It seems that we find the same "battlefield of endless controversies" in figuring out what to make of metaphysics as Kant found within the metaphysics itself. But perhaps hosting controversies is the mission of metaphysics. If we accept that knowledge is fallible and advances through improvement, its improvements must first be conceived without sufficient evidence, on a hunch. Perhaps the intractable "questions" are only questions in their surface grammar, and have much more oblique depth significance (stimulating and guiding other research, for example). In the end, philosophy is the handmaiden of the culture at large, and its clients either have ways of resolving their own controversies (science) or thriving on them (art). As long as they benefit from metaphysics as it is, why change what works as intended? Friedman's Dynamics of Reason (p.107) suggests that we should not:
"Although we do not (and I believe should not) attain a stable consensus on the results of distinctively philosophical debate, we do, nonetheless, achieve a relatively stable consensus on what are the important contributions to the debate and, accordingly, on what moves and arguments must be taken seriously... Characteristically philosophical reflection interacts with properly scientific reflection in such a way that controversial and conceptually problematic philosophical themes become productively intertwined with relatively uncontroversial and unproblematic scientific accomplishments; as a result, philosophical reflection can facilitate interaction between different (relatively uncontroversial and unproblematic) areas of scientific reflection."
In earlier times philosophers were more definitive, negatively so. Kant thought that he diagnosed the problem and found a solution, once and for all, in Critique of Pure Reason by way of Copernican turn from metaphysics to epistemology, "to seek the observed movements, not in the heavenly bodies, but in the spectator":
"Our reason (Vernunft) has this peculiar fate that, with reference to one class of its knowledge, it is always troubled with questions which cannot be ignored, because they spring from the very nature of reason, and which cannot be answered, because they transcend the powers of human reason... Thus, however, reason becomes involved in darkness and contradictions, from which, no doubt, it may conclude that errors must be lurking somewhere, but without being able to discover them, because the principles which it follows transcend all the limits of experience and therefore withdraw themselves from all experimental tests. It is the battlefield of these endless controversies which is called Metaphysic... I venture to maintain that there ought not to be one single metaphysical problem that has not been solved here."
Hegel told us that "the answer to questions that philosophy leaves unanswered is that they must be put differently". C.S. Peirce argued that the flaw was in the misguided Cartesian maxim, and that philosophy, like science, must admit its perpetual fallibility and seek self-correction, not ultimate answers. As he wrote in Some Consequences of Four Incapacities:
"The same formalism appears in the Cartesian criterion, which amounts to
this: "Whatever I am clearly convinced of, is true." If I were really convinced, I should have done with reasoning and should require no test of certainty. But thus to make single individuals absolute judges of truth is most pernicious. The result is that metaphysicians will all agree that metaphysics has reached a pitch of certainty far beyond that of the physical sciences; -- only they can agree upon nothing else. In sciences in which men come to agreement, when a theory has been broached it is considered to be on probation until this agreement is reached. After it is reached, the question of certainty becomes an idle one, because there is no one left who doubts it."
After the epistemological and scientific turns it was time for the linguistic turn. Carnap, in a distant echo of Kant, merrily declared in self-explanatorily titled Elimination of Metaphysics Through Logical Analysis of Language:
"In the domain of metaphysics, including all philosophy of value and normative theory, logical analysis yields the negative result that the alleged statements in this domain are entirely meaningless. Therewith a radical elimination of metaphysics is attained, which was not yet possible from the earlier antimetaphysical standpoints."
Wittgenstein mused about intractability of metaphysical problems in 1930-s, after parting ways with positivists. He compared irresolvable philosophical problems to "diseases of understanding" and wondered:
"Have we to do with mistakes and difficulties that are as old as language? Are they, so to speak, diseases that are bound up with the use of a language, or are they of a more special nature, characteristic of our civilization? Or also: Is the pre-occupation with the edium of language that runs through all our philosophy an ancienttrend of all philosophizing of all philosophy, an ancient struggle? Or is it new, like our science?" [MS 132, 7 quoted from Gordon Baker's late interpretation of Wittgenstein by Hacker]
What emerged out of these musings was his philosophical therapy of late period, originally modeled on Freud's psychoanalytic method (which he later came to despise, however). Instead of attacking positions with arguments remove the confusions and anxieties that move people to insist on those positions in the first place. Dissolution and conciliation in place of solutions.
But then Quine, Carnap's nemesis, adopted a far more pragmatic view of metaphysics (or ontology, as he preferred to call it) as organizing disparate parts of our knowledge for various uses. And even Peirce himself developed an elaborate speculative metaphysics of his own with universe as "effete mind, inveterate habits becoming laws" in a series of papers one of which was tellingly titled A Guess at the Riddle.