A pinch of philosophy to begin with

Try­ing to answer the title ques­tion: “In what way does the brain see col­ors?” we can­not avoid the answer to a much more basic ques­tion about the col­or­ful­ness of what the brain or the world might see. In oth­er words, the point is whether the real­i­ty around us is col­or­ful. Every­day expe­ri­ence con­stant­ly con­firms us in this belief. How­ev­er, the col­or­ful­ness of the world is not as obvi­ous as it might appear at first glance. First of all, if we want to give some infor­ma­tion about the world around us, we must first agree on where the world ends and where the one who explores it begins. A key point is decid­ing where the bound­ary between the world-which-sur­rounds me and the see­ing self, lies. In fact, it is our see­ing self that expe­ri­ences the col­or­ful­ness of the world and states some­thing about it.

My inten­tion is to sig­nal the prob­lem that is impor­tant for under­stand­ing what per­cep­tion is, rather than write some naive phi­los­o­phy paper. The most obvi­ous answer that comes to mind about the bound­ary between see­ing self and the-world-around-me is based on the obser­va­tion that what sep­a­rates me from the world is sim­ply my skin. Every­thing out­side my skin is not me, and, there­fore, it belongs to the world I recog­nise. The para­dox is that by accept­ing such an assump­tion as obvi­ous we must reject anoth­er, equal­ly obvi­ous, one.

If the skin is my bound­ary sep­a­rat­ing me from the world, then the world is col­or­less. The grass is not green and the sky is not blue. It is only an illu­sion which my eyes and the brain – locat­ed in the leather sack of my body – show me. Fur­ther­more, from this point of view, the world is not only col­or­less, but worse – it is not even dark-grey! The cat­e­gories of bright­ness – just like the cat­e­gories of col­or­ful­ness – are not the prop­er­ties of light, but of the per­son who is sen­si­tive to them.

In order to under­stand this obvi­ous­ness, let us think about what has already been indi­cat­ed on the pages of this book. Vis­i­ble light (in fact, an elec­tro­mag­net­ic wave in the range of about 400–700 nanome­tres and vari­able ampli­tude) stim­u­lates pho­tore­cep­tors in the reti­na. And that is all we know about this wave.  Although we asso­ciate col­or with one or anoth­er light length, nev­er­the­less, it is nei­ther light nor dark, nor col­or­ful. These are sim­ply some pho­tore­cep­tors that react more inten­sive­ly to a cer­tain length and ampli­tude of this wave, and the brain inter­prets their response, cre­at­ing the impres­sion of col­or with a cer­tain inten­si­ty and brightness.

Obvi­ous­ly, I do not want to be bor­ing, but once again I would remind that we still have no idea how the col­or vision is gen­er­at­ed based on reac­tion of mil­lions of neu­rons involved in the pro­cess­ing of data record­ed by pho­tore­cep­tors. Let us leave this prob­lem to pro­fes­sion­al philoso­phers. In any case, empir­i­cal sci­ences have not pro­vid­ed any ver­i­fi­able hypoth­e­sis con­cern­ing this prob­lem yet.

How­ev­er, if we do not like the idea of ​​the col­or­less­ness of the world, it is nec­es­sary to con­sid­er whether it would be pos­si­ble some­how to shift the bound­ary between see­ing self and the-world-around-me, so that the real­i­ty out­side me might receive some splen­dour. For our pur­pos­es, it would be rather dif­fi­cult to jus­ti­fy extend­ing our see­ing self beyond the skin, since we have no rea­son to doubt that it is the eyes and the brain that gen­er­ate col­ors. It would be nec­es­sary there­fore to move the bor­der of see­ing self to inside the body so that those brain struc­tures which are respon­si­ble for col­or vision would be on the world’s side.  Thanks to this, no one could ques­tion the obvi­ous expe­ri­ence of its col­or­ful­ness. Unfor­tu­nate­ly, such an approach rais­es fur­ther prob­lems only. I will draw the atten­tion to two of them. 

First­ly, all men­tal states, includ­ing the expe­ri­ence of one’s self (also the see­ing self), are more or less a deriv­a­tive of a whole, func­tion­ing brain. So how to sep­a­rate the parts of the brain that are respon­si­ble for col­or vision from the ones that gen­er­ate our sense of self?  And sec­ond­ly, the intro­duc­tion to our rea­son­ing of some imma­te­r­i­al or non-neu­ronal ele­ment that would be respon­si­ble for the exis­tence of the self (in each of its roles), in oppo­si­tion to the states of mind result­ing from its cog­ni­tive activ­i­ty, e.g. the col­or vision, would throw us beyond the bound­aries of empir­i­cal sci­ence in the sense that cur­rent­ly belongs to this concept.

There­fore, there is prob­a­bly no sense in devel­op­ing these points fur­ther, because they would exceed the scope of this book. It is enough to say that the con­cept accord­ing to which the world is col­or­less and non-bright, and the bor­der of our see­ing self is on the cornea of ​​the eye, seems to be the most sen­si­ble solu­tion, at least at the present stage of knowl­edge about vision.

Filling contours or contouring color spots?

Based on what I have writ­ten so far in rela­tion to the per­cep­tion of con­tours of observed things, one can for­mu­late a gen­er­al rule that objects in the visu­al scene are iden­ti­fied more accu­rate­ly if their con­tours appear more clear­ly by con­trast­ing sur­faces of dif­fer­ent bright­ness and if these con­tours are com­plete and not “dis­tort­ed” by means of e.g. chiaroscuro.

Mike G. Har­ris and Glyn W. Humphreys (2002) note that when asked to draw any object, we usu­al­ly start by set­ting the out­lines, not by shad­ing or col­or­ing sev­er­al adja­cent sur­faces. Apart from the fact that the lat­er­al inhi­bi­tion mech­a­nism shaped in the course of evo­lu­tion is large­ly respon­si­ble for this, nev­er­the­less since child­hood we have been trained to fill the con­tour pic­tures with dif­fer­ent col­ors (as in e.g. Fig. 107) rather than con­tour­ing the col­ored sur­faces. And we do so ignor­ing the fact that the vast major­i­ty of objects do not have any con­tour lines. That what we see are just sur­faces of dif­fer­ent col­ors or — in low light — of dif­fer­ent bright­ness, that mark the bound­aries of objects sep­a­rat­ing them from the background.

Fig­ure 107. Flin­stones, illus­tra­tion from chil­dren’s col­or­ing book. Graph­ic design: P.F.

So what role do col­ors play in per­ceiv­ing the world around us? Do we need them to iden­ti­fy objects? Or maybe col­or vision is just an extra addi­tion that evo­lu­tion has equipped us with?

On the one hand, the ratio of the num­ber of cones that are respon­si­ble for col­or vision to the num­ber of rods that respond to the bright­ness of the light in low light con­di­tions is approx. 1:20. This means that we can doubt if we need col­or vision for any­thing. On the oth­er hand, there are incom­pa­ra­bly more cones than rodes around the fovea, which is a spe­cial place because of visu­al acu­ity. There­fore, one can right­ly assume that vision medi­at­ed by col­or-sen­si­tive cones, how­ev­er, plays a spe­cial role in view­ing the world.

It appears that with­out look­ing at the mech­a­nisms respon­si­ble for col­or vision we will not solve these dilem­mas. It is worth start­ing this sto­ry by recall­ing some facts from history.

Young-Helmholtz trichromatic theory

Young-Helmholtz trichro­mat­ic the­o­ry was cre­at­ed in the 18th cen­tu­ry, and its pre­cur­sors were the British men: Sir Isaak New­ton, George Palmer and Sir Thomas Young. The first of them laid the cor­ner­stone for both mod­ern optics and the the­o­ry of col­or vision. The results of his exper­i­ments on the split­ting of white light into the col­or spec­trum negat­ed the cen­turies-old tra­di­tion, accord­ing to which it was not light, but objects that caused us to see colors.

Before, col­or was thought to be the prop­er­ty of objects, not vision. Peo­ple thought that light brings out col­ors from objects and thus they are per­ceived by the observ­er. All cats are grey after dark because there is too lit­tle light to reveal their actu­al col­or. We just need to turn on the lamp and we can imme­di­ate­ly see that a cat is gin­ger, for exam­ple. New­ton ques­tioned this idea. He point­ed out that the object itself, depend­ing on the inten­si­ty and qual­i­ty, i.e. the shade of the light falling on it, can cause com­plete­ly dif­fer­ent col­or expe­ri­ence. On this basis, he con­clud­ed that col­or vision is much more a deriv­a­tive of the vision sys­tem than of the prop­er­ties of the objects viewed.

Fas­ci­nat­ed by New­ton’s dis­cov­er­ies, George Palmer in a book pub­lished in 1777 sug­gest­ed that the nerve fibres that make up the reti­na of the eye con­tain three types of col­or­ing par­ti­cles. Each of these par­ti­cles absorbs rays of light with a length cor­re­spond­ing to one of the col­ors: red, yel­low or blue. Accord­ing to Palmer, per­ceiv­ing dif­fer­ent col­ors is the result of mix­ing of two or three col­or­ing par­ti­cles in the eye in the pro­por­tion in which they absorb light of the length they are sen­si­tive to. Palmer was a chemist and there­fore he explained the col­or vision mech­a­nism by anal­o­gy with the process of paint mix­ing, i.e. sub­trac­tive col­or syn­the­sis (Fig. 108).

Fig­ure 108. The phe­nom­e­non of sub­trac­tive syn­the­sis of three col­ors: yel­low, magen­ta (shade of pur­ple) and cyan (shade of blue). Graph­ic design: P.F.

Accord­ing to Palmer, the impres­sion of green aris­es as a result of mix­ing of two col­or­ing par­ti­cles that have absorbed rays of light of a length cor­re­spond­ing to yel­low and cyan, i.e. blue col­or hue. Red is a deriv­a­tive of a mix­ture of col­ors that absorb light of a length cor­re­spond­ing to yel­low and magen­ta, and blue — a deriv­a­tive of magen­ta and cyan. The light absorbed by all three types of mol­e­cules gives the impres­sion of black, while white is a con­se­quence of the lack of stim­u­la­tion of any of the reti­nal col­ors (Lowen­gard, 2006).

Georges Palmer built his con­cept on two pil­lars: New­ton­ian optics and the the­o­ry of col­or mix­ing, devel­oped by French chemists, Charles Dufay and Jean Hel­lot. His con­cept sur­vived for the next 30 years, but caused much con­tro­ver­sy. First of all, Palmer could not explain why the col­or­ing par­ti­cles he pos­tu­lat­ed would only react to the three wave­lengths of light indi­cat­ed by him. Is it pos­si­ble that the remain­ing wave­lengths of vis­i­ble light play no role in col­or vision? And sec­ond­ly, with the devel­op­ment of phys­i­ol­o­gy, his con­cept of nerve fibres, in which — accord­ing to Palmer — three col­or­ing par­ti­cles would mix like in tiny cru­cibles, raised more and more doubt.

