2
$\begingroup$

In the 1994 paper On the Maxwell-Klein-Gordon Equation with Finite Energy of Klainerman and Machedon, the proof of Proposition 1.1 contains the following statement. For $\phi$ (the scalar field of) a classical solution of the Maxwell-Klein-Gordon system, if $$ \| \phi \|_{L^3(\mathbb{R}^3)} \leqslant C(1+t) \left(1 +\| \phi \|_{L^3(\mathbb{R}^3)}\right)^{1/2}, $$ then $$ \| \phi \|_{L^3(\mathbb{R}^3)} \leqslant C(1+t). $$ I don't understand how they deduce this. For a start, it certainly doesn't follow if one ignores the fact that $\phi$ solves the Klein-Gordon equation: take, for example $\phi(t,x) = t^2 \psi(x)$, where $\psi(x)$ is any $L^3(\mathbb{R}^3)$ function with unit norm. The $(1+t)$ looks like it should come from some sort of linear wave equation estimate, but there's no indication they are using anything like that in that proof! What's going on?

$\endgroup$

1 Answer 1

2
$\begingroup$

There is a typo in the paper. Look at the bottom two lines of page 22 which I transcribe here

\begin{align} \|\phi\|_{L^3} & \leq \ldots \\ & \lesssim \mathscr{I}_0^{1/2} (1+t)^{1/2} ( \mathscr{I}_0 + \|A\|_{L^6} \|\phi\|_{L^3} )^{1/2}\\ & \lesssim \mathscr{I}_0 \color{red}{(1+t)} (1 + \|\phi\|_{L^3})^{1/2} \end{align}

If we use Sobolev to control $\|A\|_{L^6} \lesssim \|\nabla A\|_{L^2} \lesssim \mathscr{I}_0$, to get from the second to last line to the last line, the final line should in fact have the term that I marked in red replaced by $\color{green}{(1+t)^{1/2}}$. With this (1.5.c) follows.

$\endgroup$
1
  • $\begingroup$ Ah yes, of course! Thanks for the clarification. $\endgroup$ May 24, 2017 at 17:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.