People do often like 2nd and 3rd order distortion. We did a test in the hearing aid business once with amps that clipped with a soft clip - mostly 3rd order (about 25% THD) and people liked them better than linear amps. But even a small amount of higher order distortion (about .01%) at the waveform crossover and it sounds awful.

So anyone who says that distortion is irrelavent is absolutely wrong, but saying that there are cases where people like distortion is also quite correct. The point is that there is NO correlation between THD (or IMD) and subjective perception - the proof of this is on my web site.