The Colour Imagery of Poets

Archives Suisses de Neurologie, Neurochirurgie et de Psychiatrie, 91/1, 1963

Dedicated to Manfred Bleuler


There are reasons for thinking that normal people differ in the extent to which they use colour imagery and in the degree to which they are attracted by colours of different parts of the spectrum. Painters show a great range of differences from individual to individual in their capacity to use contrasts and harmonies of colour in a creative way. There are also popular beliefs that our liking for different colours is affected by the temporary state of mood, dark and drab colours being preferred in times of relative depres­sion. Furthermore it may well be the case that our habits of using colour imagery, our preferred range of colours, and even our actual abilities to handle problems of colour may show changes from childhood on to later ages.

    Considerations of such a kind suggested an enquiry into the use of colour by poets, in the first place to discover whether there were material differ­ences between individuals, and consistencies within individuals; in the second place to see whether any idiosyncrasies in the use of colour imagery could be related to other features of personality. What is here reported is a provisional report on the early stages of this enquiry. The method adopted consisted in reading a volume of the works of one of the poets, either from the begin­ning or, in the case of a voluminous writer such as Tennyson, from an arbitrarily chosen page. Every colour adjective or noun or even verbal form encountered from that page on was then noted, until a total of 100 colour references had been accumulated. In some cases a second sample was taken as a check on the first.

    From early on it became clear that the extent to which colour imagery was made use of was quite independent of the esteem in which the poet is held, and whether he is or is not regarded as a major poet. This quality, the addiction to the use of colour, also varies very greatly from poet to poet and seems to be largely an individual characteristic. As this emerged, counts were made of the number of lines of verse required to accumulate 100 colour references, and the mean number of lines per colour reference is given in Table 1. From this table it will be seen that Milton, Marlowe, Poe, Arnold, Browning and Shakespeare are relatively sparing in their use of colour, while Shelley and Keats are relatively abundant. The poets Hopkins and Thompson, not included in these counts, were also prolific in the use of colour, in at least some part of their work.


     This draws attention to the fact that the use of colour varies with the nature of the work. In his earlier poems, closely concerned with nature, Hopkins makes abundant use of colour; but much less so in his religious poems. Shakespeare and Marlowe use less colour, and more sombre colours, in their dramatic works than in their poems.

    Nevertheless there seems to be a good (teal of individual consistency in Hic density of colour references in the work of an individual poet. Thus a sample from the earlier poems of Keats required 1453 lines of verse to provide 100 colour references; the same number was provided by 1911 lines from later poems. Two samples from Shelley, not distinguished by an age difference, required 1460 lines in the first and 1618 lines in the second to collect 100 colours. Both of these poets are fairly abundant in their use of colour.

    Though both Shelley and Keats make much use of colour, they differe from one another in the sort of words they employ, Shelley uses for the greater part straightforward and commonplace words: yellow, blue, snowy, purple, green, grey, white, black, golden, hoary, dun, azure, etc., and very rarely such exotic terms as "moonlight‑coloured". Keats is much freer with such words, and phrases as vermeil, damask, verdurous, Tyrian, rubious‑argent, ruddy gules, volcanian yellow, etc. Furthermore he piles colour on colour in a way equalled, in the poets tested, only by Francis Thompson and Gerard ManleyHopkins:

            "There the kingfisher saw his plumage bright
            Vieing with fish of brilliant dye below;
            Whose silken fins, and golden scales' light
            Cast upwards, through the waves, a ruby glow;
            There saw the swan his neck of arched snow,
            And oared himself along with majesty;
            Sparkled his jetty eyes; his feet did show
            Beneath the waves like Afric's ebony,
            And on his back a fay reclined voluptuously."

     Francis Thompson is also a poet who revels in colour and loves to find out‑of‑the‑way expressions to paint his picture: crocean, amethystine, gold‑mailed, purple‑foaincd, greening‑sapphire, etc. His touch with colour is more subtle that than of Keats, and he can cope both with a torrent of colour and a mere whisper. Thus he has given us both:

           "Summer set. lip to earth's bosom bare,
           And left the flushed print in a poppy there: 
           Like a yawn of fire from the grass it came,
           And the fanning wind puffed it to flapping flame.
           With burnt mouth, red like a lion's, it drank
           The blood of the sun as he slaughtered sank,
           And dipped its cup in the purpurate shine
           When the Eastern conduits ran with wine."

 and also, in much quieter tones:

           "Clarified silver; greens and azures frail 
           As if the colours sighed themselves away,
           And blent in supersubtle interplay
           As if they swooned into each other's arms;
           Repured vermilion,
           Like ear‑tips 'gainst the sun."

      Hopkins goes even further than Thompson with the extravagance and tue eccentricity of the words and phrases he uses in his palette. Among his whites we find glassy white, white‑fiery, whitebeam, champ‑white, silver­surfed, lily‑coloured, swan‑fledged. Colds and yellows include gold‑water, fall‑gold, goldy, quick‑gold, gold‑vermilion, gold‑wisp. The reds include crimson‑white, blood‑light, pansy‑dark, crimson‑golden, cnmson‑cresseled, rosy‑budded, rosy‑pale, rose‑flake, blood‑vivid, dappled‑with‑damson, strawberry‑breasted. With the blues there are molten‑blue, turquoise­gemmed, hyaline and pencilled blue, jay‑blue, blue‑beating, blue‑bleak, azurous, azuring‑over, azured, glass‑blue, sapphire‑shot, etc.

