It's like this.
You really get more exposure in the shadows, and the impression that this will dilute them is right. But with optical printing, a (slightly) denser negative will require longer printing. Longer printing means more exposure in the shadows, and this puts them back in track.
Scanners behave differently. Usually, the software used to process the image will account both highlights and shadows before a positive picture is produced.
A dense negative will require lengthier printing, optically. Lengthy exposures causes more halide to be affected. When more halide is developed, more dye is produced. Lengthy printing with a normally exposed negative (with less density) will cause the hues to go dark. But when the dyes in the negative are already dense to begin with- something achieved with overexposure in colour negative stocks- the light blocking power of the dyes prevent details from darkening. The dyes aren't quite opaque enough to totally curtail exposure.
In the past when we used low grade or "budget" colour negatives (we had low end Kodak colour film which was very pastel, or the earlier Luckycolor films), the only way to give punch to their normally washed out rendering was to slightly overexpose them. The Kodak had a box speed of 100, but it gave nice results at EI 64. Luckycolor and Eracolor (both from China, and this was in the early 1990s) were supposed to be 100, but EI 50 or even EI32 did better. In fact, the "budget" Kodak 100 was eventually repackaged as KodakSP 64.
These poor grade negative stocks produced low dye densities with normal exposure. The only practical way to make them produce more dyes for more punchy hues was to increase exposure- more exposure, more development, equals more dyes.