Boundedness of a composition of dyadic paraproducts
Ana Čolović
TL;DR
The article resolves the boundedness problem for the composition of dyadic paraproducts of type $(0,1,0,1)$, providing necessary and sufficient testing conditions for $P_b^{(0,1)}\circ P_d^{(0,1)}$ and showing norm equivalence to a model operator on the upper half-plane. It introduces a transplantation framework to $\mathcal{H}$, defines a model operator $T^{(0,1,0,1)}_{b,d}$ with matching Gram matrices, and reduces boundedness to three testing conditions $A$, $B$, and $C$. The authors prove the necessity of these conditions via forward/backward testing on weighted Carleson tiles and establish sufficiency by careful decomposition of the associated bilinear form into three controllable parts, ultimately bounding the operator by the product $\|b\|_{BMO}\|d\|_{BMO}$. A direct $H^1$–$BMO$ duality argument provides the sharp $BMO$-norm control of the overall bound, completing the characterization. This work settles the remaining case in the dyadic paraproduct composition problem and links the dyadic setting to two-weight-type testing phenomena on the upper half-plane.
Abstract
We resolve the question of the boundedness of the composition of dyadic paraproducts, first posed by Pott, Reguera, Sawyer, and Wick in~\cite{PotCarSawWic}, by providing necessary and sufficient conditions for their boundedness.
