Re: Survey(?) of ConLangs' Calendars and Colors and Kinterms

Yahya Abdal-Aziz wrote:

> > It seems extremely unlikely to me that a language would distinguish
>blue,
> > blue-purple and indigo, but not eg. orange and brown; I would suspect
>that
> > "indigo" were also primarily the name of a substance - that is, the name
> > of the dye solution, not of its color. (It doesn't exactly matter that
>there's
> > a separate word for the indigo plant too. Especially since there is not
>a
> > single "the" indigo plant.)
>
>If the word 'nila' only named the dye, and not its colour
>(I don't say it does!), would it still seem as unlikely to you
>that the language would distinguish blue and [blue-]purple,
>but not orange and brown?
Not really. Distinguishing two very similar hues of bluish purple is MUCH
odder than distinguishing blue and purple.


>If you took a systematic approach to colours, you might
>start with the standard 'subtractive colours' red, blue
>and yellow as 'primary' colours.  You'd then mix pairs of
>these primaries in (optically) equal proportions to create
>the 'secondary' colours purple, green and orange.  Finally,
>you'd mix all three primaries in equal proportions to get
>brown.  Of course, you could mix any of these colours
>with white to get 'tints' or with black to get 'shades'.
>And you could vary the proportions to get different hues.
That's of course just one solution; I can think of other, equally systematic
ones. CMYK pops easily into mind next; you could also start from dividing
the spectrum in equal parts (and this can be done either by frequency, by
wavelenght, or by the projection of a prism), or from the details of human
color vision, or from the colors of certain common substances (chlorophyll,
hemoglobine, etc.)

And naturally, by mixing RBY in *optically equal* proportions, you ought to
get black or dark gray, not brown.


>But there's no real reason to believe that speakers of
>any natlang have deliberately constructed their colour
>terms in such a systematic way.  If, as I believe, humans
>have invented or adopted colour terms mostly for their
>practical utility, then the selection of colours named in
>natlangs would, one expects, be rather unsystematic and
>even patchy.  Such is the case in the various classes of
>the typology we've been discussing.
>
>
> > This of course doesn't override blue-purple (could we say that it's
> > "phonemically" just plain purple?) ...
>
>Not phonemically, I think, but semantically?
I'm using the word "phonemically" in the sense that this color, with regards
to the rest of the system, occupies the /purple/ slot, even if it's not
exactly same as [purple] as defined by the English word. If there is a word
which relates to "semantical" the same way as "phonemical" relates to
"phonetical", I would sure like to know it. I could just coin "semammical"
but that sounds too zarking ridiculous. :)


>If we suppose that, on grounds of economy, any language
>makes only those distinctions which it finds useful, then it
>would follow that the 'purple' attribute must have been more
>useful than the 'brown' attribute.  My guess is that the primary
>use of the colour 'purple' would have been in teaching young
>people to identify fruit in the forest.
Makes sense. I would still expect that color terms based on utility would
have to overlap where they are psychologically close. And I don't mean fuzzy
borderline cases; I mean that a certain color's name would in one context be
(in this case) indigo, and in another, purple, with little ambiguity.


> > Also:
> > >The most useful insights given by this classification are, I think,
> > ...
> > >2. that the most basic contrast is the most extreme - between light
> > >and dark (white and black); and the second is between 'hot' colours
> > >(red) and the rest.
> >
> > So, with system IIIa, you'd analyze "red" = "dark hot", "yellow" =
>"light
> > hot"? Interesting, but what about IIIb then?? Green seems to be somewhat
> > neutral with regards to both the light/dark and hot/cold dichotomies.
>
>'Green' is a very interesting colour!   Mostly because
>there are just so _many_ different greens.  In English
>we need to qualify our greens to get yellow-greens and
>blue-greens, and I think this is also the case in most
>natlangs.  I find this rather surprising, since perceptually,
>human vision is most responsive to greens, in two ways:
>first, that green light contributes approximately 60% to
>our sensation of brightness (red 30% and blue only 10%);
>second, that we can distinguish many more hues of green
>than of any other of the seven spectral colours ROYGBIV
>(Red Orange Yellow Green Blue Indigo Violet) formed by
>splitting natural white light by a prism or rainbow.
It is true that green has the widest wavelenght / frequence range, but this
does not necessarily imply that it also has the widest psychological range.
I'd in fact argue the opposite: since such a large part of the spectrum
appears "green", it must be comparatively harder for the the human eye to
distinguish colors of that region from each another.
And when you say that "green" light makes up 60% of brightness, I suppose
you actually mean the mid-freq cone cells, which are often called "green"?
Their intensity peak is actually in the yellow region - which is why yellow
appears the brightest of the spectral colors.