In 1807, Sir Thomas Young, dur­ing lec­tures on nat­ur­al phi­los­o­phy, pre­sent­ed the con­cept accord­ing to which the reti­na of the eye is not built of nerve fibres filled with three col­or­ing par­ti­cles, but of three types of recep­tor cells that react with vary­ing inten­si­ty to light rays cov­er­ing the entire vis­i­ble light spec­trum. Young also sug­gest­ed to, when con­sid­er­ing col­or vision, depart from the idea of ​​col­or mix­ing in the eye, as Palmer pre­sent­ed, refer­ring to the anal­o­gy of paint mix­ing. Instead, Young pos­tu­lat­ed the con­cept of cumu­lat­ing the strength of recep­tor reac­tion to stim­u­la­tion with light waves of dif­fer­ent lengths. Although he gen­er­al­ly agreed with Palmer’s idea that all col­ors are a deriv­a­tive of mix­ing of three com­po­nents, he used the col­or spec­trum obtained as a result of white light dis­per­sion as the basis for their deter­mi­na­tion. He asso­ci­at­ed three points of max­i­mum recep­tor sen­si­tiv­i­ty with blue, green and red (Fig. 109). His selec­tion was obvi­ous­ly arbi­trary, but not with­out log­ic. After divid­ing the entire spec­trum of vis­i­ble light into three parts, the col­ors dis­tin­guished by him def­i­nite­ly dom­i­nate with­in each category.

Fig­ure 109. The col­or spec­trum after dis­per­sion of vis­i­ble light as a basis for Young to dis­tin­guish three col­ors to which three types of recep­tors locat­ed in the eye are sen­si­tive. Graph­ic design: P.F.

It is worth to men­tion that although the results of sub­se­quent phys­i­o­log­i­cal stud­ies did not ful­ly con­firm Young’s intu­ition regard­ing pho­tore­cep­tor sen­si­tiv­i­ty to red, green and blue, nev­er­the­less their divi­sion into R- (red), G- (green) and B‑cones (blue) remains valid until now.

A mod­el descrip­tion of the rela­tion­ship between light wave­length and sen­si­tiv­i­ty of the three types of pho­tore­cep­tors pos­tu­lat­ed by Young was pre­sent­ed in 1851 by a Ger­man phys­i­ol­o­gist and physi­cist, Her­mann von Helmholtz. Like Young, he argued that per­ceiv­ing dif­fer­ent col­ors is asso­ci­at­ed with vary­ing pho­tore­cep­tor sen­si­tiv­i­ty to all vis­i­ble wave­lengths. He also believed that for each type of pho­tore­cep­tor there is such a wave­length of light to which the pho­tore­cep­tor is most sen­si­tive (see Fig. 110).

Fig­ure 110. Rela­tion­ships between sen­so­ry sen­si­tiv­i­ty of three hypo­thet­i­cal visu­al recep­tors and the wave­length of light in the vis­i­ble light spec­trum, accord­ing to Young-Helmholtz trichro­mat­ic the­o­ry of col­or vision. Graph­ic design: P.A. based on Palmer (1999)

The chart in Fig. 110 shows that see­ing of a col­or, e.g. orange, is asso­ci­at­ed with high activ­i­ty of recep­tors that are sen­si­tive to long wave­lengths of vis­i­ble light, much small­er activ­i­ty to aver­age wave­lengths of light and trace activ­i­ty to the short­est light waves. In turn, vision of, e.g. green, is asso­ci­at­ed with the max­i­mum activ­i­ty of recep­tors that are sen­si­tive to the aver­age wave­length of light and the slight activ­i­ty of the oth­er two types of pho­tore­cep­tors. The phe­nom­e­non of col­or mix­ing as a result of cumu­la­tion of light waves of dif­fer­ent wave­length is called addi­tive col­or syn­the­sis and, like the con­cept of sub­trac­tive col­or syn­the­sis, it allows us to pre­dict what col­or will result from a com­bi­na­tion of two or three light com­po­nents of dif­fer­ent col­or (Fig. 111).

Fig­ure 111. The phe­nom­e­non of addi­tive mix­ing of light waves of dif­fer­ent length, cor­re­spond­ing to red, green and blue. Graph­ic design: P.F.

This time, the impres­sion of the col­or yel­low results from mix­ing of two streams of light with lengths cor­re­spond­ing to red and green. Magen­ta is the col­or result­ing from two streams of light, whose length cor­re­sponds to red and blue, while the impres­sion of cyan is cre­at­ed by illu­mi­nat­ing the sur­face with light of the length cor­re­spond­ing to blue and green.

The con­cept of three types of recep­tors sen­si­tive to light waves of dif­fer­ent lengths devel­oped by Young and Helmholtz was empir­i­cal­ly con­firmed in 1956 by Gun­nar Svaetichin. He showed that in the reti­na of fish there are actu­al­ly three types of pho­tore­cep­tors (cones) that are par­tic­u­lar­ly sen­si­tive to three dif­fer­ent wave­lengths of light, rough­ly cor­re­spond­ing to blue, green and red — the col­ors pos­tu­lat­ed by Young and Helmholtz. A few years lat­er, sim­i­lar results were obtained by bio­physi­cists from Johns Hop­kins Uni­ver­si­ty (Marks, Dobelle and Mac­Ni­chol, 1964) dur­ing the stud­ies on pho­tore­cep­tors found in the reti­na of mon­keys, and, in the same year, biol­o­gists from Har­vard Uni­ver­si­ty (Brown and Wald, 1964) — in ref­er­ence to pho­tore­cep­tors locat­ed in the human retina.

R‑, G- and B- cones or L‑, M- and S‑cones?

The actu­al inten­si­ty ranges of reac­tion of the reac­tion of three hypo­thet­i­cal types of cones to elec­tro­mag­net­ic wave in the vis­i­ble light spec­trum are illus­trat­ed in Fig. 112. The chart also shows the ranges of reac­tiv­i­ty of rods, i.e. pho­tore­cep­tors active in the con­di­tions of reduced bright­ness. By the way, nei­ther Young nor Helmholtz sus­pect­ed that there could be two types of pho­tore­cep­tors in the eye, one of which is sen­si­tive to intense day­light and the oth­er is active at night. Basi­cal­ly, their con­cept only con­cerned the first type of recep­tors, i.e. cones. Max­i­mum sen­si­tiv­i­ty of rods falls on a vis­i­ble wave length of about 500 nm and a rel­a­tive­ly low inten­si­ty (almost three times low­er than the inten­si­ty of light to which the cones respond). In turn, in the group of cones there are three types of pho­tore­cep­tors that have already been men­tioned, sen­si­tive to dif­fer­ent wave­lengths of light, which are inter­pret­ed by the brain as col­ors: red, green and blue.

Fig­ure 112. Rela­tion­ships between the length of the elec­tro­mag­net­ic wave and the strength of the rods’ reac­tion (dashed line) and three groups of cones, respond­ing to light of a length cor­re­spond­ing to red ®, green (G) and blue (B) col­or. Graph­ic design: P.F.

R‑cones react most inten­sive­ly to the vis­i­ble light wave with a length of about 560 nm. As of all the pho­tore­cep­tors they are the ones that are sen­si­tive to the longest elec­tro­mag­net­ic wave in the vis­i­ble light spec­trum, they are also called L‑type cones (L for long). This wave­length does not quite match the red col­or, but the impres­sion of red is actu­al­ly cre­at­ed as a result of this pho­tore­cep­tor only in a slight­ly more com­pli­cat­ed way than described by Young-Helmholtz the­o­ry. This will be dis­cussed lat­er. G‑type cones are the ones that most inten­sive­ly react to the elec­tro­mag­net­ic wave with a length of about 530 nm. Since their max­i­mum sen­si­tiv­i­ty falls on a wave of vis­i­ble light with an inter­me­di­ate length between long and short, they are also marked with the let­ter M (mid­dle). In turn, cone of type B or, in oth­er words, type S (short) react most inten­sive­ly to blue light with a length of about 420 nm, because their max­i­mum sen­si­tiv­i­ty is asso­ci­at­ed with the short­est wave­lengths of vis­i­ble light (Matthews, 2000).

Cones of type R, G and B are not equal­ly rep­re­sent­ed in the reti­na. The fewest are the cones react­ing to the short­est elec­tro­mag­net­ic wave in the vis­i­ble light spec­trum, i.e. blue. They con­sti­tute only 5–6% of approx. 5 mil­lion cones. This pro­por­tion is rel­a­tive­ly con­stant in all peo­ple. Fur­ther­more, cones of type B are not in fovea, and occur very irreg­u­lar­ly in the rest of the reti­na (Hofer, Car­roll and Williams, 2009). We do not real­ly know the rea­sons, but it does not inter­fere with nor­mal per­cep­tion of the blue color.

On the oth­er hand, cones of type R, which react most inten­sive­ly to long wave­lengths of vis­i­ble light, are usu­al­ly much numer­ous than cones of type G, which react most inten­sive­ly to the medi­um wave­length of vis­i­ble light, although it turns out that there are sig­nif­i­cant dif­fer­ences in their pro­por­tion in dif­fer­ent peo­ple (Hofer, Car­roll, Neitz and Williams, 2005). While main­tain­ing the nor­mal abil­i­ty to dif­fer­en­ti­ate red and green, in some peo­ple the ratio of cones of type R to type G can be from 1.9:1 to 16.5:1, but there are also those in which it is invert­ed, e.g. 0, 37:1 (Fig. 113). Any­how, only R- and G‑cones are found in the fovea.

Fig­ure 113. A frag­ment of the reti­na at a dis­tance of about  1 ° from the fovea with col­ored cones sen­si­tive to long (red), medi­um (green) and short (blue) light waves, reg­is­tered in three dif­fer­ent peo­ple. The ratio of R‑cones to G‑cones is: A — 0.37:1, B — 1.9:1 and C — 16.5:1, respec­tive­ly. Graph­ic design: P.F. based on Hofer, Car­roll, Neitz and Williams (2005)

The range of color vision due to the distribution of cones on the retina

Since dif­fer­ent types of cones are so uneven­ly dis­trib­uted over the sur­face of the reti­na, it is worth real­is­ing how this relates to the range of col­or vision. Let us start by deter­min­ing how many cones are found in dif­fer­ent parts of the reti­na of the human eye. Chris­tine A. Cur­cio and her team have been deal­ing with this issue for over 30 years (1987; 1990).

