demands on arbitration committees keep growing. The committees themselves have long ago failed basic principles of legal legibility, leading to sprawling bureaucracy.
One possible solution is to fundamentally alter what we store. Rather than writing thousands of words of prose, we could use Wikidata to automatically generate articles. We already have factboxes which could be largely automatically populated, and many people only care about the factboxes. Prose would be limited to commentary and explication, but would not be the main bodies of articles. This is not just a pipe dream; LMFDB [0] exists and is worth examining as an example of how code and data can automatically generate the bulk of an encyclopedia.
But, let's be honest, the writing was on the wall when Esperanza [1] was dissolved. We are now somewhere between Bureaucracy and The Aftermath.
This is in fact being proposed at https://meta.wikimedia.org/wiki/Wikimedia_Forum#Proposal_tow... by a prominent Wikidatan. (Edit: follow up at https://meta.wikimedia.org/wiki/Wikilambda and https://meta.wikimedia.org/wiki/Talk:Wikilambda .) However generating sensible articles would require expanding the current Wikidata model, and this is something that should happen gradually and be managed by the WD community itself, not at a separate project. The whole pretty-printing-in-natural-language part is the most speculative by far, and incubating it separately makes more sense.
It's worth noting that Wikidata itself is not "deletionist" other than as implied by verifiability- and sourcing-requirements. Its model is far more general and far more "inclusionist" than even the most permissive visions for Wikipedia.
This feels like a gross misreading of the facts.
Wiktionary is separate from Wikipedia because its work product is fundamentally different -- it's a dictionary, not an encylopedia. Wikipedia has a lot of articles about things that aren't words, and Wiktionary has a lot of pages for words that wouldn't make sense to write an encyclopedia article about.
Wikidata, meanwhile, is only about raw data. Which is a part of an encyclopedia, but far from all of it. (How would you write an article about the history of Rome using only data?)
> But, let's be honest, the writing was on the wall when Esperanza [1] was dissolved.
This, too, is a gross misstatement of the facts.
Esperanza was dissolved because it was becoming a cabal. It started as a social group, but rapidly grew into its own organization, with its own decision-making processes and elected officials separate from those of Wikipedia. There was widespread agreement, even from Esperanza's founder, that the organization was no longer fulfilling its purpose, and an effort to reform it before it was shut down.
There's a decent summary at: https://en.wikipedia.org/wiki/Wikipedia:Wikipedia_Signpost/2...
Because human beings don't see reality accurately, see the science:
There are a number of issues that seem to be eternal. Even if there's an obvious right answer, I see year after year, decade after decade that they keep being discussed, with the same points being made over and over again. Is there a name for these?
I guess that for each of them, there is some kind of unspeakable reason that trumps any sound rationale.