>in the last few years, interior designers
>have moved away from using the traditional colour wheel,
>in which red, blue and yellow are spaced equally around
>the circumference of the circle, with various mixtures
>as colour gradients between them. They now use a FOUR-
>colour wheel, with 'primaries' of red, blue, green and
>yellow.  Doing so makes many more hues of green readily
>available for selection.  (As a byproduct, it also changes
>which colours are considered complements - those which
>lie directly opposite each other on the wheel.)
I recall reading that human color vision transforms the signals coming from
the eye into 3 channels which correspond with darkness/lightness,
redness/greenness and yellowness/blueness... so this makes sense, too.


>A yellow-green (think a psychedelic lime) can be a very
>hot green; a blue-green (think aquamarine, turquoise,
>cyan) can range from light (and bright) to dark, and
>among these, the darker colours are generally seen as
>cooler and the lighter as warmer.  However, specific
>experiences may change a very pale blue-green to ice!
>I think there's a couple of factors that interact to
>determine how hot we feel a particular blue-green is:
>the intensity or saturation and the experiences our
>environment expose us to.  If you've never seen ice,
>but have only ever seen a pale blue-green as the colour
>of a sky washed out by a blazing white-hot sun, the
>'icy' blue-green someone else sees may well be instead
>a 'scorching' blue-green.
>
>Regards,
>Yahya
Yes, this is now closer to my latest point: that green does not fit into a
single spot on the hot/cold spectrum. Yet you were saying that the
secondmost basic color contrast is the contrast of hot and non-hot, which
does not fit well together with BWRG color systems.
...No, wait, I just realized it does: "non-hot" is a negative definition and
therefore not as basic as "hot". So BWRG systems simply ignore non-hotness,
and instead add a new, 4th distinction of greenness. In other words, BWR is
triangular, BWRY square planar and BWRG is tetrahedral...

Hey, let's see how far this analogy goes? BWRYG is obviously square
pyramidical, and BWRYGB would be octahedral. After that a geometrical
interpretation gets harder, tho not impossible. Adding brown, we have a
tetrahedron contained within a octahedron (if the octahedron is regular, the
tetrahedron needs to be a very flat one). Orange, pink and purple would be
added on edges rather than corners, and gray in the center, which finally
break the polyhedra analogy.
But let's now take a peek at the polyhedra which do *not* appear. There is
no trigonal bipyramid system (eg. BWRYB or BWRGB), nor a cubical one. Yet,
with regular or semi-regular polyhedra, the rhombic dodecahedron is the
simplest one where the distance to the center is equal or greater than the
distance between vertexes. That polyhedra has 14 vertexes. In other words,
monomorphemic non-extreme color terms wouldn't be expected to come in before
there's at least some 10-12 color terms on the boundary. Therefore either
1) color systems are not based on psychologically equivalent distances
between color terms, or
2) colorspace is non-euchlidic.
As intriguing as the second choice is, surveys on additive and subtractive
colorspaces (where mixtures of two colors always lie on the same line) seem
to point towards a simple Euclidian space. However, this might be circular
reasoning: maybe an exotic geometry of sorts could be involved, where there
would be *two* shortest routes between any two points - one for additive and
one for subtractive mixing. I wonder what that would look like? :)
Then again, I'm not yet convinced if mixing of colors is as simple as it is
claimed to be. Different wavelenght mixtures A+B and C+D may look like the
same color, but is it really certain that when mixed with some other
wavelenght E they still remain the same color?

But I'm getting carried away. Anyway, on hues of green, personally I
actually identify cyan as a secondary color, distinct from both green and
blue but relative to both. However, I do not perceive a similar secondary
color existing between yellow and green. So in my perception, the spectrum
consists of four primary and three secondary colors: RoYGcBv. And even if I
had to add a fourth secondary color, it would be magenta, not yellow-green
(nor indigo!)

John Vertical

Reply