In Fig.  14 in the chap­ter devot­ed to the image con­tent analy­sis sys­tem, I pre­sent­ed a schemat­ic dis­tri­b­u­tion of cones and rods depend­ing on the dis­tance from the fovea. In turn, in Fig. 114 in this chap­ter, we can see how many more or less cones there are in dif­fer­ent parts of the reti­na. It can be seen at first glance that the dis­pro­por­tions are huge: from 199 thou­sands per mm2 in the fovea up to 2 thou­sands on the periph­ery of the reti­na. But the inci­dence of cones varies great­ly also on the rest of the surface.

Fig­ure 114. Dis­tri­b­u­tion of cones on the sur­face of the reti­na in thou­sands per 1 mm². A — reti­nal sur­face in the range of 0° (fovea) to 90° (the black oval on the right side of the fovea is a blind spot) and B — sur­face of the fovea in the range of 0° to 2°. Graph­ic design: P.F. based on Cur­cio, Sloan, Pack­er, Hen­drick­son, Kali­na (1987)

The quan­ti­ta­tive dis­tri­b­u­tion of cones, shown in Fig. 114 tells us some­thing about res­o­lu­tion that trans­lates into clar­i­ty of vision in good light­ing con­di­tions. How­ev­er, it says noth­ing about col­or vision, and cones R, G and B are also dis­trib­uted uneven­ly on the reti­na. This caus­es some prob­lems with accu­rate recog­ni­tion of indi­vid­ual col­ors depend­ing on how far from the fovea a light wave of a cer­tain length is pro­ject­ed onto the reti­na. These rela­tion­ships for the right eye are pre­sent­ed in Fig. 115 (the image for the left eye is symmetrical).

Fig­ure 115. The col­or recog­ni­tion field in the right eye due to the dis­tri­b­u­tion of cones of type R, G and B on the reti­na. Graph­ic design: P.F. based on Boff and Lin­coln (1988)

First of all, it turns out that at a dis­tance greater than 45o from the fovea, recog­ni­tion of almost all col­ors is sig­nif­i­cant­ly reduced. We can also state that blue and yel­low are more accu­rate­ly recog­nised than green and red. The lim­its of vis­i­bil­i­ty of indi­vid­ual col­ors are slight­ly wider. Eileen G. Anc­man (1991) deter­mined that above 83.1o the blue col­or is basi­cal­ly no longer vis­i­ble, above 76.3o the red col­or dis­ap­pears, and above 74.3o the impres­sion of green dis­ap­pears com­plete­ly. Nev­er­the­less, prob­lems with their accu­rate recog­ni­tion start much earlier.

Let’s try to trans­late this data into the sit­u­a­tion of view­ing a large paint­ing, e.g. The Bat­tle of Grun­wald by Jan Mate­jko (Fig. 116). The paint­ing is indeed impres­sive: 4.26 m high and 9.87 m wide. Using the for­mu­las that we used to esti­mate the range of sharp vision, it is very easy to deter­mine how many cen­time­ters fall with­in 15o, 30o, 45o, and 60o of the field of vision angle, if we view the paint­ing from a dis­tance of, e.g. 5 m. And so, 15o cor­re­sponds to approx. 1.3 m [2 x tg (15o / 2) x 5 m], 30o – 2.7 m, 45o – 4.1 m and 60o – 5.8 m. Now we only have to trans­fer these mea­sures to the paint­ing and dis­cov­er sev­er­al reg­u­lar­i­ties to our surprise.

Fig­ure 116. Jan Mate­jko, The Bat­tle of Grun­wald (1875–1878). Nation­al Muse­um in War­saw, War­saw [4.26 x 9.87 m]

Let’s sup­pose that we are look­ing at the cord of Mark­war von Salzbach, Bran­den­burg com­man­der on the right side of the paint­ing, approx. 9 m from its left edge. When asked sud­den­ly about the col­or of zhu­pan of Vytau­tas (about 5.5 m from the left) or about the col­or of the hood of a war­rior stand­ing with an axe by the Grand Mas­ter of the Teu­ton­ic Knights (about 3.5 m from the left), we can have a seri­ous prob­lem, because those are sit­u­at­ed on the bor­der of accu­rate recog­ni­tion of red col­or. In turn, focus­ing on the left side of the paint­ing, where against the back­ground of dis­persed red robes one can see the fig­ure of Jakub Skar­bek from Góra, attack­ing Casimir V, the Duke of Szczecin, depict­ed in a hel­met with a pow­er­ful plume, we would rather not cor­rect­ly recog­nise cor­rect­ly the col­or of the zhu­pan, but even more the col­or of green robes of the com­man­der of Tuchola, Hein­rich von Schwel­born, dying from the sword of the Czech knight Jan Žiž­ka (about 7.5 m).

By the way, it is worth recall­ing that Jan Mate­jko had a sig­nif­i­cant vision defect from birth, which prac­ti­cal­ly pre­vent­ed him from see­ing his works in full. He watched them while paint­ing only from a per­spec­tive of sev­er­al cen­time­tres, which sig­nif­i­cant­ly nar­rowed his field of col­or recog­ni­tion! He could see his paint­ings only in his imagination.

The opponent-process theory by Ewald Hering

After a short vis­it to the Nation­al Muse­um in War­saw, let’s stay in 1878, when Jan Mate­jko com­plet­ed The Bat­tle of Grun­wald. It was then that Ewald Her­ing, a Ger­man phys­i­ol­o­gist, pub­lished the the­o­ry of col­or vision, based on oppo­nent process the­o­ry. It was an alter­na­tive con­cept to the then “applic­a­ble” Young-Helmholtz trichro­mat­ic the­o­ry of col­or vision. Her­ing also based his the­o­ry on the results of New­ton’s research on the dis­per­sion of white light into var­ie­gat­ed com­po­nents. How­ev­er, he reject­ed the hypoth­e­sis of a tri­ad of recep­tors sen­si­tive to red, green and blue, claim­ing that the idea does not explain the sub­jec­tive impres­sion of yel­low, “as pure and pri­mal as red, green and blue.

Her­ing noticed that as a result of mix­ing of some col­ors in dif­fer­ent pro­por­tions, new shades are cre­at­ed, which con­sti­tute tonal tran­si­tion between them, and mix­ing oth­er col­ors does not give such effects. For exam­ple, shades of orange are a tran­si­tion­al form between red and yel­low. Just like shades of pur­ple are sat­u­rat­ed with red and blue to vary­ing degrees. How­ev­er, there are also col­or com­bi­na­tions that do not cre­ate tran­si­tion­al forms, but com­plete­ly new qual­i­ties. For exam­ple, adding yel­low and blue cre­ates a com­plete­ly new qual­i­ty, i.e. green. In turn, the com­bi­na­tion of green and red gives the col­or of mud. On the basis of the con­sid­er­a­tions con­cern­ing the sub­jec­tive expe­ri­ence of see­ing col­ors, Her­ing came to the con­clu­sion that the reti­na should con­tain cells that form pairs respon­si­ble for the per­cep­tion of oppos­ing col­ors: red and green as well as blue and yel­low, name­ly those, which cre­ate new qual­i­ties, not inter­me­di­ate forms, when they mix together. 

For­mu­lat­ing his con­cept, Her­ing drew atten­tion to New­ton’s attempts to geo­metri­cize the col­or space, which result­ed in his descrip­tion of the entire spec­trum of vis­i­ble light in the form of a col­or wheel (Fig. 117 A). Young and Helmholtz ignored this ele­ment of New­ton’s con­cept and stopped at a lin­ear descrip­tion of the col­or sequence in accor­dance with the increas­ing length of the cor­re­spond­ing light waves (see Fig. 110).

Fig­ure 117 A. Col­or wheel as pro­posed by Isaak New­ton in Optics (1704). Graph­ic design: P.F.

Her­ing, in turn, was very fond of the idea of ​​pre­sent­ing col­ors in the form of a cir­cle, because he noticed that their dis­tri­b­u­tion per­fect­ly cor­re­sponds to the red and green, yel­low and blue oppo­si­tions noticed by him. Lying on oppo­site sides of the cir­cle, oppos­ing col­ors do not cre­ate tran­si­tion­al forms with each oth­er; they cre­ate tran­si­tion­al forms with col­ors belong­ing to the sec­ond pair. Her­ing pre­sent­ed New­ton’s col­or dis­tri­b­u­tion on the wheel in a slight­ly more schemat­ic form, sug­gest­ing that the four basic col­ors — red, yel­low, green and blue — divide the col­or wheel into four parts, occu­py­ing a dom­i­nant posi­tion with­in each of them (Fig. 117 B).

Fig­ure 117 B. Col­or wheel in the ver­sion pro­posed by Ewald Her­ing in On the The­o­ry of Sen­si­bil­i­ty to Light (1878). Graph­ic design: P.F.

If Ewald Her­ing lived to be 128, he could feel ful­ly sat­is­fied after read­ing the arti­cle by Roger N. Shep­ard (1962), devot­ed to an attempt to empir­i­cal­ly repro­duce the men­tal sim­i­lar­i­ty struc­ture of 14 col­ors select­ed in the wave­length range from 434 nm to 674 nm by Gös­ta Ekman (1954). The task of the indi­vid­u­als test­ed in the Shep­ard’s exper­i­ment was to com­pare all these col­ors in pairs. It turned out that from a psy­cho­me­t­ric point of view, the two-dimen­sion­al solu­tion is com­plete­ly sat­is­fy­ing and almost exact­ly repro­duces the New­ton­ian struc­ture of the rela­tion­ship between col­ors. And most impor­tant­ly, the test result also con­firmed the col­or oppo­si­tions sug­gest­ed by Her­ing (Fig. 118).

Fig­ure 118. Col­or wheel repro­duced with mul­ti­di­men­sion­al scal­ing based on the results of empir­i­cal research. Graph­ic design: P.F. based on Shep­ard (1962)

Roger N. Shep­ard and Lynn A. Coop­er (1992) once again con­duct­ed anal­o­gous stud­ies in the group of peo­ple with nor­mal vision and in the group of peo­ple suf­fer­ing from deuter­a­nopia (inabil­i­ty to recog­nise green col­or) and protanopia (inabil­i­ty to recog­nise red col­or), as well as peo­ple blind since birth. They also reduced the num­ber of com­pared col­ors to nine and pre­sent­ed stim­uli both in visu­al form (col­ored card­boards) and in the form of names. It turned out that the dis­tri­b­u­tion of col­ors through names was iden­ti­cal in all groups, except for those who were blind from birth. How­ev­er, sig­nif­i­cant dif­fer­ences between the groups occurred when the col­ors were pre­sent­ed in a visu­al form.

Anoth­er sup­port for the con­cept of Her­ing’s oppos­ing col­ors is the expe­ri­ence of after­im­age (Fig.  119). It is a type of visu­al illu­sion result­ing from long-term stim­u­la­tion of dif­fer­ent parts of the reti­na with dif­fer­ent colors.

