Summary: In a previous work [the authors, J. Eur. Math. Soc. (JEMS) 25, No. 6, 2057–2125 (2022; Zbl 07714608)], we established a multilinear duality and factorisation theory for norm inequalities for pointwise weighted geometric means of positive linear operators defined on normed lattices. In this paper, we extend the reach of the theory for the first time to the setting of general linear operators defined on normed spaces. The scope of this theory includes multilinear Fourier restriction-type inequalities. We also sharpen our previous theory of positive operators. Our results all share a common theme: estimates on a weighted geometric mean of linear operators can be disentangled into quantitatively linked estimates on each operator separately. The concept of disentanglement recurs throughout the paper. The methods we used in the previous work – principally convex optimisation – relied strongly on positivity. In contrast, in this paper we use a vector-valued reformulation of disentanglement, geometric properties (Rademacher-type) of the underlying normed spaces, and probabilistic considerations related to \(p\)-stable random variables.