Glen Gordon writes
> But why a distinction between "grue" and "dark"? Is there
> anything in IE myth to suggest _four_ colours? Everything
> seems to point to three.
PIE probably, yea, even most likely, had only three color-terms. I
mentioned grue (green-blue) because EIEC hints that some scholars
seem
say there were four, or at least, at the very last stage, was moving
to four colors.
With only 3 or 4 color-words, the actual spread of 'color' would be
larger than we today tolerate. I suspect 'row' (red-yellow) included
what we would call orange, pink, and even 'tan'. I suspect what we
call light gray would have been called 'light'.
Undoubtedly, of course, they could say something like 'light like new
grass', 'dark like pine needles' or 'light like a clear cloudless sky
after a rainstorm'. But these are not real color-words.