Why does the visible light spectrum appear to be circular [duplicate]

I'm not sure if this is an acceptable place to ask this question, as it may have more to do with the biological workings of human eyes than with the physical properties of light, but I'd rather hear a physical explanation than a biological one (if one exists).

Essentially, we often describe the visible spectrum with a color wheel. In this wheel, Red appears next to Purple (violet), but Red and Violet are at opposite ends of the visible light spectrum. What accounts for this? That is, why do we perceive the Visible Light Spectrum as circular, instead of linear?

marked as duplicate by Ben Crowell, Qmechanic♦Aug 14 '13 at 18:40

• You might find this interesting. youtube.com/watch?v=S9dqJRyk0YM – joshphysics Aug 2 '13 at 19:16
• @joshphysics I actually had seen that video (though didn't remember it). Henry's video was entertaining, but I don't think it really gave an insight as to why we roll up all the "left over" light into pink. Furthermore, it didn't even attempt to say why rolling up the electromagnetic spectrum makes any sense. Why should EM radiation with the shortest imaginable wavelength be anything like EM radiation with the longest imaginable wavelength? – Daniel Rosenthal Aug 2 '13 at 19:28
• duplicate of physics.stackexchange.com/q/40763 – Ben Crowell Aug 14 '13 at 14:12

Humans are trichromats which means that we see color by processing information from three different color channels. Specifically, we have three different types of cone cells, each of which sends a signal to our brain, and those signals are processed by our brain yielding a perceived color. As a result, perceived human color is a function $f(S,M,L)$ of the response values $S$, $M$, and $L$ of the so-called "$S$hort," "$M$edium," and "$L$ong" cones. The cones are given these names because the short cone has a peak response at relatively short wavelengths of incoming light (blue), the medium cone has a peak response at medium wavelengths, etc.