We have used density-functional methods including explicit spin-orbit corrections, to calculate the P-31 nuclear magnetic resonance (NMR) chemical shifts of the tetrahalophosphonium cations PX4+ (X = F, Cl, Br, I). The agreement between theory and experimental literature data for PF4+, PCl4+, and PBr4+ is good. For PI4+, the calculations predict an extremely negative (high-field) shift of approximately -520 ppm, due to particularly large spin-orbit contributions from the four heavy iodine substituents, transmitted to the phosphorus nucleus by a very effective Fermi-contact mechanism. No experimental data were available for PI4+. We have, therefore, prepared the compounds PI4AsF6, PI4SbF6, PI4AlI4, and PI4GaI4 and recorded their solid-state P-31 NMR spectra, both with and without magic-angle spinning of the sample. Using the noncoordinating AsF6- and SbF6- anions, the measured isotropic shifts are -519 and -517 ppm, respectively, in good agreement with the predicted extreme value for the isolated cation. In contrast, delta(31)P values of only -304 and -295 ppm are found for PI4AlI4 and PI4GaI4, respectively. The large deviation from the isolated-cation limit in the latter two compounds is probably related to significant I ... I secondary bonding interactions, as found in the solid-state structure of PI4AlI4. The observed solid-state shift tensors are discussed. The present results disagree clearly with previous claims for the synthesis of PI5.