Well, please be more specific about the molecule(s) in question for that quotation. d-orbitals are certainly involved in some bonding, especially for the transition metals, but you are right that the participation of d-orbitals in the bonding of main group covalent molecules and polyatomic ions has been overstated in the past and is in fact fairly minimal.
As you mentioned, there is very little, if any, d-orbital participation in the bonding of sulfate. As for d-orbitals being involved in polarization but not bonding, that literally does not make sense for something like sulfate. Sulfur's d-orbitals are empty, so it is impossible for them to polarize without bonding. Only when there are occupied d-orbitals (as in transition metals) can non-bonding d-orbitals affect things like the shape and polarity of the compound (as in crystal field theory/LFT).
So, long story short, I think that comment you read is simply wrong, and your original understanding is correct. See the below references if you want to read more about it.
- Reed, A. E.; Schleyer, P. v. R. Chemical Bonding in Hypervalent Molecules. The Dominance of Ionic Bonding and Negative Hyperconjugation over d-Orbital Participation. J. Am. Chem. Soc. 1990, 112 (4), 1434–1445. https://doi.org/10.1021/ja00160a022.
- Cunningham, T. P.; Cooper, D. L.; Gerratt, J.; Karadakov, P. B.; Raimondi, M. Chemical Bonding in Oxofluorides of Hypercoordinate Sulfur. Faraday Trans. 1997, 93 (13), 2247–2254. https://doi.org/10.1039/a700708f.