You are 100% correct: it is computer generated.

One can usually tell, and if you want a breakdown of how, I can give that, too.

But you find that consistent with anything? :)

Like the title says, accidents. See the first line of my reply? It runs consistently as the topic - whether as a result of giving it a topic to generate about, perhaps as a result of sentence-fragment feed, perhaps by random. But, it manages to express - poorly - actual, established concepts.

truth, that is, the ideas of approximation

the idea of "... this could never be proved right",

truth exposed to repeated testing should acknowledge and account for ... the effect could never suffice nor are definitely right.

This approximation... is often summarized by the expedient of truth

(though "as" would have been better)

Truth is a quality not quantity, and they hold common concordances that truth based on mere correspondence is not enough to described to it: We never will be proved right

I don't know what the hell you fed it, but it spit out a pretty consistent critique of real-world Bayesian approximation - "mere correspondences," if one will - versus an objective notion of truth. That variance happens to be written deeply into the scientific method, formal logic, and several other traditions of philosophy in science.

You can even check it; I don't know what the hell you fed it, but it wrote (and then, btw, ruined) a perfectly solid treatise. Go pit the italicized quotes up there against the first line or so of my reply. :)

Like I said, I can show you where the CGT broke its Turing plausibility, as well.