Fig­ure 119. Exam­ple of an after­im­age of oppo­si­tion col­ors. Expla­na­tions in the text. Graph­ic design: P.F.

To evoke the after­im­age, focus your eyes for 30 sec­onds on the black point lying at the junc­tion of four col­ored squares on the left. Then move to the right pan­el and focus on the cen­tre point as well. Unex­pect­ed­ly, four squares in com­ple­men­tary col­ors will appear on the white plane, i.e. in the place of the col­or yel­low we will see a col­or sim­i­lar to blue, in the place of red — green, etc. Accord­ing to Her­ing, the fact that after­im­ages use oppos­ing col­ors is anoth­er sup­port for his con­cept of col­or vision based on oppos­ing processes.

In addi­tion to pos­tu­lat­ing the two mech­a­nisms under­ly­ing the view of oppos­ing col­ors, Her­ing also drew atten­tion to the third oppo­si­tion, this time in terms of lumi­nance, i.e. bright­ness of light­ing. He claimed that the loss of col­or vision, e.g. in low-light con­di­tions, is inde­pen­dent of the abil­i­ty to dis­tin­guish degrees of image bright­ness in the range of white to black. Accord­ing to Her­ing, the third type of recep­tor must be respon­si­ble for cod­ing of bright­ness, and there­fore also the third oppos­ing mech­a­nism, inde­pen­dent of those that under­lie the per­cep­tion of oppos­ing colors.

Her­ing locat­ed all three per­cep­tive mech­a­nisms in recep­tors locat­ed in the reti­na of the eye. Reject­ing the divi­sion of recep­tors into three types, pos­tu­lat­ed by Young and Helmholtz, spe­cialised in the per­cep­tion of red, green and blue, he also pro­posed a triple divi­sion of recep­tors, but accord­ing to com­plete­ly new cri­te­ria. He assumed that there are recep­tors in the reti­na respon­si­ble for encod­ing red and green, yel­low and blue as well as white and black (Fig. 120).

Fig­ure 120. Hypo­thet­i­cal pho­tore­cep­tors respon­si­ble for col­or vision, pos­tu­lat­ed by Edwald Her­ing as part of his the­o­ry of oppos­ing process­es. Graph­ic design: P.A. based on Palmer (1999)

In his opin­ion, the prin­ci­ple of oper­a­tion of all types of recep­tors is based on the recon­struc­tion or decay of the same chem­i­cal inside them. For exam­ple, an increase in red-green recep­tors, while zero­ing their state in blue-yel­low and black-and-white recep­tors would be the basis for the sen­sa­tion of red. If this was accom­pa­nied by decay at the same time, i.e. a decrease in the amount of chem­i­cal sub­stance in the blue-yel­low recep­tors, then we would have the impres­sion of vio­let or pur­ple, etc. The direct rea­son for the decay or accu­mu­la­tion of a chem­i­cal sub­stance in indi­vid­ual recep­tors is — accord­ing to Her­ing — light of a cer­tain length.

Müller-Hurvich-Jameson zone theory of color vision

Until the 1930s, the Young-Helmholtz trichro­mat­ic the­o­ry and the Her­ing’s the­o­ry of oppos­ing process­es were treat­ed as two con­tra­dic­to­ry con­cepts explain­ing the mech­a­nism of col­or vision. George Elias Müller pro­posed a solu­tion to this the­o­ret­i­cal dead­lock (1930 — after: Klein, 2010). He pre­sent­ed the con­cept of col­or cod­ing zones on the visu­al path­way (zone the­o­ry). Accord­ing to his sug­ges­tion, the first zone is asso­ci­at­ed with the reac­tiv­i­ty of reti­nal pho­tore­cep­tors sen­si­tive to red, green and blue light. The sec­ond zone is the visu­al path­way on the seg­ment of the reti­na of the eye — the cere­bral cor­tex and the third, asso­ci­at­ed with the activ­i­ty of cor­ti­cal struc­tures. Müller’s intu­itions proved to be extreme­ly accurate.

The first zone, cov­er­ing struc­tures inside the reti­na of the eye, includ­ing R, G and B recep­tors, has been quite decent­ly described by Young and Helmholtz as part of the trichro­mat­ic the­o­ry of col­or vision. The sec­ond zone is asso­ci­at­ed with the struc­ture of con­nec­tions and activ­i­ty of gan­glion cells, which out­put the sig­nal from the eye reti­na to LGN and fur­ther to the cor­ti­cal struc­tures of the brain. Data from pho­tore­cep­tors is ordered in this sec­tion exact­ly as Her­ing described it in his the­o­ry of oppos­ing process­es. Final­ly, a third col­or cod­ing zone, cov­er­ing the cor­ti­cal areas of the brain, rang­ing from the blob areas in V1, to the V4 struc­ture in the tem­po­ral lobes which was described by Semir Zeki (1973; 1993; 1999; 2003), and by Edwin H. Land and John J. McCann (1971) in the the­o­ry of col­or vision sta­bil­i­ty retinex.

The con­cept of col­or vision zones pro­posed by Müller in the 1930s was large­ly con­firmed empir­i­cal­ly in phys­i­o­log­i­cal and neu­ro­log­i­cal stud­ies only thir­ty years lat­er. Before this hap­pened, how­ev­er, Leo Hur­vich and Dorthea Jame­son (1957), based on Müller’s ideas, con­duct­ed empir­i­cal research whose pur­pose was to ver­i­fy the the­o­ry of oppos­ing process­es by estab­lish­ing the rela­tion­ship between per­ceived col­ors and the cor­re­spond­ing wave­lengths of vis­i­ble light.

Hur­vich and Jame­son attempt­ed to quan­ti­fy the course of oppos­ing process­es using psy­chophys­i­cal meth­ods. Just as Müller, they assumed that in the sec­ond zone of the visu­al path­way data from three types of pho­tore­cep­tors sen­si­tive to red, green and blue light gen­er­ate impres­sions in the range of four pri­ma­ry col­ors that are mutu­al­ly exclu­sive in pairs (red vs green and blue vs yel­low). If this is the case, then in the full spec­trum of vis­i­ble light it is pos­si­ble to empir­i­cal­ly deter­mine the elec­tro­mag­net­ic wave­lengths that cor­re­spond to both the max­i­mum and min­i­mum val­ues ​​of col­or sat­u­ra­tion in both pairs. By anal­o­gy, we can also deter­mine the wave­lengths that are respon­si­ble for dif­fer­ent degrees of bright­ness in the dimen­sion: white vs black.

Per­sons par­tic­i­pat­ing in the exper­i­ments car­ried out by Hur­vich and Jame­son made a series of assess­ments regard­ing the con­tent of a giv­en col­or, e.g. blue, in all col­ors from the entire spec­trum of vis­i­ble light. A sim­i­lar task con­cerned the oth­er three col­ors: yel­low, red and green.

On the basis of the col­lect­ed data, Hur­vich and Jame­son described not only the rela­tion­ship between the length of the light wave and the per­ceived inten­si­ty of each of the four col­ors, but above all deter­mined the elec­tro­mag­net­ic wave­lengths that cor­re­spond­ed to the zero inten­si­ty of the oppos­ing col­ors in pairs. Sim­i­lar­ly, the rela­tion­ship between the wave­length of the light and the per­ceived bright­ness of light has been described (see Fig. 121).

Fig­ure 121. Rela­tion­ships between vis­i­ble wave­length and sat­u­ra­tion of red, green, yel­low and blue, with marked wave­lengths that cor­re­spond to the max­i­mum and min­i­mum (zero) val­ues ​​of sat­u­ra­tion of oppo­si­tion col­ors. The dashed line indi­cates the per­ceived bright­ness of indi­vid­ual col­ors. Graph­ic design: P.F. based on Hur­vich and Jame­son (1957)

Based on empir­i­cal­ly deter­mined func­tions shown in Fig. 121 it can be seen that four basic col­ors are per­ceived in two oppo­si­tions: blue-yel­low and red-green. The impres­sion of a pure blue col­or is gen­er­at­ed by a vis­i­ble wave of 445 nm and a wave of 554 nm — yel­low. In turn, the 492 nm wave­length gives the impres­sion of a green col­or, which is com­plete­ly devoid of both yel­low and blue. This is a neu­tral tran­si­tion point sug­gest­ed by Her­ing and Müller between these two oppos­ing colors.

The sit­u­a­tion regard­ing the oppo­si­tion of red and green is slight­ly more com­plex. The impres­sion of pure green cor­re­sponds to a wave­length of 524 nm, while the impres­sion of pure red is a com­po­nent of two col­ors: red, with a vis­i­ble light length of 610 nm and pur­ple, with a length of 440 nm. Col­ors com­plete­ly devoid of red and green, i.e. neu­tral col­ors for this oppo­si­tion, are blue (472 nm) and yel­low (573 nm).

The last of the func­tions pre­sent­ed in Fig. 121 is a bright­ness func­tion. Based on the course of this func­tion, we can con­clude that the max­i­mum bright­ness is relat­ed to the wave­length of the light cor­re­spond­ing to pure yel­low and it decreas­es sym­met­ri­cal­ly with the elon­ga­tion and short­en­ing of the elec­tro­mag­net­ic wave­length towards the ends of the vis­i­ble light spectrum.

The results of research con­duct­ed by Hur­vich and Jame­son con­firmed the the­o­ry of Her­ing’s oppos­ing process­es in psy­chophys­i­cal stud­ies. The issue of neu­ro­phys­i­o­log­i­cal foun­da­tions of these process­es remains to be clar­i­fied, i.e. we need the answer to the ques­tion how it hap­pens that on the basis of data from three recep­tors sen­si­tive to red, green and blue light, the impres­sion of yel­low col­or vision and the impres­sion of dif­fer­ent degrees of bright­ness arise. Appro­pri­ate the­o­ry was already coming.

From stimulation of R‑, G- and B‑cones to opponent-color vision

The dis­cov­ery of the rela­tion­ship between the sen­si­tiv­i­ty of three types of cones in the reti­na of the eye to the wave­length of vis­i­ble light and the vision of oppos­ing col­ors, described in the the­o­ry of oppos­ing process­es, was made on the basis of the results of research on the reac­tions of gan­glion cells that dis­charge the sig­nal from the reti­na of the eye to the lat­er­al genic­u­late nucle­us (LGN). As we remem­ber, LGN is locat­ed about halfway between the reti­nas and the cere­bral cortex.

