To me, the main appeal of pixel art is to be able to clearly see the individual pixel boundaries. To a lesser degree, I also like the look created by using a limited palette. Antialiasing makes it harder to tell where one pixel ends and the next begins, and dithering makes it harder to identify the individual colors in an image. While technically impressive, both of those strategies are attempts to imitate traditional mediums instead of embracing pixel art on its own merits.
On the contrary, I'd say they're core concepts of pixel art as a medium. It's not about emulating something else any more than crosshatching with ink is emulating woodcuts. It's about achieving the best results with what you have.
In fact, certain kinds of dithering in combination with PAL colorblending on the C64 will produce completely new color values not available in the fixed palette. Similar tricks were employed with composite CGA output.
