President Barack Obama said over the weekend that he wants Congress to grant him the “authorization” to do what he feels he already has the “authority” to do: take military action against Syria.

This linguistic self-contradiction, highlighted by my National Review colleague Matthew Franck, is symptomatic of a widespread confusion about the president’s war powers. It is also a reversal of candidate Obama’s 2007 view that the Constitution lets the president initiate a military attack only to stop “an actual or imminent threat to the nation.” Nobody believes that the Syrian regime is such a threat.

Arguments about the limits of the president’s powers often get sidetracked by the War Powers Act, which requires the president to get congressional approval for military action within 60 days of taking it. No president has acknowledged that statute, passed over President Richard Nixon’s veto, as constitutional. That act gives the president both more and less power than the Constitution does. It subjects him to a congressional veto even if he is exercising his duties as commander in chief to respond to an imminent threat. At the same time it implicitly allows him to engage in temporary military actions even if he isn’t responding to a threat to the country and has no congressional authorization.

The Constitution trumps any statute. The question is whether its grant to Congress of the power to “declare war” means that Congress has to authorize discretionary military action -- that is, action that is not a response to a past or impending attack on Americans -- before the president can act. It would be foolish to read the constitutional text to mean that Congress has to use the words “declare war” to make military action legitimate: The Constitution itself nowhere says anything to that effect.

The arguments that the president need not get congressional approval, however that approval is worded, have always struck me as sophistry: attempts to narrow the meaning of “declare” or “war” beyond any plausible construction. ("If it’s not World War II, it’s not really a war …") Franck points out that Alexander Hamilton, the Founding Father who took the most expansive view of presidential power, thought congressional approval was necessary. “I am not ready to say that [the president] has any other power than merely … to repel force by force,” he wrote in 1798, and to go further “requires the sanction of that Department which is to declare or make war.”

Or look at Federalist 69, in which Hamilton tries to build support for ratifying the Constitution by assuaging fears that the executive it creates would be too powerful. His power, he writes, will be “in substance much inferior” to that of the British king: “It would amount to nothing more than the supreme command and direction of the military and naval forces, as first General and admiral of the Confederacy; while that of the British king extends to the declaring of war and to the raising and regulating of fleets and armies -- all which, by the Constitution under consideration, would appertain to the legislature.”

Hamilton went on to say that the governors of New Hampshire and Massachusetts may well have more power over their militias. An artful proponent of broad presidential war powers could try to twist that passage, I suppose: Obama would just be “directing” the military to act in Syria, and “declaring” war doesn’t really mean what we think it means.

But the most natural reading of the evidence is that it’s Congress that moves us from a state of peace to a state of war. If Congress votes against a strike on Syria and the president acts anyway, he will be violating his oath of office.

(Ramesh Ponnuru is a Bloomberg View columnist. Follow him on Twitter.)