First, Edward F. Mac­Ni­chol and Gun­nar Svaetichin (1958), con­duct­ing research on gold­fish, and a few years lat­er Rus­sell L. de Val­ois (1960) in sim­i­lar stud­ies con­duct­ed on mon­keys, dis­cov­ered that the activ­i­ty of some gan­glion cells in LGN sig­nif­i­cant­ly increas­es when light of the length cor­re­spond­ing to the col­or, e.g. red, falls on the reti­na, while the reac­tion of these cells decreas­es under the influ­ence of green light. In turn, oth­er cells react in the oppo­site way: green light clear­ly stim­u­lates them, but under the influ­ence of red light their activ­i­ty decreas­es. What is more, gan­glion cells have also been found to react inten­sive­ly to blue light, and their activ­i­ty decreas­es marked­ly when the light turns yel­low. And final­ly, there are cells that react inten­sive­ly to yel­low light while their activ­i­ty is slowed down under the influ­ence of blue light. These stud­ies also iden­ti­fied cells that respond to bright light and do not respond to weak light, and cells that acti­vate under low light and do not respond to bright light.

Stephen E. Palmer (1999) defines the reac­tions of all these types of gan­glion cells as a neu­ronal imple­men­ta­tion of Her­ing’s oppo­nent process­es the­o­ry. The basis of action of these cells is the mech­a­nism of activ­i­ty of bipo­lar and gan­glion cells form­ing the so-called ON and OFF chan­nels. Let us remem­ber that pho­tore­cep­tors are inter­con­nect­ed in cir­cu­lar recep­tive fields of hor­i­zon­tal cells on the major­i­ty of the reti­nal sur­face. Pho­tore­cep­tors locat­ed in the cen­tral parts of these fields addi­tion­al­ly con­nect with bipo­lar ON or OFF cells, and those in turn with gan­glion cells also of ON or OFF type, respec­tive­ly. The mech­a­nism for trans­mit­ting data on the length of vis­i­ble light in the spec­trum of oppo­nent col­ors via these chan­nels is anal­o­gous to the mech­a­nism for cod­ing of the bright­ness of light falling on the recep­tive field, described in one of the pre­vi­ous chapters.

If the recep­tion area of ​​a gan­glion cell, e.g. ON type, is organ­ised so that its cen­tral part reacts to a wave of light, cor­re­spond­ing to the green col­or, then when it is illu­mi­nat­ed with such light, infor­ma­tion about the pres­ence of green is trans­mit­ted to LGN, and then to V1, and infor­ma­tion about the pres­ence of red is inhib­it­ed. How­ev­er, if the cen­tral part of the gan­glion cell recep­tive field is of the OFF type, then the infor­ma­tion about the illu­mi­na­tion of this part of the reti­na with light cor­re­spond­ing to the green col­or is inhib­it­ed, and the infor­ma­tion about the pres­ence of the red col­or is trans­ferred. ON and OFF chan­nels work sim­i­lar­ly in rela­tion to the pair of blue and yel­low col­ors (Daw, 2008).

What remains to be clar­i­fied is the rela­tion­ship between the reac­tion of gan­glion cells encod­ing one of the four pri­ma­ry col­ors and the reac­tiv­i­ty of R, G and B pho­tore­cep­tors. This issue has not been explained to this day, but Rus­sell L. de Val­ois, Israel Abramov and Ger­ald H. Jacobs (1966) pro­posed a solu­tion that has not yet been chal­lenged. Its graph­ic illus­tra­tion is shown in Fig. 122.

Fig­ure 122. Scheme of com­bi­na­tions of R‑, G- and B- pho­tore­cep­tor stim­u­la­tions with­in the recep­tive fields of gan­glion cells encod­ing two pairs of oppo­nent col­ors and lumi­nance. Graph­ic design: P.A. based on de Val­ois, Abramov and Jacobs (1966)

Accord­ing to the scheme shown in Fig. 122, the impres­sion of red may arise as a result of:

  • acti­va­tion of the ON chan­nel for the recep­tive field, which in the cen­tral part receives data on the lev­el of stim­u­la­tion of R- and B‑cones or
  • acti­va­tion of the OFF-chan­nel for the recep­tive field, which receives data on the lev­el of stim­u­la­tion of R- and B‑cones from the areola.

In turn, the impres­sion of green can arise as a result of:

  • acti­va­tion of the ON chan­nel for the recep­tive field, which in the cen­tral part receives data on the lev­el of stim­u­la­tion of G‑cones or
  • acti­va­tion of the OFF-chan­nel for the recep­tive field, which receives data on the lev­el of stim­u­la­tion of G‑cones from the areola.

The impres­sion of yel­low also aris­es in one of two situations:

  • acti­va­tion of the ON chan­nel for the recep­tive field, which in the cen­tral part receives data on the lev­el of stim­u­la­tion of R- and G‑cones, or
  • acti­va­tion of the OFF-chan­nel for the recep­tive field, which receives data on the lev­el of stim­u­la­tion of R- and G‑cones from the areola.

Also the impres­sion of blue can arise in two situations:

  • acti­va­tion of the ON chan­nel for the recep­tive field, which in the cen­tral part receives data on the lev­el of stim­u­la­tion of B‑cones or
  • acti­va­tion of the OFF-chan­nel for the recep­tive field, which receives data on the lev­el of stim­u­la­tion of B‑cones from the areola.

Obvi­ous­ly, the impres­sion of any col­or aris­es only when all the sub­sys­tems respon­si­ble for sen­so­ry data pro­cess­ing at each stage of visu­al path­way work effi­cient­ly. If not, the final effects of data pro­cess­ing may even dif­fer sig­nif­i­cant­ly from those described in the above schemes.

And final­ly, a reminder on how the bright­ness (lumi­nance) detec­tion mech­a­nism works. The impres­sion of bright­ness can arise as a result of:

  • acti­va­tion of the ON chan­nel for the recep­tive field, which in the cen­tral part receives data on the lev­el of stim­u­la­tion of all types of cones or
  • acti­va­tion of the OFF-chan­nel for the recep­tive field, which receives data on the lev­el of stim­u­la­tion of all types of cones from the areola.

Due to the dif­fer­ent light bright­ness cou­pled with dif­fer­ent elec­tro­mag­net­ic wave­length, the stim­u­la­tion lev­el of R- and G‑cones will have a much greater impact on bright­ness than B‑cones, because their max­i­mum sen­si­tiv­i­ty lev­el – 560 and 530 nm respec­tive­ly – coin­cides with such a light wave­length that is the bright­est, i.e. approx. 550 nm (see graph of the bright­ness func­tion in Fig. 121).


In the his­to­ry of art, the his­to­ry of col­or and its the­o­ry is an oscil­la­tion between degra­da­tion to the role of an orna­ment or dec­o­ra­tion and a pro­mo­tion to the basic truth of paint­ing (Melville, 1994)

On Claude Monet’s visual experiments

When equipped with the knowl­edge of col­or vision, can we bet­ter under­stand why painters or pho­tog­ra­phers use col­or in one way or anoth­er and what visu­al effects they achieve? The first exam­ple of a paint­ing that I would like to draw atten­tion to is Impres­sion, Sun­rise by Claude Mon­et (Fig. 123 A).

Fig­ure 123 A. Claude Mon­et, Impres­sion, Sun­rise (1872). Musée Mar­mot­tan Mon­et, Paris, France [48 x 63 cm]

When admir­ing the repro­duc­tion of this paint­ing in the album, Mar­garet Liv­ing­stone (2002) noticed an inter­est­ing phe­nom­e­non. The blood-red sphere of the ris­ing sun, almost impos­ing on the observ­er, ceas­es to arouse inter­est when it is viewed in a mono­chrome ver­sion (Fig 123 B). In her opin­ion, remov­ing col­ors makes the sun kind of fade, although its con­tours are not com­plete­ly invis­i­ble in the mono­chrome version.

Fig­ure 123 B. Claude Mon­et’s paint­ing, Impres­sion. Sun­rise (1872) in mono­chrome ver­sion. Pro­ce­dure: Image/Mode/Lab Col­or [Lumi­nance]

To find out if Mon­et’s paint­ing is real­ly seen dif­fer­ent­ly in the orig­i­nal and mono­chrome ver­sion, we con­duct­ed a study togeth­er with Anna Szpak and Patryc­ja Kot, dur­ing which we showed 19 repro­duc­tions of the col­or­ful Impres­sion to 19 women and 19 men, and its mono­chrome ver­sion — to 26 women and 17 men. All indi­vid­u­als were approx. 22 years old. The images were dis­played on a 23″ Apple Cin­e­ma HD Dis­play (1920 x 1200 pix­els). Dur­ing the exper­i­ment, we record­ed the eye move­ments of the sub­jects using the SMI iView X Hi Speed ​​1250 Hz oculograph.

We expect­ed that inter­est in the area of ​​the sun in a col­or image would be much greater than in the anal­o­gous area in a mono­chrome ver­sion. We also won­dered how much atten­tion the respon­dents would devote to oth­er parts of the pic­ture, e.g. a boat in the fore­ground, which is very clear in both ver­sions of the pic­ture. Espe­cial­ly with regard to the mono­chrome ver­sion, we assumed that it was the dark boat in the fore­ground that strong­ly con­trasts with the bright sea back­ground, should arouse much greater inter­est of the sub­jects than the area of the sun.

The ocu­lo­graph record­ing ful­ly con­firmed Liv­ing­stone’s obser­va­tions and our expec­ta­tions (Fig. 123 C and D). While watch­ing the col­or­ful image, the sub­jects devot­ed most time to sun (Fig. C), while when watch­ing the mono­chrome repro­duc­tion, they were par­tic­u­lar­ly inter­est­ed in the boat in the fore­ground (Fig. D). All dif­fer­ences in times of visu­al fix­a­tion on the same frag­ments of both ver­sions of the paint­ing and com­pared to oth­er frag­ments with­in the same ver­sions are clear and sta­tis­ti­cal­ly sig­nif­i­cant (Tukey’s HSD test; p <0.001).

Fig­ure 123 C. Claude Mon­et’s paint­ing, Impres­sion. Sun­rise (1872) in the col­or­ful (orig­i­nal) ver­sion with an atten­tion map. Elab­o­ra­tion of own research results
Fig­ure 123 D. Claude Mon­et’s paint­ing, Impres­sion. Sun­rise (1872) in the mono­chrome ver­sion with an atten­tion map. Elab­o­ra­tion of own research results

If we assume that the time spent on look­ing at an object is relat­ed to inter­est in it, then it is worth con­sid­er­ing what aroused so much inter­est of the sub­jects in this sharply out­lined red sphere against the back­ground of the morn­ing sky. As a curios­i­ty, it can be added that while watch­ing a paint­ing in a muse­um and star­ing longer into the sun, some peo­ple expe­ri­ence a kind of puls­ing, which prob­a­bly makes them even more inter­est­ed and stops their sight in this place.

