Lewis held that our common-sense-beliefs have greater initial plausibility than every philosophical argument against them, be it in mathematics ("there are numbers") or metaphysics ("there is time"), philosophy of mind ("there are beliefs"), ethics, etc.

Philosophy can help to find a realiser - a best candidate - for the role of numbers, beliefs, etc., but the price for "losing our moorings" (after g.e. moore), i.e., denying common-sense propositions, is almost always too high.

There is at least one case, of course, where Lewis was willing to pay: modal realism.

Is he? I actually didn't get that impression.