To avoid this getting into an unpleasant exchange when I don't think any such is necessary (or intended), I will risk putting forth how I read his statements (correct me if I'm wrong, Richard).
It isn't that such adaptations are, per se, wrong, bad, or deleterious; but that by the very nature of providing the visualization themselves, they eliminate the reader's exercise of those faculties in favor of a more passive response to a different type of material. As an example: when listening to a piece of "narrative" music, such as Saint-Saenz' "Danse Macabre", the listener forms their own visual reflections of the notes, tones, chords, and relationships of the piece, culled from the imaginative and combinative faculties in their own minds. On the other hand, no matter how great the quality, when listening to the same piece accompanied by images chosen or created by another person, that very process will interfere with, by blending with, altering, or simply replacing, those which the listener would have created on their own. It's like an interference pattern of waves canceling out or replacing a prior or only nascent set. Depending on how intensely an individual remembers such things, that effect can either modify or completely eradicate their own native images from that point on; but that it will have some effect (whether of improvement or degradation depending on the quality of the art concerned) is almost absolutely certain. In turn, this renders any future visualizations connected to that material suspect as entirely one's own.
Hence, while the effect may not be, in the strictest sense, deleterious, it nonetheless interferes with the creations of one's own naturally-occurring visualization.