Accord­ing to Liv­ing­stone (2002), this weird impres­sion comes from the fact that the visu­al sys­tem, by pro­cess­ing data on the col­or and image bright­ness inde­pen­dent­ly, can­not make the most fun­da­men­tal deci­sion regard­ing the pres­ence or absence of things in the visu­al scene. The part of the visu­al sys­tem that is evo­lu­tion­ar­i­ly much old­er, respon­si­ble for lumi­nance detec­tion, is help­less in the face of a fun­da­men­tal lack of con­trast between adja­cent planes of the sun and sky. In turn, the evo­lu­tion­ar­i­ly younger part sen­si­tive to col­ors reg­is­ters in the paint­ing the pres­ence of a red cir­cle against the blue sky. The result is an amaz­ing eerie puls­ing expe­ri­ence, which is the reac­tion of the lumi­nance detec­tion sys­tem to the dif­fi­cul­ty in deter­min­ing the posi­tion of an object whose pres­ence is report­ed by the col­or detec­tion sys­tem. In this way, we can explain the effect of the insta­bil­i­ty of the image frag­ment built of two adja­cent col­ors of the same bright­ness. In short, the bright­ness of the sun is almost the same as the bright­ness of the back­ground on which it was placed by “uncon­scious neu­ro­sci­en­tist”, Claude Mon­et. And that’s the mag­ic behind this painting.

It is easy to see that the bright­ness of the sun and the sur­round­ing sky is very sim­i­lar. In Fig.  124 A there is an enlarged frag­ment of the Mon­et’s paint­ing show­ing the sun. Below is the same frag­ment in the mono­chrome ver­sion (Fig. 124 B). I esti­mat­ed and applied to it numer­i­cal val­ues ​​cor­re­spond­ing to the bright­ness of the areas around them. As we can see, the dif­fer­ences between the bright­ness inside and out­side the sun are small. It is worth remem­ber­ing that some of the dark­er places inside and on the edges of the sun were not cre­at­ed due to the use of a dark­er shade of red, but are a shad­ow cast by the thick­ly laid paint.

Fig­ure 124 A. Enlarged frag­ment of Claude Mon­et’s paint­ing, Impres­sion. Sun­rise (1872)
Fig­ure 124 B. Enlarged frag­ment of Claude Mon­et’s paint­ing, Impres­sion. Sun­rise (1872) in mono­chrome ver­sion with aver­aged val­ues ​​on the black sat­u­ra­tion scale (0 = white, 100 = black), esti­mat­ed for forty-eight areas (101 x 101 pix­els) lying inside and out­side the sun

Analy­sis of the Mon­et’s paint­ing reveals one more very impor­tant fea­ture of the visu­al sys­tem, which uses the ben­e­fits of col­or reg­is­ter­ing. Since the observer’s lumi­nance sys­tem has dif­fi­cul­ty iden­ti­fy­ing the sun in Mon­et’s paint­ing, and yet sees it, it can only mean that the col­or, like chiaroscuro, also codes the shape of the objects we look at. To con­firm it, I iso­lat­ed from the orig­i­nal ver­sion of the Mon­et’s paint­ing two paint­ings com­posed only of oppos­ing col­ors: red and green, and yel­low and blue, set­ting their lumi­nance at the same, medi­um lev­el. This allowed to deter­mine how much red, green, blue and yel­low is con­tained in the col­ors that Mon­et used to paint it. The result­ing images allow us to see the bound­aries between col­or­ful spots, and thus to read the shapes that they code. In any case, these are the pre­dic­tions aris­ing from the the­o­ry of oppo­nent process­es, built on apt intu­itions of Ewald Her­ing. The effect of sim­u­la­tion of oppo­nent process­es is pre­sent­ed in Fig. 125 A‑D.

Fig­ure 125 A. Mon­et’s paint­ing, Impres­sion. Sun­rise, after remov­ing all col­ors except red and green and equal­iz­ing their lumi­nance. Graph­ic design: P.F. Pro­ce­dure: (1) Image/Mode/Lab Col­or, (2) use the Rec­tan­gle Tool [L=50] to over­ride the chan­nels: Light­ness and b
Fig­ure 125 B. Mono­chrome ver­sion of Mon­et’s paint­ing pre­sent­ed in Fig. 123 A, illus­trat­ing the dis­tri­b­u­tion of red (black) and green (white) sat­u­ra­tion on the sur­face of the paint­ing. Graph­ic design: P.F. Pro­ce­dure: (1) Image/Mode/Lab Col­or, (2) Delete Chan­nel [Light­ness], Delete Chan­nel [b], (3) Image/Auto Con­trast, (4) Image/Adjustments/Invert.
Fig­ure 125 C. Mon­et’s paint­ing, Impres­sion. Sun­rise, after remov­ing all col­ors except yel­low and blue and equal­iz­ing their lumi­nance. Graph­ic design: P.F. Pro­ce­dure: (1) Image/Mode/Lab Col­or, (2) use the Rec­tan­gle Tool [L=50] to over­ride the chan­nels: Light­ness and a
Fig­ure 125 D. Mono­chrome ver­sion of Mon­et’s paint­ing pre­sent­ed in Fig. 123 A, illus­trat­ing the dis­tri­b­u­tion of yel­low (black) and blue (white) sat­u­ra­tion on the sur­face of the paint­ing. Graph­ic design: P.F. Pro­ce­dure: (1) Image/Mode/Lab Col­or, (2) Delete Chan­nel [Light­ness], Delete Chan­nel [a], (3) Image/Auto Con­trast, (4) Image/Adjustments/Invert.

Both pairs of oppos­ing col­ors in Fig. 125 A and C have the same lumi­nance. This caus­es that the bound­aries of col­ors are hard­ly vis­i­ble, but cer­tain­ly the brain has no prob­lems to dif­fer them, if only because they are cod­ed with the help of var­i­ous chan­nels, made of gan­glion cells that deal with the trans­mis­sion of sen­so­ry data deep into the brain. For illus­tra­tive pur­pos­es only, I have also devel­oped both ver­sions of the paint­ing in a high con­trast mono­chrome ver­sion (Fig. 125 B and D). There should be no doubt now that the basis for see­ing the sun in Mon­et’s paint­ing is not the con­trast of bright­ness, but the con­trast of oppos­ing col­ors that clear­ly encode the shape of the sun (see Fig. 124 B vs Fig. 125 B and D).

It is also worth not­ing that while watch­ing a col­or­ful repro­duc­tion of the paint­ing, the sec­ond longest observed object was the dark out­line of the boat in the fore­ground. In the con­text of what I wrote about the diver­gence of infor­ma­tion flow­ing from the col­or chan­nels and the lumi­nance chan­nel, it is dif­fi­cult to deter­mine why the test sub­jects spent so much time look­ing at this frag­ment of the paint­ing. On the one hand, there is no doubt that due to the con­trast of bright­ness it is the clear­est object in the whole paint­ing, which is clear­ly seen in its mono­chrome ver­sion (see Fig. 123 B). From this point of view, we can say that the lumi­nance chan­nel “allows” to unam­bigu­ous­ly iden­ti­fy the object in the paint­ing, due to its high contrast.

On the oth­er hand, how­ev­er, the boat sim­ply dis­ap­pears when the paint­ing is fil­tered through both oppos­ing col­or chan­nels (Fig. 125 A‑D). Could longer look­ing at the boat be anoth­er man­i­fes­ta­tion of the visu­al con­flict between the chan­nel cod­ing the shape based on dif­fer­ences in lumi­nance and based on the diver­si­ty of oppos­ing col­ors, which increas­es the inter­est of observers? Or maybe it is only a man­i­fes­ta­tion of the nat­ur­al reac­tion of the visu­al sys­tem, expressed by an increased con­cen­tra­tion of atten­tion on those ele­ments of the visu­al scene that will most accu­rate­ly recre­ate its sense.

Well, we cer­tain­ly will not be able to dis­pel all doubts imme­di­ate­ly. Yet, it is worth to try to solve the prob­lem of the sun and the rea­sons for spe­cial inter­est in it. Cer­tain­ly, the analysed paint­ing of Mon­et can­not be includ­ed in the set of ful­ly con­trolled exper­i­men­tal stim­uli. To draw sci­en­tif­ic con­clu­sions based on it, a slight­ly more pre­cise mod­el is required.

Sun by Monet under scrutiny

The essence of the so-called Mon­et effect, as I called the analysed phe­nom­e­non, con­sists in the increased inter­est of the observ­er in those areas of the visu­al scene that are char­ac­terised by dis­crep­an­cy between the data encod­ing the shapes of objects there based on lumi­nance and two col­or chan­nels. To paint this frag­ment, Mon­et used red (sun) and blue (sky) with sim­i­lar lumi­nance, equal to about 40 units (see Fig. 124). Fol­low­ing this idea, togeth­er with Piotr Szyj­ka, we con­struct­ed stim­uli and con­duct­ed an exper­i­ment in which an ocu­lo­graph was used (French­man, Szyj­ka, 2011).

The stim­u­lus mate­r­i­al con­sist­ed of 48 boards. On each board there were 24 dif­fer­ent-col­ored stars of the same size [32.3 mm; 125 pix­els]. The stars were arranged ran­dom­ly on a grey back­ground of the same bright­ness [Lab = 50:0:0; RGB = 119:119:119]. Each star had its assigned ran­dom rota­tion angle around its axis. The dis­tance of the cen­tres of indi­vid­ual stars from each oth­er was equal to or greater than the range of the field of vision through the fovea, i.e. at least 52.41 mm; [201.57 pix­els]. In half of the boards all stars had 7 arms, while in the oth­er half one of the 24 stars had 8 arms. On each of these boards it was always an aster­isk with a dif­fer­ent col­or (see an exam­ple of a board with an octag­o­nal star in Fig. 126 A).

Fig. 126 A. Exam­ple board con­tain­ing 24 mul­ti­col­ored stars, among which 1/3 has low­er bright­ness, 1/3 — equal and 1/3 high­er than the back­ground. Among them, there is one octag­o­nal star with the same lumi­nance as the back­ground. Graph­ic design: P.A.
Fig. 126 B. Mono­chrome ver­sion of the board shown in Fig. 126 A, illus­trat­ing the effect of “dis­ap­pear­ing” stars whose bright­ness is iden­ti­cal to the back­ground bright­ness. Pro­ce­dure: Image/Mode/Lab [Light­ness]

Bright­ness and col­or of indi­vid­ual stars were ful­ly con­trolled. We adopt­ed three lev­els of star bright­ness: L = 25 (dark­er than the back­ground), L = 50, the same as the back­ground bright­ness and L = 75 (brighter than the back­ground). Each board had eight stars brighter than the back­ground, 8 stars dark­er than the back­ground, and 8 stars of the same bright­ness as the back­ground. In Fig.  126 B illus­trates the mono­chrome effect cre­at­ed after remov­ing the col­or from the board on the left.

