In his online lectures on Computational Science, Prof. Gilbert Strang often interprets divergence as the "transpose" of the gradient, for example here (at 32:30), however he does not explain the reason.
How is it that the divergence can be interpreted as the transpose of the gradient?