For what it’s worth, in the German, Schopenauer wrote: “Ein Mensch kann zwar tun, was er will, aber nicht wollen, was er will.” My non-expert literal translation is: A person can indeed do what he wants, but not want as he wants. In German, wollen is the infinitive of the verb to want, while tun is the infinitive of to do (an alternative of machen, or to make). Query: How do you interpret Schopenhauer’s comment in the context of his work, The World as Will and Representation? (Die Welt als Wille und Vorstellung”?)
For us anglophones, to will is to want something rather explicitly in conscious terms. To want may alternatively have to do with underlying tastes and preferences beyond conscious control. I take issue with your English translation, though native German speakers may have something more insightful to say about this.
If we are looking for “true” interpretations, that is, interpretations mostly likely to have been intended by Schopenhauer, then I would count your third interpetation as being the closest. None of them actually reflect what he is literally saying. I believe, from what I remember barely of Schopenhauer, he viewed ultimate reality as a pervasive underlying Will, a thing in itself (das Ding an Sich) which did not even know what it wanted.
Schopenhauer could not have intended as true more modern interpetations relying upon modern facts about the world, such as cognitive systems, because such facts were not accessible to him writing as he did in the 19th century.
There is a story of Carneades of Cyrene, a post-classical philospher, who came to Rome as an ambassador of Athens. As a member of the New Academy, Carneades was well versed in sceptical argumentation, and stood steadfastly against dogma of any kind. Once in Rome he proceeded to deliver a spellbiding address, arguing that justice should top a list of human motives. The following he day, in service of his real argument concerning the uncertainly of human knowledge (deep scepticism), he proceeded to contradict the argument given the previous day, arguing instead that justice should rank much lower on the scale of human motives. Both declamations were more or less equally compelling. Cato the elder sent Carneades packing, presumably out of Cato’s concern that scepticism widely practised would undermine the Roman military culture and muddle popular thinking.
But somewhere I believe Dawkins is reported as having quipped that it is no bad thing to have an open mind, so long as it is not so open that one’s brains spill out. Doesn’t rationality require one to respect one’s evidence, which one cannot expect to do if one does not know what it is? Therefore, doesn’t one have to posit knowledge as unanalysable, in contrast to belief which may be true or false?