Col­ors of stars were cre­at­ed from a com­bi­na­tion of para­me­ters cor­re­spond­ing to lim­it and zero val­ues ​​on col­or scales: from red to green and from blue to yel­low. As a result, we obtained 8 dif­fer­ent col­ors, which in com­bi­na­tion with three lev­els of lumi­nance gave 24 dif­fer­ent-col­ored and dif­fer­ent-bright­ness stars, which we placed ran­dom­ly on indi­vid­ual boards.

The num­ber of boards pre­sent­ed to the exam­ined indi­vid­u­als result­ed from the fact that we want­ed every col­or of each bright­ness lev­el to occur at least once in the 8‑arm ver­sion. Because the sub­jects’ task was to answer the ques­tion whether the board has an 8‑arm star, half of the boards con­tained it, and the oth­er half did not. Final­ly, we showed all the boards three times, in ran­dom order. We used the pro­ce­dure writ­ten in the e‑Prime soft­ware to present stim­uli and col­lect data on the respons­es of the subjects.

All boards were dis­played on 23″ Apple Cin­e­ma HD Dis­play (1920 by 1200 pix­els). Sub­jects sat at a dis­tance of about 50 cm from the mon­i­tor and react­ed using an Ergodex key­board with vari­able key arrangement.

We exam­ined 23 indi­vid­u­als (13 women and 11 men) aged around 23 years. As I men­tioned, their task was (three times in ran­dom order) to view each board and answer the ques­tion whether the shapes of all stars on the board are the same. Dur­ing the study, we record­ed the eye move­ments of the sub­jects using the SMI iView X Hi Speed ​​1250 Hz ocu­lo­graph. If the test sub­ject claimed that all stars had the same shape, then the next board appeared. If, on the oth­er hand, the sub­ject claimed that not all stars had the same shape, then such a per­son was asked to focus on the star she claimed to be dif­fer­ent. It was an indi­ca­tor of the accu­ra­cy of the response. In addi­tion to ocu­lo­mo­tor behav­iour, we also record­ed reac­tion time.

Although in the study we asked sub­jects to active­ly look for stars with dif­fer­ent shapes, nev­er­the­less, we were pri­mar­i­ly inter­est­ed in how much atten­tion they would devote to search those stars whose lumi­nance was the same as the back­ground bright­ness com­pared to those whose lumi­nance was high­er or low­er than the back­ground. If the Mon­et effect is real, then the time spent look­ing at stars with a bright­ness equal to the back­ground bright­ness should be longer than the time spent look­ing at oth­er stars.

The results of the analy­sis of vari­ance ful­ly con­firmed these assump­tions, Not only did it turn out that the sub­jects looked at stars with a bright­ness equal to the back­ground bright­ness longer than those brighter or dark­er [F (2, 46) = 9.77; p < 0.001; η2 = 0.30; Fig. 127 A], but this rela­tion­ship was also con­firmed in rela­tion to 8‑arm stars with a lumi­nance of 50 units [F(2, 46) = 8.79; p < 0.001; η2 = 0.28]. More­over, it turned out that the Mon­et effect occurs regard­less of the col­or of the star. The impor­tant thing is for its bright­ness to be the same as the back­ground brightness.

Fig­ure 127 A. Rela­tion­ship between the bright­ness of the stars’ col­or and the aver­age time of visu­al fix­a­tion on them. Own research results based on Fran­cuz and Szyj­ka (2011)
Fig­ure 127 B. Rela­tion­ship between the bright­ness of the stars’ col­or and the cor­rect­ness of recog­ni­tion of 8‑arm stars. Own research results based on Fran­cuz and Szyj­ka (2011)

Addi­tion­al­ly, we found that, although the test sub­jects devot­ed a lot of atten­tion to 8‑arm stars with a bright­ness equal to the back­ground bright­ness, nev­er­the­less, regard­less of the col­or of the stars, sub­jects rarely cor­rect­ly recog­nised them as dif­fer­ent from the oth­ers. For some col­ors, the dif­fer­ences turned out to be sta­tis­ti­cal­ly sig­nif­i­cant [e.g.  for the green col­or: F (2. 46) = 5.58, p= 0.007, η2 = 0.19; Fig. 127 B). It looks a bit like solv­ing the prob­lem of the pres­ence of the star on the board due to lumi­nance so strong­ly dom­i­nat­ed sub­jects’ atten­tion so strong­ly that they could not cor­rect­ly state that a giv­en stars has 7, and not 7 arms. Here is an exam­ple of how skil­ful­ly nar­row­ing the observer’s field of atten­tion by giv­ing him a dif­fi­cult per­cep­tu­al task can dis­rupt his nor­mal cog­ni­tive process­es. When the stars are bright and clear­ly stand out from the back­ground, then the prob­a­bil­i­ty of their accu­rate recog­ni­tion increas­es to 0.8, but when their bright­ness becomes the same as the back­ground — recog­ni­tion accu­ra­cy drops to less than 0.4.

How many shapes are in one object?

In the light of the pre­sent­ed results of Mon­et’s paint­ing analy­sis, it is worth con­sid­er­ing what oth­er con­se­quences of dis­crep­an­cies between the shape cod­ed through the lumi­nance chan­nel and the col­or chan­nels can be?

Let’s start with the exam­ple of the paint­ing Apol­lo and two Mus­es, paint­ed in the 18th cen­tu­ry by Pom­peo Giro­lamo Batoni (Fig. 128 A). The paint­ing comes from the col­lec­tion of the Wilanów Palace Muse­um in War­saw. It depicts three mytho­log­i­cal char­ac­ters: Apol­lo, a sym­bol of per­fect divine music — the har­mo­ny of the heav­en­ly spheres and the mus­es look­ing at him — Euterpe and Ura­nia, who embody Music and Astron­o­my (Gutows­ka-Dudek, 2005). This, of course, is the per­spec­tive of an art his­to­ri­an. On the oth­er hand, a lay­man in the field of art his­to­ry, watch­ing this paint­ing, will cer­tain­ly notice a young undressed men accom­pa­nied by girls in equal­ly airy clothes, while the behav­ior of all three char­ac­ters “stopped in the frame” can give at least ambigu­ous impres­sion (French­man, 2012).

Regard­less of the observer’s knowl­edge, we have no doubt about the char­ac­ters and objects depict­ed in the paint­ing. Refer­ring to the clas­sic canons of beau­ty, we can even say that the paint­ing is impec­ca­ble, but it also does not cre­ate ten­sions that held our atten­tion for longer when watch­ing the paint­ing by Mon­et. And this is not sur­pris­ing, because the visu­al sys­tem has very lit­tle puz­zles to solve here. Shapes are cod­ed with both bright­ness and two col­or chan­nels (red-green and blue-yel­low) and con­firm each oth­er almost in 100% (see Fig. 128 C‑D).

Fig­ure 128 A. Pom­peo Giro­lamo Batoni, Apol­lo, and two Mus­es (approx. 1741). Muse­um of King Jan III’s Palace at Wilanów, War­saw [122 x 90 cm]
Fig­ure 128 B. Batoni’s paint­ing, Apol­lo and two Mus­es (approx. 1741) in mono­chrome ver­sion. Devel­oped based on Fig. 128 A. Pro­ce­dure: Image/Mode/Lab Col­or [Lumi­nance]
Fig­ure 128 C. Mono­chrome ver­sion of Batoni’s paint­ing pre­sent­ed in Fig. 128 A, illus­trat­ing the dis­tri­b­u­tion of red (black) and green (white) sat­u­ra­tion on the sur­face of the paint­ing. Graph­ic design: P.F. Pro­ce­dure: (1) Image/Mode/Lab Col­or, (2) Delete Chan­nel [Light­ness], Delete Chan­nel [b], (3) Image/Auto Con­trast, (4) Image/Adjustments/Invert
Fig­ure 128 D. Mono­chrome ver­sion of Batoni’s paint­ing pre­sent­ed in Fig. 128 A, illus­trat­ing the dis­tri­b­u­tion of yel­low (black) and blue (white) sat­u­ra­tion on the sur­face of the paint­ing. Graph­ic design: P.F. Pro­ce­dure: (1) Image/Mode/Lab Col­or, (2) Delete Chan­nel [Light­ness], Delete Chan­nel [a], (3) Image/Auto Con­trast, (4) Image/Adjustments/Invert

To recog­nise the scene depict­ed by Batoni in the pic­ture, it does not real­ly mat­ter which ver­sion of it we see. Sim­i­lar­ly, all three visu­al chan­nels encode the shapes of the char­ac­ters paint­ed on it. This is the most typ­i­cal way in which the visu­al sys­tem reg­is­ters and inter­prets the dis­tri­b­u­tion of light enter­ing the observer’s eye. It is clos­est to every­day expe­ri­ence. In this way, for cen­turies, West­ern Euro­pean painters have tried to cap­ture real­i­ty in paint­ings. It was not until the inven­tion of pho­tog­ra­phy in the first half of the 19th cen­tu­ry that this direc­tion of paint­ing devel­oped deci­sive­ly. As a result, visu­al arts freed them­selves from too nar­row frames of eras and styles and tru­ly explod­ed with rev­o­lu­tion­ary changes with­in the so-called con­tem­po­rary art.

Paint­ing by Claude Mon­et, Impres­sion. Sun­rise is one of the most impor­tant man­i­fes­ta­tions of this rev­o­lu­tion. Mon­et broke with tra­di­tion, show­ing that the goal of visu­al art is not to accu­rate­ly depict real­i­ty, but to cre­ate it. In fact, each inten­tion­al­ly cre­at­ed paint­ing does not so much repro­duce real­i­ty, but presents it trans­formed, just from the fact that the visu­al scene is kept with­in its frames. And it is not only about the con­tent of the per­for­mance, but above all about its form. In any case, this way of think­ing about the paint­ing dom­i­nat­ed almost all con­tem­po­rary art. In this con­text, I would also like to draw atten­tion to oth­er, extreme­ly inter­est­ing paint­ing, which illus­trate var­i­ous types of devi­a­tions from the prin­ci­ple of striv­ing for obses­sive con­fir­ma­tion of con­tent using all three visu­al data analy­sis chan­nels that inter­est us.

Coloring pages by Picasso 

The first exam­ple is almost the sec­ond end of the com­pli­ance scale for shapes encod­ed by the lumi­nance chan­nel and oppo­nent-col­or chan­nels. A beau­ti­ful paint­ing by Pablo Picas­so, Moth­er, and Child, depicts his first wife, Olga Kok­lowa, with their one-year-old son Paul (Fig. 129).

