I don't understand the argument that "the power of the country's government to tax" is a point in favor of a currency. I would say that it is pretty clear that the opposite is true for almost every significant tax; that a country collecting taxes in a currency would be a disadvantage of a currency. The fact that countries need to pass laws allowing taxation of non-local currencies seems to support that.
For a limited class of taxes, namely, taxes directly denominated in a currency (like property taxes), the taxability of a currency adds to the demand for that currency. But for the most part, since non-local currency transactions and barter transactions are taxed at the effective exchange rate / fair value, the fact that taxes are collected in dollars would seem to neither hurt nor help the currency.
That said, in summary, the "artificial supply limitation" applies to both Bitcoin and dollars, with dollars relying purely on trust that America will not increase the money supply unnecessarily. That inflation has actually been mild is a "bug" in the system -- the Fed is deliberately trying to spur inflation, and has not been having success, for reasons that nobody (including themselves) completely understands. The calculable probability of return that you refer to has been pretty reliable (just like MBS's in 2006, couldn't resist), so that is true, but I don't see that being related to the taxability of the currency so much as to the lack of apparent inflation compared to the nominal returns on sovereign debt.
The monetary base argument also seems specious; M0, the "monetary base", is not generally considered (except by non-mainstream economists) to be all that significant compared to the higher-order money supply factors caused by dollar-denominated assets of varying liquidity. Once (or if) Bitcoin develops a credit market, then the medium-term effective money supply will grow and shrink as the market demands.