    Those colours which a poet uses to a greater extent than his fellow‑poets are also, not unnaturally, those in which he shows greater facility arid variety. Thus Matthew Arnold has a penchant for white, and he uses also silver, snow, pearl, milk, moon‑silvered and moon‑blanched. Brown, a colour few are fond of, has an appeal for Francis Thompson, who uses the varieties wood‑browned, rosied‑brown, dun, tan, russet, rust, bronzed, dusk and dusked.

    The use of colour expressions by Thompson and Hopkins seems to have been much influenced by other sensory qualities, for instance the musical and euphonious quality of the word of word‑combination. Physical and tactile qualities seem to have played a part with Milton, a poet who is not abundant with his use of colour. He shows, however, a preference for mineral images, and goes hardly at all, as so many poets do, to flowers or living things for the source of an image. Thus in the sample tested gold, golden and gilds occurred 30 times, amber and crocus once each, and the word yellow not at all. White is represented not only by white and snowy, but also with much greater richness by argent, silver, diamond, pearl, crystal, alabaster, ivory and opal. Other precious stones used for colour images are sapphire, jasper, topaz, chrysolite, ruby and carbuncle. The only flower colours which appeared were crocus, hyacinthine, iris, violet, roses and amarant.

    If we leave out of account individualities of word form, and concentrate on a grouping of the colours used in rather broad categories, we arrive at the figures shown in Table 1. The first thing to strike one on looking at this table is the differences between the colours themselves. It is, perhaps, not very surprising that dull and drab colours like grey and brown find little employment; but purple, a colour of great richness, is also little used. In respect of these colours, the differences from poet to poet are small on the absolute scale, but proportionately large; however, with such small numbers a good part of the variation may be random. The colours best suited for testing variation from poet to poet are the colours classified together under the general names: white (388 mentions), gold (336), red (312), green (200), black (175), blue (161). It is very striking that the warm colours, the reds and golds, are so much more used than the cool colours, blue being used even less than its warmer companion green.

    However, there are large differences between individuals in this respect. The minor poet Edward Lear, who was also an artist in line and colour, shows striking preference for blues and greens; and the same tendency is shown hi lesser degree by Coleridge and Shelley. There is a strong suggestion in the figures that variation in this respect is not unimodal, and that the poets separate into two classes, the larger group with preference for warnt colours, the smaller one with preference for blues and greens. This point can be checked by a table of intercorrelations (see Table 2).


    This shows a general tendency for intercorrelations to be negative iii sign, which is not unexpected in view of the fact that a large count in respect of any one colour necessarily brings with it a smaller total of counts for all other colours. Despite this tendency, the uses of blue and green colour references show a large positive correlation. A smaller positive correlation is also shown between white and black. The table suggests the possibility of further examination by means of factor analysis; but such a procedure would be better based on larger samples.

    Reviewing our findings so far, we find evidence of quantitative variation in the extent to which colour imagery is used, and possibly of qualitative variation in respect of the colours held by the individual poet in particular J)relerence. If we attempt to find some association between these qualities aim) the mental qualities of the individual poets, there is a suggestion thai the total extent to which colour is used is positively correlated with the sensuous or eidetic qualities of the poet's mind, negatively with his tendency to go into the abstract. Thus the poets who make relatively little use of colour, Browning, Milton, Shakespeare, are of an intellectual stamp; whereas those who use colour abundantly, Shelley, Keats, Thompson, are rather mcii of feeling. The qualitative difference, on the other hand, distinguishing the lovers of red and yellow and the lovers of blue and green, does not show any obvious sign of being correlated with temperamental traits. In this case H. seems probable that we are dealing with a constitutional difference of an independent kind. It would be a matter of much interest to invrstig;t lr this phenomenon along genetical lines.

    The suggestion arising from this enquiry, that each poet is is to some extent a law unto himself, and that with his use of colour he sets something in the way of a signature on his work, is one which is readily tested. A chance of such a kind arose in the material which was examined in the course of this enquiry with the poem 'Hero and Leander". The first two sestiads of this poem are known to have been written by Marlowe; but he left the poem unfinished, and it was eventually completed by Chapman. Although Chap­man clearly endeavoured to model himself on Marlowe, he was not so success­ful as a poet. Experts on style would no doubt be able to distinguish the work of the one author from the work of the other; and it is worth asking whether the two sections of the poem are also distinguishable in the use made of colour imagery. The relevant data are set out in Table 3.


      As will be seen from this table, Chapman made much more abundant use of colour than Marlowe did; the proportion, total number of colour references divided by total number of lines, is for Chapman 0.06615, for Marlowe 0.04401. The difference, 0.02214, is 2.32 times as great as its standard error, with probability as a chance finding less than 0.02. The colour differences may also be checked, in view of the suggestive observation that Marlowe makes extremely little use of black in comparison with Chapman. Throwing together the three colours purple, blue and green, all with small numbers of observations, there is a 72 for 3 degrees of freedom of 7.26, with probability slightly exceeding 0.05. This example suggests that, in suitable circumstances, the use of colour by an author might be worth investigating in cases where there was a question of disputed authorship.