Fig­ure 129. Pablo Picas­so, Moth­er, and Child (1922). Bal­ti­more Muse­um of Art, Bal­ti­more, USA [100 x 80 cm]

It is an excel­lent exam­ple of the non­cha­lance with which the artist filled the con­tours of the drawn fig­ures with col­ors. The con­tour and col­or­ful spots blend freely as if they were not con­nect­ed with each oth­er. The col­ors mere­ly sug­gest some­thing: green means the back­ground, blue — Paul’s blouse, and even their absence means that Olga’s dress is sim­ply brighter than the jack­et. Just 30 years ear­li­er no one would pay atten­tion to such a paint­ing, it would be con­sid­ered a sketch for a paint­ing, and not a fin­ished work. And yet this care­less­ness in the loca­tion of col­or­ful spots does not both­er you, but on the con­trary, it brings the whole scene to life.

After sep­a­rat­ing the draw­ing from the col­ored spots, we can real­ly see how the con­tours of the moth­er and child do not coin­cide with the edges of the col­ored sur­faces (Fig. 130 A and B). If we had not seen Picas­so’s work before, we would prob­a­bly have hard time guess­ing what these spots rep­re­sent. After fil­ter­ing through oppo­nent-col­or chan­nels, we can even bet­ter see the loss of col­or-cod­ed shapes (Fig. 130 C and D). Black and white, ambigu­ous spots resem­ble pro­jec­tion screen on which the mind of the observ­er can cast almost any asso­ci­a­tion. It is prob­a­bly thanks to them that Picas­so’s work takes on new mean­ings and is more interesting.

Fig­ure 130 A. Sil­hou­ette out­lines in Picas­so’s paint­ing, Moth­er and Child, sep­a­rat­ed from col­or spots. Graph­ic design: P.A.
Fig­ure 130 B. Col­or­ful smears on the Picas­so’s paint­ing, Moth­er and Child, sep­a­rat­ed from the con­tours of the char­ac­ters. Graph­ic design: P.A.
Fig­ure 130 C. Mono­chrome ver­sion of Picas­so’s paint­ing pre­sent­ed in Fig. 129, illus­trat­ing the dis­tri­b­u­tion of red (black) and green (white) sat­u­ra­tion on the sur­face of the paint­ing. Graph­ic design: P.F. Pro­ce­dure: (1) Image/Mode/Lab Col­or, (2) Delete Chan­nel [Light­ness], Delete Chan­nel [b], (3) Image/Auto Con­trast, (4) Image/Adjustments/Invert
Fig­ure 130 D. Mono­chrome ver­sion of Picas­so’s paint­ing pre­sent­ed in Fig. 129, illus­trat­ing the dis­tri­b­u­tion of yel­low (black) and blue (white) sat­u­ra­tion on the sur­face of the paint­ing. Graph­ic design: P.F. Pro­ce­dure: (1) Image/Mode/Lab Col­or, (2) Delete Chan­nel [Light­ness], Delete Chan­nel [a], (3) Image/Auto Con­trast, (4) Image/Adjustments/Invert

Henri and Amélie Matisse on the bend of life

Anoth­er exam­ple of an out­stand­ing paint­ing that I would like to draw atten­tion to is Woman with a Hat. Hen­ri Matisse exhib­it­ed this paint­ing in 1905 at the Salon d’Au­tomne in Paris (Fig.  131 A). It depicts his wife, Amélia Parayre (Fig.  131 B), in a sophis­ti­cat­ed hat.

Fig­ure 131 A. Hen­ri Matisse, Woman with a Hat (1904−05). Muse­um of Mod­ern Art, San Fran­cis­co, USA [79.4 x 59.7 cm]
Fig­ure 131 B. The pho­to­graph of Amélia Parayre, Hen­ri Matisse’s wife. Devel­op­ment of pho­tog­ra­phy P.F. based on a frag­ment of a frame from a doc­u­men­tary film made by BBC One with­in the Mod­ern Mas­ters series (2010). Hen­ri Matisse — The King of Color

Matis­se’s paint­ing, just like Mon­et’s, caused a lot of con­fu­sion in the cir­cles of artists and art experts at the time. Before, no one dared to show a paint­ing cre­at­ed in this way. It was a slap in the face for stu­dents and good taste.

But what real­ly annoyed the art crit­ics? What made them call Matisse and his fel­low painters “beasts” (fauves) and the Woman with a Hat “an impres­sive daub”? “Shown at the Paris exhi­bi­tion, it is still a kind of scan­dal to this day. It attacks with sharp con­trasts and a rest­less, bro­ken sur­face. When we final­ly see Amelia Matis­se’s face under an over­whelm­ing extrav­a­gant hat, we are struck by the impres­sion of uncer­tain­ty and depres­sion” — writes Anna Arno (2012) in an arti­cle about Gertrude Stein and her pas­sion for col­lect­ing Matis­se’s works. But is it just a face full of “uncer­tain­ty and depression”?

A glance at Ameli­a’s face under a red hat con­firms Arno’s obser­va­tions regard­ing the mood of the por­trayed mod­el. How­ev­er, when I bal­anced the bright­ness of all the col­ors in the image, and then fil­tered out the pair of oppos­ing col­ors and changed the results of these fil­tra­tions to a mono­chrome form, a shock­ing image appeared in front of my eyes. Here a ghost skull emerged from under a huge hat, with emp­ty eye sock­ets and frag­ments of a rot­ting body. Break­ing the paint­ing into two oppo­nent-col­or chan­nels revealed a much more dra­mat­ic truth about the paint­ed woman than it would appear from a quick look (Fig. 132 A and B).

Fig­ure 132 A. Mono­chrome ver­sion of Matis­se’s paint­ing pre­sent­ed in Fig. 131 A, illus­trat­ing the dis­tri­b­u­tion of red (black) and green (white) sat­u­ra­tion on the sur­face of the paint­ing. Graph­ic design: P.F. Pro­ce­dure: (1) Image/Mode/Lab Col­or, (2) Delete Chan­nel [Light­ness], Delete Chan­nel [b], (3) Image/Auto Con­trast, (4) Image/Adjustments/Invert
Fig­ure 132 B. Mono­chrome ver­sion of Matis­se’s paint­ing pre­sent­ed in Fig. 131 A, illus­trat­ing the dis­tri­b­u­tion of blue (black) and yel­low (white) sat­u­ra­tion on the sur­face of the paint­ing. Graph­ic design: P.F. Pro­ce­dure: (1) Image/Mode/Lab Col­or, (2) Delete Chan­nel [Light­ness], Delete Chan­nel [a], (3) Image/Auto Con­trast, (4) Image/Adjustments/Invert

Were the sharp col­ors used by Matisse to cov­er his wife’s face only an extrav­a­gance of a painter seek­ing new forms of expres­sion? Or maybe they were not ful­ly con­scious­ly used by him in this way, because of strong neg­a­tive emo­tions. There were plen­ty of rea­sons, as the begin­ning of the cen­tu­ry was a real night­mare for the Maisse fam­i­ly. Four years after their wed­ding, they were com­plete­ly deprived of their liveli­hood due to the bank­rupt­cy of Amélie’s par­ents. They left Paris and set­tled in the coun­try­side. Crit­ics com­plete­ly reject­ed the works of the not-so-young painter. Matisse was close to sui­cide. The last resort was the sib­lings of Amer­i­can art col­lec­tors Gertrud and Leo Stein. They recog­nised the val­ue of work by Matiss and began to buy his paintings.

The Steins helped not only Matiss and his fam­i­ly, but also anoth­er excep­tion­al artist who with great effort broke through the con­crete ceil­ing of art crit­i­cism in the ear­ly 20th cen­tu­ry. He was Pablo Picasso.

Where is the cross in the Crucifixion scene?

In con­clud­ing this chap­ter, I would like to recall one more exam­ple of a paint­ing illus­trat­ing the diver­gence of forms cod­ed with oppos­ing col­ors and lumi­nance. It is about the Cru­ci­fix­ion, a pic­ture paint­ed by Pablo Picas­so in 1930 (Fig. 133 A). Picas­so paint­ed almost all the char­ac­ters and objects on it with sharp col­ors. Only the cen­tral­ly locat­ed scene of the title cru­ci­fix­ion is white. The con­tour of the cross and the dis­tort­ed sil­hou­ettes of Christ and most like­ly Mary on a white back­ground pro­vokes to fill it with col­or, as in a chil­dren’s col­or­ing book or stained glass.

Fig­ure 133 A. Pablo Picas­so, Cru­ci­fix­ion (1930). Musée Picas­so, Paris, France [51.5 x 66.5 cm]

The mono­chrome ver­sion of the paint­ing, like the Woman with a Hat by Hen­ri Matisse, reflects the loca­tion of all objects, char­ac­ters and parts of them in the scene, just as in the col­ored ver­sion (Fig. 133 B). From this point of view, it can be con­clud­ed that bright col­ors are only an extrav­a­gance of a painter who con­tests the clas­si­cal stan­dards of fig­u­ra­tive paint­ing. Only see­ing this work through the lens of a red-green fil­ter (Fig. 133 C) and yel­low-blue fil­ter (Fig. 133 D) reveals a sur­pris­ing reg­u­lar­i­ty. In both ver­sions the cen­tral scene of the cru­ci­fix­ion dis­ap­pears. We can see a kind of cur­tain or just an emp­ty sur­face. This is the oppo­site sit­u­a­tion to the effect obtained by Claudie Mon­et in Impres­sion. While there, the bright­ness of the sun was very sim­i­lar to the bright­ness of the grey-blue back­ground, and only the col­ors sig­nalled its pres­ence in the paint­ing, in the work of Picas­so, con­trasts of bright­ness allow to eas­i­ly iden­ti­fy the cen­tral scene, but not the color.

Fig­ure 133 B. Picas­so’s paint­ing in a mono­chrome ver­sion. Graph­ic design: P.F. based on Fig. 133 A. Pro­ce­dure: Image/Mode/Lab Col­or [Lumi­nance]
Fig­ure 133 C. Mono­chrome ver­sion of Picas­so’s paint­ing pre­sent­ed in Fig. 133 A, illus­trat­ing the dis­tri­b­u­tion of red (black) and green (white) sat­u­ra­tion on the sur­face of the paint­ing. Graph­ic design: P.F. Pro­ce­dure: (1) Image/Mode/Lab Col­or, (2) Delete Chan­nel [Light­ness], Delete Chan­nel [b], (3) Image/Auto Con­trast, (4) Image/Adjustments/Invert
Fig­ure 133 D. Mono­chrome ver­sion of Picas­so’s paint­ing pre­sent­ed in Fig. 133 A, illus­trat­ing the dis­tri­b­u­tion of yel­low (black) and blue (white) sat­u­ra­tion on the sur­face of the paint­ing. Graph­ic design: P.F. Pro­ce­dure: (1) Image/Mode/Lab Col­or, (2) Delete Chan­nel [Light­ness], Delete Chan­nel [a], (3) Image/Auto Con­trast, (4) Image/Adjustments/Invert

Table of content